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
|
|---|---|---|---|---|---|
Toll-like receptor signaling requires functional Toll/interleukin-1 ( IL-1 ) receptor ( TIR ) domains to activate innate immunity . By producing TIR homologous proteins , microbes inhibit host response induction and improve their own survival . The TIR homologous protein TcpC was recently identified as a virulence factor in uropathogenic Escherichia coli ( E . coli ) , suppressing innate immunity by binding to MyD88 . This study examined how the host MyD88 genotype modifies the in vivo effects of TcpC and whether additional , TIR-domain containing proteins might be targeted by TcpC . In wild type mice ( wt ) , TcpC enhanced bacterial virulence , increased acute mortality , bacterial persistence and tissue damage after infection with E . coli CFT073 ( TcpC+ ) , compared to a ΔTcpC deletion mutant . These effects were attenuated in Myd88−/− and Tlr4−/− mice . Transcriptomic analysis confirmed that TcpC inhibits MYD88 dependent gene expression in CFT073 infected human uroepithelial cells but in addition the inhibitory effect included targets in the TRIF and IL-6/IL-1 signaling pathways , where MYD88 dependent and independent signaling may converge . The effects of TcpC on bacterial persistence were attenuated in Trif −/− or Il-1β −/− mice and innate immune responses to ΔTcpC were increased , confirming that Trif and Il-1β dependent targets might be involved in vivo , in addition to Myd88 . Furthermore , soluble TcpC inhibited Myd88 and Trif dependent TLR signaling in murine macrophages . Our results suggest that TcpC may promote UTI-associated pathology broadly , through inhibition of TIR domain signaling and downstream pathways . Dysregulation of the host response by microbial TcpC thus appears to impair the protective effects of innate immunity , while promoting inflammation and tissue damage .
Toll-like receptors ( TLRs ) control innate host responses to mucosal and systemic infections and signaling involves the intracellular Toll/interleukin-1 receptor ( TIR ) domain [1] . Following ligand binding , signaling is initiated by the recruitment of adaptor proteins to the TIR domain [2] , [3] , [4] , including myeloid differentiation factor-88 ( MYD88 ) , MYD88 adapter-like protein ( Mal ) , TIR domain-containing adaptor protein inducing IFNβ ( TRIF ) , TRIF-related adaptor molecule ( TRAM ) and the sterile α- and armadillo-motif-containing protein ( SARM ) . Negative regulators of TLR signaling include SIGIRR , MyD88s and IRAK-M , which block MyD88 dependent activation , or Triad3A and SARM , which block the TRIF dependent pathway . The SIGIRR TIR domain resembles MyD88 but lacks two amino acids needed for signaling to occur [5] , [6] . However , TIR-TIR interactions between SIGIRR and TLR4 prevent the recruitment of IRAK and TRAF6 to MyD88 [6] . MyD88s is a splice variant inhibiting MyD88 dependent TLR4 activation by allowing MyD88 to bind the intermediate IRAK-binding domain without inducing IRAK phosphorylation and NF-κB activation [7] . IRAK-M prevents IRAK and IRAK-4 dissociation from MyD88 and TRAF6 complex formation [8]; Triad3A interacts with TIR domains of TLRs , TRIF , TIRAP and RIP1 [9]; and SARM blocks gene induction downstream of TRIF [10] . Competition at the level of the TIR domain is thus used by host cells to modify TLR signaling in response to specific agonists [6] , [7] , [11] , [12] . Pathogens have also evolved mechanisms to inhibit the TLR dependent host defense and to increase their fitness and virulence for a specific host niche [12] . The TIR domain plays a crucial role in the mammalian innate immune response and recently proteins containing TIR domains have been described in a wide variety of bacteria , fungi , archaea and viruses [13] . Whole genome sequencing and structural studies have revealed that several pathogens carry TIR-domain homologous sequences , including two proteins from Vaccinia virus A46R and A52R , which interfere with IL-1 and TLR4 mediated activation of NF-κB [14] . Similar proteins were identified in Salmonella , Brucella and uropathogenic E . coli ( UPEC ) [12] , [15] , [16] . On the other hand , Spear and co-workers suggested that most TIR domains in bacteria did not evolve to subvert the function of eukaryotic cells but simply to function as general purpose protein-protein interaction domains for diverse uses [13] . We recently showed that the TIR homologous protein TcpC in the UPEC strain E . coli CFT073 acts as a virulence factor by suppressing innate host responses in the kidney , enhancing bacterial persistence and tissue damage [12] . Epidemiologic studies of patient E . coli isolates showed that TcpC occurred more frequently in strains causing severe kidney infections than in E . coli causing other forms of urinary tract infection ( UTI ) [12] . The UTI model is therefore quite suitable to investigate the mechanisms of TcpC-modulation of the innate host response in vivo and the consequences for bacterial persistence and disease severity [11] . TcpC binds to MyD88 [12] but it remains unknown if other TIR-domain containing molecules of the host are influenced by TcpC in vivo . This study addressed this question with the aim to define the genetic control of TcpC mediated immune inhibition in vivo . We used the murine UTI model and mice lacking specific innate immune response genes to examine whether Myd88 controls the TcpC dependent response to UPEC infection and if additional Myd88 independent signaling pathways might modify the effects of TcpC in vivo . Furthermore , transcriptomic and proteomic analysis of infected human epithelial cells was used to define targets of TcpC and effects on innate immune activation by UPEC and inhibition of TLR signaling by soluble TcpC protein was defined in murine macrophages . The results suggest that TcpC partially inhibits TIR domain dependent signaling in vivo and in host cells including pathways downstream of MYD88 , where MYD88 dependent and independent innate immune responses may converge . In this way , broad but incomplete suppression by microbial TcpC , may impair innate immune protection , while promoting inflammation and tissue damage .
Experimental UTI was established in wild type ( wt ) C57BL/6 mice by intravesical infection with E . coli CFT073 ( CFT073 ) or the CFT073tcpC::kan mutant ( ΔTcpC ) [12] and was monitored for seven days . A higher bacterial burden was present after CFT073 infection compared to ΔTcpC ( p<0 . 001 , Figure 1B ) in urine samples obtained daily after infection . This difference was reproduced in kidneys ( p<0 . 001 ) obtained on days four or seven after infection ( Figure 1A ) . MIP-2 chemokine concentrations and neutrophil responses in urine were higher in mice infected with CFT073 compared to ΔTcpC ( p<0 . 001; Figure 1C , D ) . Tissue damage was extensive in mice infected with CFT073; the kidneys were large , pale and soft and abscesses were present in 75% of the organs ( Figure 1E , F ) . In contrast , kidneys of mice infected with ΔTcpC were normal in size , color and texture and lacked detectable abscesses ( p<0 . 05 , Figure 1E , F ) . Experimental UTI with CFT073 or the ΔTcpC mutant was subsequently established in Tlr4 and Myd88 adaptor protein knockout mice in the C57BL/6 background ( Tlr4−/− and Myd88−/− respectively ) . Total bacterial counts and the TcpC dependent difference in bacterial persistence were reduced in the mutant mice , compared to wt mice ( Figure 1 ) . Bacterial counts in urine were lower in Myd88−/− ( p<0 . 05 ) and Tlr4−/− ( p<0 . 05 ) mice as compared to wt mice ( Figure 1 ) . In addition , the renal bacterial counts were significantly lower in mutant mice on day seven than in wt animals which were infected with CFT073 ( p<0 . 05 , Figure 1A ) . However , there was no difference in bacterial counts in wt and mutant mice which were infected with ΔTcpC ( Figure 1A ) . The MIP-2 and neutrophil responses were absent in infected Myd88−/− mice and drastically reduced in Tlr4−/− compared to C57BL/6 wt mice ( p<0 . 0001 , Figure 1C , D ) , confirming that these aspects of the early innate immune response require Myd88 and Tlr4 dependent signaling . Kidneys of Tlr4−/− and Myd88−/− mice infected with CFT073 or the ΔTcpC mutant were normal in size , color and texture and had no abscesses ( Figure 1E , F ) . TcpC-related differences in bacterial persistence were observed also in the mutant mice ( Figure 1 ) . Tlr4−/− and Myd88−/− mice developed significant bacteriuria ( ≥105 CFU/ml of urine ) six hours after infection with CFT073 or ΔTcpC and bacteria persisted in urine until the experimental end point . Higher numbers of CFT073 than ΔTcpC were observed on day one in Tlr4−/− mice ( p<0 . 01 ) ; and on days one , four and six in Myd88−/− mice ( p<0 . 01 , Figure 1B ) . In addition , CFT073 numbers were higher than ΔTcpC in kidneys of Myd88−/− and Tlr4−/− mice ( p<0 . 01 ) on days four and seven ( Figure 1A ) . The difference between CFT073 and ΔTcpC was reduced compared to wt mice in Tlr4−/− ( p<0 . 01 ) and Myd88−/− mice ( p<0 . 01 ) by more than two logs in urine ( Figure 1A ) . TcpC dependent increases in MIP-2 and neutrophil responses observed in wt mice were lost in Tlr4−/− mice , the , confirming the essential role of TLR4 and its TIR domain for innate immune responses to UTI ( Figure 1C and D ) . The Myd88−/− mice did not mount MIP-2 and neutrophil responses to CFT073 or ΔTcpC infection . These results suggest that TcpC affects bacterial persistence and pathology , in part , via Tlr4 and Myd88 dependent but also via Myd88 independent pathways . Kidney sections from CFT073 infected wt mice ( htx-eosin , day 7 ) showed swollen collecting ducts and inflammatory cell infiltrates into the kidney parenchyma ( Figure 2A ) . P-fimbriated bacteria were present in the tissues , from the pelvic region to the cortex , as shown by PapG specific antibody staining . Neutrophils were abundant in the abscesses and scattered throughout the tissue , as shown by a neutrophil specific antibody ( areas AI-AIII in Figure 2A ) . By dual staining , P fimbriae and neutrophils were detected in the abscesses and collecting ducts ( Figure 2A ) . In contrast , kidney sections from mice infected with ΔTcpC showed normal structure , no bacteria and few inflammatory cells ( areas BI-BII in Figure 2B ) . Figure 2C shows htx-eosin stained sections from an uninfected control kidney and the inset shows a negative control section stained with the anti-neutrophil and anti-PapG antibodies . The results suggest that TcpC reduces host resistance and increases inflammation , resulting in a high bacterial burden and tissue damage as these sequels become attenuated in mice infected with the ΔTcpC mutant . Htx-eosin stained sections from Myd88−/− and Tlr4−/− mice infected with CFT073 showed normal collecting ducts , few inflammatory cells and no bacteria in the medulla or cortex of kidneys from either host ( Figure 3A ) . P fimbriae or neutrophils were not detected in infected kidneys by immunohistochemistry ( Figure 3B ) . Similarly , there was no tissue pathology in Myd88−/− and Tlr4−/− mice infected with the ΔTcpC mutant . The results suggest that a host response involving Tlr4 and Myd88 leads to TcpC dependent kidney pathology after CFT073 infection and that hosts lacking Tlr4 or the adaptor are protected from such tissue damage . While the in vivo experiments confirmed that TcpC mediated virulence depends on pathways controlled by Myd88 , they also suggested that TcpC modifies additional host response pathways . To identify such pathways , human A498 kidney epithelial cells were infected in vitro with CFT073 or ΔTcpC ( 4 hours , 108 CFU/ml ) and complementary RNA was hybridized to Illumina whole genome microarrays . A heat map of significantly regulated genes is shown in Figure 4A ( means of triplicate spots ) . In identification of CFT073 or ΔTcpC-specific genes , fold changes of ≥2 were used . For the comparison of CFT073 or ΔTcpC , a fold change in response to either strain >1 . 5 relative to the respective control was used . Infection stimulated marked changes in gene expression and three major groups of genes were altered; 734 regulated genes were shared , 21 genes responded only to CFT073 , while 11 responded only to ΔTcpC ( Figure 4B , Supplementary Table S1 ) . Differentially expressed genes between unstimulated and CFT073 or ΔTcpC treated cells were then characterized using Ingenuity Pathway Analysis . Notably , signaling downstream of pattern recognition receptors , interferon induction , interferon response and IL-6/IL-1 signaling pathways were among the top-scoring pathways identified ( Figure 4C ) . To further study the mechanisms of TcpC mediated innate immune inhibition , differentially transcribed genes were assigned to known response pathways ( Figure 5 , Supplementary Table S1 ) . Consistent with the proposed role of TcpC as a MYD88 inhibitor [12] , in vitro infection with ΔTcpC upregulated MYD88 dependent transcripts involved in pathogen pattern recognition , compared to CFT073 infected cells ( p>0 . 01 ) . These include the inflammatory cytokines IL-6 , IL-8 , TNF-α , IL-1α/β and the transcription factors IRF7 and NF-κB1 , 2 ( Figure 5A , Green ) . In addition , the expression of downstream pro-inflammatory genes including IL-1α/β , TNFAIP6 and TNFRSF11b were upregulated in ΔTcpC compared to CFT073 infected cells ( Figure 5B , Green ) , while the transcription of negative regulators IL-1RA , NFkBIa and NFkBIb was reduced , consistent with activation of the MYD88 pathway ( Figure 5B , Red ) . Interestingly , a corresponding reduction in the MAP kinase MAP2K3 , JUN and FOS transcripts was observed , suggesting a partial rather than complete suppression of the MYD88 pathway by TcpC ( Figure 5B , Red ) . NF-κB and IRF7 transcript levels , which are both MYD88 and TRIF dependent , were significantly increased after ΔTcpC infection compared to CFT073 ( p<0 . 01 ) , suggesting that also TRIF-dependent signaling , might be inhibited by TcpC . Other genes downstream of TRIF maintained their activity regardless of TcpC , however . In order to validate the transcriptomic analysis , expression levels of selected genes were confirmed by RT-PCR in the cells infected with CFT073 or ΔTcpC ( Figure 6A ) . Significant differences were observed for IL-8 , IL-6 , NFkB1 , TNF- α and c-FOS . The pattern reflected a trend similar to that revealed by microarray analysis for the gene products tested ( Figure 5B* , Red ) . To generate further insights into the effects of TcpC , cultured human kidney cells were stimulated with either CFT073 or ΔTcpC and culture medium was assayed using the Procarta human cytokine kit ( Figure 6B ) . ΔTcpC stimulation resulted in significantly elevated levels of MYD88 dependent proinflammatory cytokines ( IL-6 , IL-8 , TNF-α ) ( p<0 . 01 ) compared to the wild type strain , corroborating the transcriptomic analysis and suggesting that MyD88 and TRIF dependent pathways are modified by TcpC . In addition , ΔTcpC stimulation resulted in increased expression of the neutrophil chemoattractants MCP-1 , GRO-α and MIP-3α ( Figure 6B ) , while in vivo infection with ΔTcpC caused a lower MIP-2 response than CFT073 . The apparent discrepancies between the in vivo data and in vitro results were further examined , by establishing murine tubular kidney cell lines from wt and Myd88−/− or Trif Lps2/Lps2 mutant mice . The cells were infected with CFT073 or ΔTcpC and MIP-2 secretion was quantified by ELISA , four hours after infection ( Supplementary Figure S1 ) . MIP-2 secretion was reduced in ΔTcpC infected compared to CFT073 infected cells from wt mice , consistent with the increased response to CFT073 in wt mice . In Trif Lps2/Lps2 cells , MIP-2 secretion was also reduced in ΔTcpC compared to CFT073 infected cells ( Supplementary Figure S1 ) . Furthermore , the in vitro response of cells from Myd88−/− mice was very low , both to CFT073 and ΔTcpC , thus reproducing the lack of response in mutant mice . By RT-PCR , the MIP-2 response to ΔTcpC was reduced compared to CFT073 in cells from wt mice and also very low in cells from Myd88−/− mice but not different in cells from Trif Lps2/Lps2 mice ( data not shown ) , Thus most but not all of the in vitro results were compatible with the in vivo data , either in human or murine kidney cells . Differences between in vivo data and in vitro assays are expected to occur , as the in vivo response of an entire organ system is unlikely to be reflected by a single cell type in vitro . The cellular studies and findings in Myd88−/− and Tlr4−/− mice suggested , that bacterial TcpC might inhibit TLR4 dependent signaling , also through the Trif adaptor . To examine this hypothesis , Trif adaptor protein knockout mice ( Trif −/− ) were infected with CFT073 or the ΔTcpC mutant and compared to wt C57BL6/129 mice . The ΔTcpC mutant established higher bacterial counts in urine than CFT073 in Trif −/− mice ( p<0 . 001 ) , in contrast to C57BL6/129 wild type mice ( Figure 7B , p<0 . 001 ) . Furthermore , the difference in kidney counts between CFT073 and ΔTcpC in wt mice ( Figure 7A , p<0 . 001 ) was not observed in Trif −/− mice , indicating that the effects of TcpC on bacterial persistence are neutralized in mice with defective Trif signaling . Trif −/− mice also exhibited a significant MIP-2 response to ΔTcpC compared to CFT073 ( p<0 . 05 , day 3 and 6 ) ( Figure 7C ) and the neutrophil response to ΔTcpC was significantly higher ( p<0 . 001 , Figure 7D ) . The effects of Trif on TcpC inhibition were confirmed in Trif Lps2/Lps2 mice , which carry a non-functional co-dominant Trif allele , induced by N-ethyl-N-nitrosourea mutagenesis on a pure C57BL/6 background [17] ( Figure 7 ) . As in Trif −/− mice , the ΔTcpC mutant established higher bacterial counts in urine than CFT073 in contrast to C57BL/6 wt mice ( Figure 7B , p<0 . 001 ) . The TcpC related difference in kidney counts between CFT073 and ΔTcpC in wt mice ( Figure 7A , p<0 . 001 ) was not observed in Trif Lps2/Lps2 mice and the MIP-2 and neutrophil responses to ΔTcpC were significantly higher compared to CFT073 ( Figure 7C , D; p<0 . 01 ) . However , neither CFT073 nor ΔTcpC induced kidney abscesses in Trif −/− mice or in Trif Lps2/Lps2 mice ( Figure 7E , F ) . The results suggest that the Trif adaptor protein is involved in the innate immune mechanisms controlling the persistence of TcpC expressing bacteria . To confirm that the TIR domain of TcpC impaired MyD88-dependent TLR signaling , bone marrow derived macrophages ( BMDMs ) from wild type or Myd88−/− mice were stimulated with different TLR ligands in the presence of titrated amounts of TIR-TcpC , the c-terminal half of TcpC containing the TIR domain . TIR-TcpC impaired TNF secretion induced by the different TLR ligands with the exception of TLR3 mediated stimulation , as expected from the MyD88 independence of TLR3 ( Figure 8A ) [12] . Also as expected only poly ( I:C ) and LPS were able to induce TNF secretion in Myd88−/− BMDMs , presumably via Trif . Interestingly , TIR-TcpC impaired this pathway as well , consistent with the in vivo observation that the Trif pathway was influenced by TcpC ( Figure 8B ) . In addition , control experiments showed that TcpC is secreted into the urine of infected mice ( data not shown ) . The transcriptomic analysis suggested that TcpC strongly regulates the pro-inflammatory cytokines IL-6 and IL-1α/β , as well as downstream signaling pathways . Enhanced expression of IL-1α/β ( p<0 . 03 for IL-1α and p<0 . 02 for IL-1β ) in ΔTcpC infected human cells suggested that the inhibitory effect of TcpC includes IL-1 dependent inflammation . To address this question , Il-1β−/− mice were infected with CFT073 or ΔTcpC , as described . The TcpC dependent difference in bacterial persistence and host response induction was reduced in these mice ( p<0 . 0001 , Figure 9 ) . Renal abscess formation or tissue pathology was not observed . IRF3 is a transcription factor controlling interferon responses to viral infection [18] . More recently , the involvement of IRFs in immunoregulation by TLRs has received more attention , since NF-kB , IRF3 and AP-1 form transcriptional complexes that regulate innate immune responses in monocytes [19] . We have recently identified IRF3 as an essential transcription factor in UTI , acting downstream of TLR4/TRAM ( unpublished observation ) . To examine if Irf3 might be involved in TcpC mediated innate immune suppression , infection with CFT073 or ΔTcpC was established in Irf3 mutant mice ( Irf3−/− ) , using wt C57BL/6 mice as controls . The difference in persistence between CFT073 and ΔTcpC in wild type mice was not observed during the early phase of infection in Irf3−/− mice ( <day 5 ) ( Figure 9A , B ) . Furthermore , in Irf3−/− mice , the early chemokine and neutrophil responses were delayed compared to responses in wt controls ( Figure 9C , D ) . Significantly reduced responses to ΔTcpC were only observed from day five post-infection in Irf3−/− mice ( Figure 9C , D ) , showing that the response kinetics differed from wt mice . Abscess formation in response to CFT073 was as frequent in the Irf3−/− as in wt mice ( Figure 9E , F ) . The results suggest that the effect of TcpC on bacterial persistence and on the MIP-2 response are attenuated during the early phase of infection in IRF3-deficient mice and that IRF3 signaling is differentially activated depending on the TcpC status of the infecting strain ( Please see figure 9B–D ) . Taken together , the results show that bacterial TcpC modifies the innate host response broadly through inhibition of Tlr4 , Myd88 , Trif , IL-6/IL-1 and Irf3 dependent antibacterial effector functions .
Bacterial TIR-like proteins are important virulence factors , which act by inhibiting innate immunity , thus facilitating the survival and persistence of several pathogens . The Salmonella enteritica TlpA protein enhances bacterial survival in macrophages and mice [15] and the Brucella TcpB protein inhibits TIRAP mediated signaling and reduces systemic spread of bacteria during the early stages of infection [16] . The E . coli TcpC protein increases virulence in the urinary tract and we have previously proposed that TLR signaling is impeded through the MYD88 adaptor via direct binding of TcpC to MYD88 [12] . This study addressed the mechanism through which TcpC modifies the innate immune response in the infected host , by varying the innate immune genetic repertoire . The results show that TcpC requires Myd88 to act as a virulence factor in vivo . Transcriptomic analysis identified additional targets for TcpC in human cells , including the TRIF and IL-6 pathways , as well as IL-1α/β . Tlr4−/− , Trif −/− , Lps2−/− and Il-1β−/− mice exhibited markedly different immune responses to TcpC stimulation and the TcpC dependent change in bacterial persistence and pathology was attenuated in these mice . The results thus suggest that pathways for host-defense can be fine-tuned by a bacterial virulence factor in order to paradoxically promote bacterial replication and pathology . MyD88 was the first TLR adaptor to be identified and is shared by the TLRs as well as the interleukin-1 receptor [20] , [21] . By targeting MyD88 , TcpC would thus be expected to impair both TLR- and IL-1 dependent signaling pathways as well as related antibacterial effector functions . This interpretation was supported by the in vivo effects , which were not restricted to Myd88 but regulated by a group of genes encoding proteins with a TIR domain or regulated by such proteins . In Myd88−/− mice , essential chemokine and neutrophil responses to CFT073 infection were completely abrogated consistent with the importance of MyD88 for the overall innate immune response to these infections . The response to infection was strongly reduced also in Tlr4−/− mice confirming our previous findings that Tlr4 signaling is essential for the innate immune response to UTI and suggesting that that inhibition of TLR4 responses may be protective at the mucosal level [11] , [22] , [23] , [24] . We have recently extended the analysis of potential eukaryotic interaction partners and have shown that the TIR-domain of TcpC binds to TLR4 in addition to MyD88 ( unpublished observation ) . TcpC did not bind to TRIF or TLR2 , however [12] . Thus , as expected by the differing aminoacid sequences , only selected TIR-domains bind TcpC . The difference in bacterial persistence , while reduced , was not abrogated in Tlr4−/− or MyD88−/− mice , however , possibly reflecting the involvement of pathways that remains intact in Myd88−/− mice . The interaction with TLR4 provides a molecular basis for inhibition of the TRIF-dependent arm of the TLR4 signaling cascade , thus possibly explaining , in part , the effects of TcpC in Trif −/− mice . This mode of action of TcpC was also supported by the experiments in murine macrophages , where responses to most of the TLR specific ligands were inhibited by purified TcpC-TIR protein . TcpC inhibited LPS-driven , MyD88 dependent and LPS-driven MyD88-independent TNF secretion , including TLR4-TRIF signaling . However , TcpC did not influence poly ( I:C ) -induced TNF-secretion , whether MyD88 dependent or not . The fact that poly ( I:C ) -induced TNF-secretion in the absence of TcpC was lower in MyD88-deficient cells compared to wild type cells cannot be interpreted to imply that poly ( I:C ) stimulates MyD88-dependent TNF-secretion , since resting MyD88-deficient macrophages are impaired in their basal expression of several cytokines including TNF [25] . Cytokines like TNF are in general harder to induce in MyD88-deficient cells , as also reported by Sun et al . [26] . In preliminary experiments , we have also analyzed the secretion of the chemokine KC and have not found differences in wild type or MyD88-deficient cells after stimulation with poly ( I:C ) . Thus , poly ( I:C ) stimulates cells in a MyD88-independent but TRIF-dependent manner and the results are compatible with the in vivo effect . TRIF has a well conserved TIR domain and several TRAF6 binding regions within the N- and C-terminal RIP homotypic interaction motifs ( RHIM ) [27] , [28] , [29] . TRAM is TLR4 specific [30] , [31] and the myristoylated N-terminus associates to the plasma membrane and protein kinase Cε phosphorylation of Serine 6 and 16 is essential for TRAM activation [32] , [33] . In contrast to MYD88 , TRIF dependent signaling activates both NF-κB and IRF3/7 [30] . After TRAM dependent TLR4 activation , TRIF forms a complex with TRAF3 , IRAK1 and an IKK-like kinase . TRAF family member associated NF-κB activator ( TANK ) -binding kinase 1 ( TBK1 ) and the IKK homolog , IKKε , phosphorylate IRF3 at its C terminus [30] , [34] , initiating IFN-α/β transcription [29] . IRF7 becomes activated in a similar manner [4] . In this study , TcpC suppressed transcription of NF-κB and IRF7 as well as IL-8 , TNF-α , IL-1α/β and IL-6 , which is consistent with the effects on MYD88 and possibly TRIF . The in vivo response to infection supports the conclusion that TcpC also suppresses Trif dependent effector functions , however , as Trif −/− mice have a functional Myd88 response , except for the cooperative TLR4 response to LPS [35] . The lack of pathology in infected Trif −/− mice further suggests that Trif signaling is essential for efficient innate immune-mediated clearance of UPEC infection . The transcriptomic analysis of infected human kidney cells suggested that TcpC also modifies proinflammatory cytokine signaling downstream of MYD88 and TRIF . The IL-6 pathway was strongly regulated and IL-1α/β expression was reduced by TcpC . The involvement of IL-1 was confirmed by infection of Il-1β−/− mice and the phenotype of the Il-1β−/− mice was quite convincing , as the difference in bacterial persistence between CFT073 and the TcpC deficient mutant in wt mice was abrogated in Il-1β−/− mice and the innate immune response to the two isogenic strains was similar . Bacterial clearance was rapid , further suggesting that IL-1 may be a significant factor in the mucosal response to UPEC and in the establishment of tissue pathology . The mechanism of TcpC modulation of IL-1 is not clear , however and may either relate to a TIR domain dependent interaction of TcpC with the IL-1 receptor , effects on upstream signaling involving Myd88 and Trif and the resulting IL-1 response or to other mechanisms . Recently , Trif dependent innate immune responses have been shown to activate IRFs that regulate the transcription of pro-inflammatory genes , including IL-8 , IL-6 and TNF , in addition to interferon genes [18] . In a parallel study , we have characterized a new TLR4/TRAM dependent pathway that mediates innate responses to P-fimbriated , uropathogenic E . coli through Irf3 ( unpublished observation ) . Using a combination of transcriptomics , phosphorylation arrays and imaging technology , we detected TRAM phosphorylation , activation of MAP kinases including p38 , CREB phosphorylation and nuclear IRF3 translocation . Irf3−/− mice lacking this pathway , developed rapid lethal kidney infections with extensive tissue damage and patients prone to acute pyelonephritis were shown to carry IRF3 promoter polymorphisms that reduce transcription efficiency . In the present study , interferon dependent pathways were differentially regulated by TcpC in vitro ( data not shown ) and in Irf3−/− mice the effects of TcpC were significantly delayed compared to wt mice . While this effect was transcient , the results further support the results suggesting that TcpC may modify the effects of the TRIF/Irf3 pathway and the progression to disease and pathology . Innate immune activation by uropathogenic E . coli is orchestrated by specific virulence factors and essential aspects of the mucosal response show pathogen specificity [36] . Such interactions are needed , as innate immune responses are not activated by asymptomatic carrier strains in the epithelial tissue , which forms a barrier against interactions with inflammatory cells , with broader reactivity . For example , epithelial cells lack surface expressed CD14 and do not respond to conserved bacterial PAMPS like LPS [11] , [36] . In addition , asymptomatic carrier strains may actively suppress the mucosal immune response , but the mechanisms are not fully understood . We have previously shown that pathogen specific TLR4 activation favours TRIF or MyD88 , depending on the surface fimbriae , which are expressed in a pathogen specific manner and serve as crucial virulence factors involved in attachment and host tissue perturbation [11] . P fimbriated bacteria preferentially activate TLR4/TRIF signalling while Type 1 fimbriae trigger TLR4 responses mainly involving MyD88 . The adaptor protein usage in infected host cells can even be shifted from TRIF to MyD88 by a change in fimbrial expression [11] . E . coli CFT073 , used in the present study , expresses both P and type 1 fimbriae , thus activating TLR4 signalling involving TRIF and MyD88 responses in infected cells , providing a basis for independent targeting of these pathways by TcpC . In conclusion , our results suggest that TcpC may act as a broad microbial innate immune response modulator in vivo , by preventing bacterial clearance and dysregulating inflammation , with destructive effects on infected tissues . This adds TcpC to an increasing number of components that regulate TLR and MYD88 dependent responses . On the host side , MYD88s inhibits the recruitment of IRAK-4 , thus acting as a negative regulator of TLR signaling [37] . SIGIRR , IRAK-M , SOCS1 , Triad3A and SARM , and the cytosolic domain of SIGIRR block TLR4 signaling through TIR-TIR interactions preventing the recruitment of IRAK and TRAF6 to MYD88 . TcpC binds to TLR4 ( unpublished data ) and MYD88 [12] through TIR domain interactions and inhibits TLR4 and MYD88 dependent signaling in vivo , as well as downstream effector functions . While many molecular interactions remain to be defined , it is clear that this microbially induced suppression of the host defense differs , in the classical sense , from mucosal tolerance , which may be triggered by microbial or other mucosal antigens , but is defined by the involvement of specific immunity , with T cells as the main effector cells . In the case of TcpC , the innate immune response is modified and the “tolerant” state appears to result from active microbial inhibition , rather than from a lasting change in the immune status of the host , and from a direct modification of host resistance rather than by inducing tolerance . Further insights into these mechanisms may be helpful to distinguish “bad” from “good” inflammation and to understand how partial inhibition of MYD88 , TRIF and TLR4 by TcpC results in pathology while complete gene deletion is protective . These findings also illustrate the importance of the host response as a generator of pathology and suggest the possibility of intervention to support “good” host responses , promoting bacterial clearance and tissue integrity while inhibiting pathology .
C57BL/6 ( wt ) , C57BL6/129 ( wt ) , C57BL/6 Tlr4−/− , C57BL/6 Myd88−/− , C57BL/6 Irf3−/− , C57BL/6 Il-1β−/− , C57BL/6 Trif Lps2/Lps2 and C57BL6/129 Trif −/− mice were bred and housed in the specific pathogen-free facilities of the MIG animal facilities ( Lund , Sweden ) with free access to food and water . All procedures were approved by the Animal Experimental Ethics Committee at the Lund District Court , Sweden ( numbers M166-04 and M87-07 ) , following Institutional , National , and European Union guidelines . The human epithelial cell line A498 ( ATCC HTB44 , human kidney carcinoma , Manassas , VA , USA ) was cultured in RPMI-1640 supplemented with 1 mM sodium pyruvate , 1 mM non-essential amino acids , gentamycin ( 50 µg mL−1 ) and 5% fetal calf serum ( PAA Laboratories , Pasching , Austria ) . Alexa-fluor 568-goat anti-rabbit IgG and Alexa-fluor 488-goat anti-rat were from Invitrogen ( Eugen , Oregon , USA ) . Anti-rat NIMP-R14 ( ab2557 ) were from Abcam ( Cambridge , UK ) and goat normal serum were from Dako ( Denmark ) . The reagents paraformaldehyde ( Merck KGaA , Darmstadt , Germany ) , triton X-100 ( VWR International Ltd , England ) , isopentane ( Sigma-Aldrich , Germany ) , VECTASHIELD mounting medium ( Vector Laboratories , USA ) , poly-L-lysine-coated glass slides ( Thermo Scientific , Waltham , USA ) , DAPI ( Invitrogen , Oregon , USA ) , tryptic soy agar ( Difco , Detroit , USA ) and hematoxylin-eosin stain ( Histolab Products AB , Gothenburg , Sweden ) were used . Mouse MIP-2 quantification kit was from R&D systems ( Abingdon , UK ) . TargetAmp Nano-g Biotin-aRNA Labelling kit was from Epicentre Biotechnologies ( Madison , USA ) , RNeasy clean-up from Qiagen ( Alabama , USA ) and the Procarta Human Cytokine 50-plex kit from Panomics ( Fremont , USA ) . Tlr4−/− [38] , Myd88−/− [39] , Irf3−/− [40] and Il-1β−/− [41] mice were generated in the C57BL/6 genetic background and Trif −/− [28] mice were generated in the C57BL6/129 genetic background . Lps2 , a non-functional codominant allele of Trif , was induced by N-ethyl-N-nitrosourea mutagenesis on a pure C57BL/6 mouse background [17] in the Scripps Institute animal facilities ( La Jolla CA ) . Il-1β−/− mice were provided by Max Planck ( Institute for Infection Biology , Berlin , Germany ) . Wild type C57BL6/129 mice are the 50% backcross of C57BL/6 mice . All the knock out mice ( Tlr4−/− , Myd88−/− , Irf3−/− and Il-1β−/− ) are backcross in 100% C57BL/6 wt mice and are pure . E . coli CFT073 was isolated from the blood and urine of a woman admitted to the University of Maryland Medical System for treatment of acute pyelonephritis [42] . This hly1+ , pap1+ , sfa1+ and pil1+ strain expresses P fimbriae , hemolysin , and Type 1 fimbriae and is highly virulent in the CBA mouse model of ascending UTI . It is cytotoxic for cultured human renal proximal tubular epithelial cells [43] . CFT073 expresses the TcpC protein and the ectcp sequences , which encode TcpC and were deleted in the ectcp::kan mutant ( ΔTcpC ) , as described [12] . The strains were grown to stationary phase overnight on tryptic soy agar with appropriate selection and suspended in 10 ml of phosphate-buffered saline ( PBS ) ( pH 7 . 2 ) to generate the bacterial suspension used for inoculation ( 109 CFU , colony forming units/ml ) . The bacteria were tested for virulence factor expression , including P and Type 1 fimbriae [44] . Female C57BL/6 , C57BL6/129 , Tlr4−/− , Myd88−/− , Trif −/− , Trif Lps2/Lps2 , Il-1β−/− and Irf3−/− mice were used at 8–12 weeks of age . E . coli urinary tract infection was established as described [45] . In brief , 0 . 1 ml of the bacterial suspension was administered into the bladders of anesthetized mice ( 109 CFU/ml ) with the help of a soft polyethylene catheter ( inner diameter 0 . 28 mm , outer diameter 0 . 61 mm; Clay Adams , New Jersey USA ) . Prior to inoculation , urine samples were collected and cultured on blood agar plates to ensure that the mice were uninfected . After infection , urine was collected into sterile tubes through gentle pressure on the mouse's abdomen ( 5 h , 24 h and up to 6 days ) and neutrophils were quantified by light microscopy using a Burker chamber . The urine samples and serial dilutions were quantitatively cultured on tryptic soy agar ( TSA ) plates . Mice were sacrificed by cervical dislocation while anesthetized . Bladders and kidneys were removed under sterile conditions , placed into a plastic bag containing 5 ml PBS ( pH 7 . 2 ) , homogenized in a Stomacher 80 homogenizer ( Seward medical , London , UK ) and plated on TSA plates . Subsequently , blood agar and TSA plates were incubated at 37°C overnight and visually scored for bacterial colonies . Kidneys were also prepared for hematoxylin-eosin staining or immunohistochemistry . Kidneys were fixed in freshly prepared 4% paraformaldehyde soon after dissection and incubated overnight at 4°C . Further , the fixed tissues were incubated in 15% sucrose ( 4°C/24 hr ) and washed in 25% ice cold sucrose ( 4°C/24 hr ) . Tissues were then frozen in isopentane at −40°C and stored at −80°C for further use . Cryostat sections were made with a steel knife ( 10 µm ) , mounted onto poly-L-lysine-coated glass slides and stained . The kidney sections from all mouse groups were stained for dual immunohistochemistry as described [46] . Briefly , tissue sections were dried at 37°C for 15 min , washed in PBS-0 . 2% Triton X-100 ( pH 7 . 2 ) ( 2×10 min/RT ) and incubated ( 30 min/RT ) with PBS-0 . 2% Triton X-100+ 5% goat normal serum ( Dako , Denmark ) . Then the sections were incubated with primary anti-rat NIMP-R14 ( ab2557 , Abcam , Cambridge , UK; 1∶200 ) and antiserum to a synthetic peptide within the PapG adhesin ( 1∶200 ) for 2–3 hr at RT . Negative controls consisted of only normal goat serum ( 1∶200 ) . The kidney sections were washed in PBS ( 3×5 min ) and incubated ( 1 hr/RT ) with secondary goat anti-rat immunoglobulins ( 1∶200 ) , conjugated with Alexa 488 dye ( A488; 495ext/519em nm ) , and secondary goat anti-rabbit immunoglobulins ( 1∶200 ) , conjugated with Alexa 568 dye ( A568; 578ext/603em nm ) , as fluorochrome ( Invitrogen , USA ) . After washing in PBS ( 2×5 min ) , specimens were counterstained ( 3 min/RT ) with DAPI ( 0 . 05 µM ) for nuclei staining . Sections were washed again in PBS ( 10 min ) and mounted with VECTASHIELD mounting medium ( Vector Laboratories , USA ) and kept in the dark . Sections were analyzed by fluorescence microscopy ( AX60 , Olympus Opticals , Hamburg , Germany ) in the Department of Pathology , Lund University , Sweden . Urine samples were collected at 0 . 6 hr and after 1 , 2 , 3 , 4 , and 6 days and stored at −20°C . MIP-2 in urine was quantified by ELISA using the Mouse MIP-2 quantification kit ( R&D systems , Abingdon , UK ) according to the manufacturer's instructions and urine was diluted five fold in sample buffer . The ELISA plates were read at 450 nm in a Labsystems Multiskan PLUS reader ( Analytical Instruments LLC , Golden Valley , USA ) . For the microarray study , 350 , 000 A498 cells were seeded in 6-well plates and infected with CFT073 or ΔTcpC ( 108 CFU/ml ) upon confluency . Total RNA was extracted 4 hr after stimulation by acid guanidinium thiocyanate-phenol-chloroform extraction ( Trizol , Invitrogen , USA ) followed by a Qiagen RNeasy clean-up procedure . RNA was reverse-transcribed to double stranded cDNA and converted to biotin-labelled cRNA using a TargetAmp Nano-g Biotin-aRNA Labelling kit ( Epicentre Biotechnologies , Madison , USA ) . Labelled cRNAs were hybridized onto an Illumina Human WG6v3 Expression Beadchip for 16 hours at 58°C . The arrays were then washed and stained based on the Illumina Wash Protocol and then scanned using a BeadArray Scanner 500GX . The background subtracted data were pre-processed to correct negative and non significant intensities . Pre-processed data was normalized using the cross correlation [47] and genes with a fold change of 2 were identified as differentially expressed . Data was preprocessed using RMA implemented in the free software packages R and Bioconductor ( http://www . r-project . org ) . In identification of CFT073 or ΔTcpC-specific genes , fold change of ≥2 was used for the comparison CFT073 or ΔTcpC with its control and fold change of ≤1 . 5 was used for the comparison ΔTcpC or CFT073 with its control , respectively . In addition to the above fold change criteria , statistical t-test with p≤0 . 05 was further used in selection of differentially expressed genes or CFT073/ΔTcpC specific genes . To further study signaling pathways altered by TcpC , the differentially expressed genes were analyzed using the Ingenuity Pathway Analysis software with default settings ( Ingenuity Systems , Redwood City , CA ) . In parallel , cDNA was also quantified by RT-PCR using human primers IL-6 ( Hs00174131_m1 ) , IL-8 ( Hs00174103_m1 ) , NFKB1 ( Hs00231653_m1 ) , cFOS ( Hs01119266_g1 ) , IRF3 ( Hs01547283_m1 ) , STAT1 ( Hs01014002_m1 ) , IRF7 ( Hs00185375_m1 ) and TNF-α ( Hs01113624_g1 ) from Applied Biosystems . Kidney A498 cells were stimulated with CFT073 or ΔTcpC as described above . Culture supernatants were collected and cleared by centrifugation at 20 , 000×g before storage at −80°C . Cytokine profile analysis was performed in triplicate using the Procarta Human Cytokine 50-plex kit ( Panomics , Fremont , USA ) according to manufacturer's protocol . Cytokine levels were evaluated using a Luminex 100 system ( Luminex , Austin , TX , USA ) . Primary murine renal tubular epithelial cells ( MRTEC ) were harvested from murine kidneys ( C57BL/6 , Trif Lps2/Lps2 , Myd88−/− ) following microdissection and brief collagenase digestion ( 2 mg/ml of Collagenase Type II ) . Cells were grown in hormonally defined RPMI supplemented with epidermal growth factor ( Sigma , 50 pg/ml ) , insulin-transferrin-sodium selenite media supplement ( Gibco , diluted 1∶100 ) , Prostoglandin E1 ( Sigma , 1 . 25 ng/ml ) , T3 ( Sigma , 34 pg/ml ) , hydrocortisone ( Sigma , 18 ng/ml ) , 10% heat-inactivated FCS , and 0 . 5 µg/ml gentamycin at 37°C . The cells were stimulated 5 days after the primary explantation at 100% confluence and infected with CFT073 or ΔTcpC ( 108 CFU/ml ) for 4 hrs . Total extracted mRNAs were converted to cDNA using RT2 First Strand Kit ( SABioscience Corporation , Fredrick , MA , USA ) . The mouse primers used for RT-PCR GAPDH ( QT01658692 ) and MIP-2 ( QT00113253 ) were from Qiagen . Gene expression levels were calculated by the ΔCt method and normalized to house-keeping genes . cDNA was quantified by RT-PCR using a Rotor gene 2000 instrument ( Corbett Life Science , Sydney , Australia ) and normalized against GAPDH for each gene . In parallel , MIP-2 in culture supernatants was also quantified by ELISA using the Mouse MIP-2 quantification kit ( R&D systems , Abingdon , UK ) according to the manufacturer's instructions . Wt or Myd88−/− BMDM ( bone marrow derived macrophages ) were cultured in 6-well plates at a concentration of 0 . 5×106/well in DMEM supplemented with 10% FCS , 1% penicillin-streptomycin and 0 . 1% 2-mercaptoethanol . Cells were stimulated with the TLR ligands Pam3Cys , poly ( I:C ) , LPS , flagellin or CpG-ODN 1826 in the absence or presence of titrated amounts of the purified recombinant TIR domain of TcpC ( TIR-TcpC ) for 3 h . TNF was quantified in the culture supernatant using ELISA Duo sets ( R&D Systems ) as described by the manufacturer . Bacterial counts and immune responses are presented as geomean ± SEM . Fisher's exact test and two-tailed T test was used for the analysis of bacterial counts and immune response . Student's t-test was used for protein array analysis . Wilcoxon's matched pairs test was used for paired comparisons . The level of significance was set at p<0 . 05 for all tests . Gene ID number for human TLR4 is 7099 , mouse Tlr4 is 21898 , human MYD88 is 4615 , mouse Myd88 is 17874 , human TRIF is 148022 , mouse Trif is 106759 , human IRF3 is 3661 , mouse Irf3 is 54131 , human IL-1B is 3553 and mouse Il-1b is 16176 .
|
The clinical manifestations of infection range from beneficial , asymptomatic states to life threatening disease , depending on the arsenal of virulence factors carried by the bacteria and the host immune defence repertoire . Pathogenic bacteria have evolved many sophisticated ways of avoiding the host defence and especially the immune response to infection . In this study , we present a very interesting case where bacteria actively inhibit the immune response by producing a host defence like protein , TcpC , which acts by promoting bacterial survival and corrupting the tissue response to infection such that the tissues are damaged rather than protected . The importance of TcpC is demonstrated in a mouse model of urinary tract infection ( UTI ) and in isolated human and murine kidney cells . The results suggest that TcpC expressing bacteria cause death , inflammation and tissue damage in normal hosts by creating a dysfunctional innate immune response and that partial inhibition of adaptor proteins turns the normally protective defense into lethal inflammation , followed by kidney tissue damage . In human cells , TcpC was a broad innate immune inhibitor , acting via the MYD88 , TRIF and IL-1/IL-6 pathways . Our report increases the understanding of how TcpC and microbial proteins with similar targets succeed in shifting the balance in favor of the pathogen , thus promoting disease . These data are fundamentally important in showing pathways for host defense that can be fine tuned by a bacterial virulence factor in order to paradoxically promote bacterial replication thereby illustrating the host response as a generator of pathology .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases/bacterial",
"infections",
"microbiology/innate",
"immunity"
] |
2010
|
Inhibition of TIR Domain Signaling by TcpC: MyD88-Dependent and Independent Effects on Escherichia coli Virulence
|
Pseudokinases lack essential residues for kinase activity , yet are emerging as important regulators of signal transduction networks . The pseudokinase STRAD activates the LKB1 tumour suppressor by forming a heterotrimeric complex with LKB1 and the scaffolding protein MO25 . Here , we describe the structure of STRADα in complex with MO25α . The structure reveals an intricate web of interactions between STRADα and MO25α involving the αC-helix of STRADα , reminiscent of the mechanism by which CDK2 interacts with cyclin A . Surprisingly , STRADα binds ATP and displays a closed conformation and an ordered activation loop , typical of active protein kinases . Inactivity is accounted for by nonconservative substitution of almost all essential catalytic residues . We demonstrate that binding of ATP enhances the affinity of STRADα for MO25α , and conversely , binding of MO25α promotes interaction of STRADα with ATP . Mutagenesis studies reveal that association of STRADα with either ATP or MO25α is essential for LKB1 activation . We conclude that ATP and MO25α cooperate to maintain STRADα in an “active” closed conformation required for LKB1 activation . It has recently been demonstrated that a mutation in human STRADα that truncates a C-terminal region of the pseudokinase domain leads to the polyhydramnios , megalencephaly , symptomatic epilepsy ( PMSE ) syndrome . We demonstrate this mutation destabilizes STRADα and prevents association with LKB1 . In summary , our findings describe one of the first structures of a genuinely inactive pseudokinase . The ability of STRADα to activate LKB1 is dependent on a closed “active” conformation , aided by ATP and MO25α binding . Thus , the function of STRADα is mediated through an active kinase conformation rather than kinase activity . It is possible that other pseudokinases exert their function through nucleotide binding and active conformations .
Pseudokinases are classified as protein kinases that lack key catalytic residues within their kinase domain [1] , [2] . These proteins are emerging as important regulators and scaffolding components of various signal transduction networks [2] . Despite being predicted to lack intrinsic kinase activity , several “pseudokinases” such as WNK , CASK , and IRAK2 still possess the ability to phosphorylate substrates . In the case of WNK isoforms , the missing conserved catalytic Lys residue in subdomain-II is substituted by another Lys residue located in subdomain-I [3] . CASK , despite lacking the conserved Mg2+ binding Asp residue in the DFG motif of subdomain-VII , folds into an active conformation capable of binding ATP and phosphorylating substrates in the absence of Mg2+ ions [4] . Interestingly , recent studies have shown that mutations in CASK affect brain development and cause mental retardation in humans [5] . Recent data also indicate that the IRAK2 pseudokinase , despite lacking the Mg2+ binding DFG motif as well as the catalytic HRD motif , still possesses activity [6] . These results suggest that some of the other proteins in the human genome that are classified as pseudokinases may still possess catalytic activity and thus function as normal kinases . The STe-20 Related Adaptor ( STRAD ) pseudokinase forms a 1∶1∶1 heterotrimeric complex with the LKB1 tumour suppressor kinase and the scaffolding protein MO25 [7] , [8] . In humans , there are two closely related isoforms of STRAD ( STRADα and STRADβ ) and MO25 ( MO25α and MO25β ) that similarly interact with and activate LKB1 . Loss-of-function mutations in the LKB1 kinase in humans result in the inherited Peutz-Jeghers cancer syndrome [9] . Inactivating mutations in LKB1 are also increasingly being reported in sporadic cancers , in particular lung cancer [10] . LKB1 exerts its tumour-suppressing effects by phosphorylating and activating AMP-activated protein kinase ( AMPK ) as well as a number of other related kinases [11] . LKB1-mediated activation of AMPK occurs when cellular energy levels are low , and activation of AMPK inhibits cell growth and proliferation through multiple pathways , including suppressing activity of mTOR [12] , [13] . Recently , it was reported that a severe human developmental and epileptic syndrome termed polyhydramnios , megalencephaly , symptomatic epilepsy ( PMSE ) , was caused by a homozygous partial deletion in the STRADα gene ( LYK5 ) , truncating 180 C-terminal residues of the protein [14] . Individuals affected by this condition suffer from severe mental retardation , gross movement disorders , and childhood mortality [14] . How this mutation affects STRADα function and its ability to interact with LKB1 is unknown , although histological staining of neuronal tissues of PMSE patients has suggested elevated mTOR pathway activity , which could potentially result from loss of LKB1 kinase activity . Unlike the majority of kinases that require phosphorylation of their T-loop , LKB1 is activated through direct interaction with STRADα/β isoforms [7] , [8] . The kinase domain of LKB1 binds to the pseudokinase domain of STRAD [7] . At least 12 point mutations located in the LKB1 kinase domain that prevent LKB1 from interacting with STRAD isoforms have been identified in human cancers [15] . Activation of LKB1 and interaction with STRAD isoforms is markedly enhanced in the presence of MO25α/β isoforms , indicating that MO25 stabilizes the interaction between STRAD and LKB1 . The C-terminal Trp-Glu-Phe residues ( WEF motif ) of STRADα bind to MO25α , and mutations of these residues abolish this interaction [8] . Structural analysis of MO25α revealed a helical repeat , horseshoe-shaped protein that interacts with the WEF motif of STRADα through a hydrophobic pocket located on its convex C-terminal surface [16] . In contrast , proteins that are distantly structurally similar to MO25α , such as the Armadillo repeat proteins PUM1 , β-catenin , and importin-α , interact with their binding partners through their concave surface [17]–[19] . Many of the surface-exposed residues on the MO25α concave surface are conserved between species , suggesting that these may mediate interactions with ( an ) unknown regulator ( s ) [16] . Although STRADα mutants lacking the C-terminal WEF motif are unable to interact with MO25α alone , they can still form a heterotrimeric complex with LKB1 and MO25α , demonstrating that STRADα possesses additional interactions with LKB1 and/or MO25α , separate from the WEF motif [15] . All studies undertaken to date suggest that STRADα expressed in bacteria is incapable of autophosphorylating or phosphorylating other substrates tested ( MBP , histone 2A , or LKB1 ) when assays were undertaken in the presence of Mg2+ ions [7] , [15] ( J . Boudeau , unpublished data ) . Despite lacking detectable kinase activity , STRADα is still capable of interacting with ATP as well as ADP in a magnesium-independent manner [15] . Mutations that abolish ATP binding do not affect the ability of STRADα to activate LKB1 in the presence of MO25α . Thus , the role of ATP-binding to STRAD is unclear . Here , we report the structure of STRADα as part of the STRADα/MO25α heterodimer . The data show that despite being inactive , STRADα folds into an ATP-bound , closed conformation with an ordered activation loop similar to that of fully active protein kinases . Our data establish that STRADα is indeed deficient in intrinsic catalytic activity because it lacks most essential catalytic residues . Moreover , we observe that STRADα does not only interact with MO25α through its WEF motif as previously envisaged , but forms an extensive network of interactions with the highly conserved concave surface of MO25α . Binding studies and mutagenesis data show that the closed/“active” conformation that STRADα assumes is maintained through cooperative binding of ATP and MO25α . STRADα mutants incapable of interacting with ATP and MO25α are unable to activate LKB1 , despite interacting with it . We conclude that the ability of STRADα to activate LKB1 is dependent on an active conformation rather than catalytic phosphoryltransferase activity . Our results also indicate that the human mutation that causes PMSE syndrome destabilizes STRADα and prevents it from binding to , and activating LKB1 .
STRADα comprises a pseudokinase domain ( residues 58–401 ) , two nuclear export sequences ( residues 21–29 and 417–426 ) [20] , and a C-terminal WEF motif ( residues 429–431 ) previously shown to interact with MO25α [8] , [16] . We focused on the interaction between the STRADα pseudokinase domain ( residues 59–431 ) and full-length MO25α ( residues 1–341 ) . These proteins were coexpressed in Escherichia coli and the STRADα/MO25α complex eluted as a heterodimer of the expected size from a gel filtration column , yielding approximately 60 mg of the complex from 4 l of culture ( Figure S1 ) . Initial crystals of the STRADα/MO25α complex in space group P212121 diffracted only to 4 . 8 Å resolution ( Table 1 ) . With the help of chemical lysine methylation [21] , diffraction of these crystals ( retaining the same space group and unit cell dimensions ) improved to 2 . 35 Å ( Table 1 ) . The structures of both methylated and unmethylated crystals were solved by molecular replacement and revealed the same packing/intermolecular interactions . The high-resolution , methylated form of the complex was refined to a final model with good statistics ( Rfree/Rwork of 0 . 254/0 . 206; Table 1 ) . The structure of STRADα exhibits the classical bilobal protein kinase fold , with the N-terminal lobe ( residues 59–152 ) organized around a central β-sheet , and a C-terminal lobe ( residues 153–401 ) that is largely α-helical ( Figure 1A ) . A well-resolved molecule of ATP was observed in the cleft between the small and large lobes of the pseudokinase ( Figure 1A and 1B ) . The ATP molecule displays the canonical binding mode and retains a similar conformation to that of ATP molecules bound to active kinases ( root mean square deviation [RMSD] = 0 . 9 Å on all atoms compared to ATP bound to PKA [22] ) . Sequence comparison reveals that STRADα lacks numerous essential catalytic residues found in active protein kinases , namely a conserved Gly residue in the glycine-rich loop ( subdomain-I ) , the Lys residue of the VAIK motif ( subdomain-II ) , the catalytic Asp residue of the HRD motif ( subdomain-VIb ) , a conserved Asn residue ( subdomain-VIb ) , as well as the entire DFG motif in subdomain-VII ( Figure 1B and 1C ) . Despite missing these key residues , STRADα adopts a similar overall conformation to that of TAO2 ( sharing 25% sequence identity and 37% sequence similarity ) , an active protein kinase of known structure [23] ( RMSD = 1 . 4 Å on 197 Cα atoms ) . Comparison of the STRADα and TAO2 structures reveals that a number of substitutions of key catalytic residues are found in STRADα . Met83 replaces one of the conserved Gly residues in the glycine-rich loop , Arg100 substitutes the catalytic Lys residue in the VAIK motif , Ser195 replaces the Asp residue in the HRD motif , His200 substitutes for the conserved Asn in subdomain-VIb , and the entire DFG motif is replaced by GLR ( residues 213–215 ) . In active protein kinases , the DFG motif plays a pivotal role in coordinating two Mg2+ ions: one that orients the γ-phosphate into the position required for phosphoryl transfer and the other that controls ATP conformation by interacting with the β/γ phosphates . Consistent with the lack of the DFG motif in STRADα , no Mg2+ ions were observed in the STRADα-ATP complex , despite 1 mM MgCl2 being present in the crystallization mother liquor . However , despite the absence of Mg2+ ions , the positioning of the β/γ phosphates in STRADα was similar to that of active TAO2 kinase complexed to MgATP ( Figure 1B ) . The β-phosphate is tethered through interactions with Arg215 from the GLR ( DFG ) motif , and His200 ( subdomain-VIb ) , basic residues that may substitute for one of the positively charged Mg2+ ions ( Figure 1B ) . The second Mg2+ ion and its coordinating residues are also missing; instead , the γ-phosphate only interacts with a conserved lysine ( Lys197 ) in the catalytic loop . Thus , STRADα appears to have evolved a novel , Mg2+-independent mechanism to bind the phosphate groups of ATP . The presence of the two hydrogen bonds between N1 and N6 atoms of the ATP adenine ring and the protein backbone , observed in all active protein kinase structures , further illustrates the conservation of the ATP binding pocket . Thus , the STRADα structure explains previous observations that STRADα can bind ATP in the absence of Mg2+ , and its similar affinity for ADP and ATP [15] . Although the activation loop of STRADα ( residues 212–245 ) is not phosphorylated , it is well ordered , a feature normally observed only in structures of activated protein kinases that are phosphorylated on their activation loop ( Figure 1A ) . Remarkably , Asp232 in the activation loop occupies a position similar to the activating phosphorylated residue found in active kinases , e . g . , ( phospho ) Ser181 in TAO2 ( Figure 1B ) . Asp232 appears to play the same structural role as the activating phosphate group , coordinating the conserved arginine from the catalytic HRD motif ( Arg194 in the STRADα HRS motif ) ( Figure 1B and 1C ) . Further evidence that STRADα adopts the canonical active conformation stems from the presence of a short antiparallel β-sheet between the β6 and β9 strands , which is a characteristic feature of the active state of kinases [24] . Furthermore , the STRADα αC-helix is rotated into the “closed” conformation found in active kinases [25] , [26] , with the conserved ion pair between the Glu118 on the αC-helix and Arg100 in subdomain-II formed via two water molecules ( Figure 1B ) . Despite STRADα binding ATP in the correct orientation for activity and folding into an active conformation , STRADα ( residues 59–431 ) expressed in E . coli did not autophosphorylate or phosphorylate myelin basic protein ( Figure S2 ) . We have attempted to detect activity in the presence and absence of MO25α and/or 10 mM MgCl2 . We have also generated mutations converting all the missing catalytic residues on the STRADα pseudokinase discussed above to the equivalent residues found in the active kinase TAO2 ( Figure S2 ) . However , none of these mutants showed autophosphorylation or phosphorylated myelin basic protein in the presence or absence of Mg2+ ions and/or MO25α ( Figure S2 ) . We also tested whether STRADα possessed ATPase activity , employing a highly sensitive ATPase assay kit ( Innova Biosciences ) , but no activity was observed ( E . Zeqiraj , unpublished data ) . Nevertheless , it is impossible to categorically rule out that STRADα will not , highly specifically , phosphorylate an as-yet unidentified substrate . The asymmetric unit of the STRADα/MO25α complex crystals contains one molecule of MO25α , with a conformation similar to the previously published MO25α/WEF peptide complex structure [16] ( RMSD = 0 . 6 Å on 292 Cα atoms ) , and one molecule of STRADα . The position and conformation of the WEF motif is similar to that in the previously described MO25α/WEF complex [16] ( RMSD = 0 . 3 Å on 35 atoms , Figure S3A ) . Due to tight crystal contacts ( total buried surface on MO25α by STRADα and its symmetry mates = 2 , 833 Å2 ) , it was not immediately apparent which contacts represented biologically relevant interactions and which were crystallographic packing artefacts . Whereas clear electron density is present for the last six amino acids of STRADα ( residues 426–431 , including the WEF motif that interacts with MO25α , Figure S3A ) , residues 402–425 of STRADα were not visible in the electron density maps , and it was thus not possible to directly identify the appropriate symmetry mates of STRADα and MO25α that make up the biologically relevant binary complex . Analysis of the crystal contacts between symmetry-related molecules suggested that there were four possible ways in which STRADα could interact with MO25α ( Figure 1D ) . We studied all four possible STRADα/MO25α complexes and ranked these in terms of total buried surface area , a possible method for distinguishing crystallographic from biological contacts [27] . Discounting the WEF motif interaction ( 800 Å2 buried surface area ) , identical in all four possible complexes , the buried surface area in each of the possible complexes is 1 , 550 Å2 , 225 Å2 , 58 Å2 , and 200 Å2 for complexes A , B , C , and D , respectively ( Figure 1D ) . In addition , the distances between the last well-defined residue of the STRADα C-terminal lobe and the first well-defined residue of the WEF motif at the extreme C-terminus of STRADα were measured for the four possible complexes . This yielded direct distances of 52 , 83 , 65 , and 55 Å for complexes A , B , C , and D , respectively ( Figure 1D ) . Taken together , it appears that complex A is the most likely biological interaction , since STRADα binds to the ( highly conserved ) concave surface of MO25α and has the largest buried surface area , while also possessing the shortest distance from the C-terminal lobe to the WEF motif ( Figure 1D ) . Similarly , analysis of the possible complexes with PISA [28] yields the highest ( 1 . 0 ) complexation significance score ( CSS ) for complex A , while predicting that complexes B , C , and D will not be stable in solution . The 6-His purification tag that extends from the N-terminus of STRADα ( 450 Å2 buried surface area in complex A ) forms additional contacts between MO25α and STRADα . SPR studies demonstrate that His-tagged STRADα binds MO25α in vitro with the same affinity as STRADα lacking the His tag ( Figure S4 ) . Furthermore , MO25α residues 2–5 ( Pro-Phe-Pro-Phe , termed the PFPF motif here ) make hydrophobic contacts in a pocket adjacent to the STRADα ATP binding pocket on a symmetry-related copy of STRADα ( Figure S3B ) . This is unlikely to constitute a physiological STRADα/MO25α interaction , as deleting this motif did not impair the in vivo interaction of MO25α with either STRADα alone or a complex of STRADα and LKB1 ( Figure S3C and S3D ) . Moreover , we were unable to affinity purify overexpressed STRADα or LKB1 from a cell extract employing a PFPF motif containing biotinylated peptide ( Figure S3E ) . A complex of LKB1/STRADα/MO25α ( ΔPFPF ) still activated the heterotrimeric AMPK complex expressed in E . coli with similar efficiency as wild-type LKB1/STRADα/MO25α ( Figure S3F ) . Nevertheless , it is possible that the PFPF docking site on STRADα does play a role in enabling STRADα to interact with other regulators or substrates of the LKB1 complex . Intriguingly , a similar crystallographic interaction can be observed in the structure of the mammalian AMPK heterotrimeric complex [29] . In this case , a similar hydrophobic N-terminal motif “MYAF” from the β2 domain interacts with the kinase domain from the neighbouring molecule in the crystal lattice , albeit not near the phospho-nucleotide binding site . MO25α is composed of seven structurally similar α-helical repeats ( named R0–R6 ) that form a horseshoe-shaped molecule with a concave and a convex surface [16] . MO25α helical repeats R1–R6 consist of three α-helices ( H1–H3 ) each , whereas repeat R0 consists of only two helices [16] . Helices H3 from repeat R1–R5 are arranged in an almost parallel fashion and make up the concave surface of MO25α ( Figure 1E ) . Other helical repeat adaptor proteins , such as PUM1 , β-catenin , and importin-α , make use of a similar concave surface to interact with macromolecular partners [17]–[19] . Strikingly , the crystal structure of the STRADα/MO25α complex reveals that , in addition to the interaction through the WEF motif , a major additional binding interface involves the STRADα N-terminal kinase lobe and the MO25α concave surface ( Figures 1E and 2 ) . Part of the interaction surface on STRADα is N-terminal to the αC-helix and comprises the loop between the αB/αC helices ( residues 104–109 ) , termed the “αB site” here ( Figures 1E and 2A ) . This region forms an extensive hydrogen-bonding network centred on Arg227 from the R5-H3 of MO25α ( Figure 1E ) , burying a total of 245 Å2 surface area . Residues Tyr223 , Arg227 , Lys231 , and Asn269 of MO25α engage the side chains of residues Glu105 and Asn109 of STRADα , whereas Leu104 , Ala106 , Cys107 , and Ser108 contribute to the interaction via their backbone atoms . The αC-helix of STRADα runs along the concave surface of MO25α facing the H3 helixes of the MO25α repeats R4 , R3 , and R2 ( Figures 1E and 2A; termed the “αC site” here ) . Tethered by hydrophobic and hydrogen-bonding interactions ( Figure 1E ) , the αC-helix forms the major interaction surface , contributing a total of 405 Å2 buried surface area on the MO25α concave surface . C-terminal to the αC helix , a second hydrogen-bonding network with comparable buried surface area ( 270 Å2 ) to the αB site is present , and involves residues Leu124 , Asn126 , and Tyr185 from the STRADα helix αE ( Figures 1E and 2A; termed the “αE site” here ) . This region interacts with Glu93 , Lys96 , and Phe92 from the R1-H3 helix of MO25α ( Figure 1E ) . Together , the αB site and the αE site appear to act as anchor regions , positioning the αC-helix to run along the H3 helices of R1–R5 of MO25α . Additional interactions are found between Phe178 of MO25α , forming hydrophobic stacking interactions with residues from the N-terminal β4 and β5 strands of STRADα ( termed the “β4/β5 site” here; Figures 1E and 2A ) . STRADα and STRADβ also possess an insertion of ten residues ( 221–229 ) in the activation loop that is not observed in TAO2 or other STE20 kinases ( Figure 1A ) . Within this insertion , His223 , Gly224 , and Arg226 show weak interactions with the R0 and R1 helical repeat of MO25α ( termed the “activation loop site” here; Figure 2A ) . This interaction perhaps explains why the STRADα activation loop is ordered . All of the key interacting interface residues are highly conserved between species of STRADα and MO25α ( Figures 2 and S5 ) . The structure of the STRADα/MO25α complex shows that , in addition to the WEF binding pocket on the convex surface of MO25α , a major network of interactions between STRADα and the concave surface of MO25α is observed over the αB , αC , αE , β4/β5 , and activation loop sites . To test the importance of these additional interactions , we investigated how mutations of residues located on the MO25α concave surface affected interaction with STRADα . We mutated residues in MO25α in the novel αB , αC , αE , β4/β5 , and activation loop binding sites as well as the previously characterised WEF pocket ( Figure 3A ) . As reported previously , mutation of Met260 in the WEF pocket of MO25α abolishes its ability to interact with STRADα in HEK293 cells [16] . However , we also observed that mutations in the two anchor regions ( Phe92 , Glu93 , and Lys96 from the αE site and Tyr223 and Arg227 from the αB site ) abolished MO25α binding to STRADα ( Figure 3A ) . Similarly , mutating Phe178 in the β4/β5 site , Ile145 and Ser182 in the αC site , or Arg107 in the activation loop site markedly disrupted the MO25α-STRADα interaction . Mutations of Leu141 , Lys231 , and Asn269 in the αC site did not significantly affect binding ( Figure 3A ) . Mutation of the reciprocal interacting residues on STRADα , including Glu105 , Asn109 , Asn126 , Ile138 , and Tyr185 , also abolished or markedly reduced binding to MO25α ( Figure 3B ) . These results confirm the importance of the network of interactions between the concave surface of MO25α and STRADα in enabling the stable association between these two proteins , at least in the absence of LKB1 . Previous work has shown that MO25α mutants in which the WEF pocket was disrupted , and that were no longer able to form a complex with STRADα , were still capable of forming a heterotrimeric complex with LKB1 and STRADα [8] , [15] . Similarly , MO25α mutants in which key STRADα binding residues located within the concave surface were mutated are still capable of interacting with the LKB1/STRADα complex ( Figure 3C ) . Even double MO25α mutants in which both the WEF pocket and the αB , αE , or β4/β5 sites were disrupted were capable of associating with the LKB1/STRADα complex ( Figure 3C ) . Moreover , the specific activity of LKB1/STRADα complexes associated with these MO25α mutants was either normal or only moderately reduced ( Figure 3C ) . This suggests the presence of additional interactions between MO25α and LKB1 in the presence of STRADα . Earlier studies revealed that mutation of a conserved Arg240 residue located on the concave surface of MO25α reduced interaction with LKB1 complexed to STRADα lacking the WEF motif [15] . Arg240 might be involved in interaction with LKB1 , as this residue is located on the concave surface of MO25α , distant from STRADα ( Figure 2 ) . To further investigate the role of Arg240 in enabling MO25α to associate with LKB1/STRADα , we mutated Arg240 alone or in combination with residues in either the WEF pocket ( Met260 ) or the αB STRADα binding sites ( Arg227 ) . We found that mutation of Arg240 alone does not prevent MO25α from interacting with LKB1/STRADα ( Figure 3D ) . However , a double MO25α mutant lacking Arg240 and a key concave surface-binding site in the αB site ( Arg227 ) , markedly impaired binding to LKB1/STRADα ( Figure 3D ) . A triple mutant of MO25α lacking Arg240 , Arg227 , and the WEF pocket site failed to associate with LKB1/STRADα and stimulate LKB1 activity ( Figure 3D ) . These observations indicate that MO25α possesses three sites with which it can interact with the LKB1/STRADα complex ( Figure 2 ) , namely two STRADα binding regions ( extensive concave MO25α surface and WEF pocket ) as well as a putative LKB1 binding site ( Arg240 ) . Inspection of the STRADα/MO25α complex reveals an unexpected resemblance to the interaction between activated cyclin-dependent kinase 2 ( CDK2 ) and its activating regulatory subunit cyclin A ( Figure 4A and 4B ) [30] . Although MO25α/β isoforms are not related to cyclins at the primary sequence level , both proteins consist of multiple α-helical repeats . Crystal structures of CDK2/cyclin A complex have revealed cyclin A binds to the so-called “PSTAIRE ( αC ) helix” of CDK2 kinase as well as the loop immediately preceding this helix [30] . Comparisons between free CDK2 and CDK2/cyclin A complex structures have shown that the cyclin molecule orients a conserved glutamate residue ( Glu51 ) from the αC-helix of the protein kinase to allow formation of an ion pair with a lysine residue ( Lys33 ) from the conserved VAIK motif [30] , which keeps the CDK2 kinase in a closed conformation ( Figure 4B ) . Similarly , the position of MO25α in the STRADα/MO25α complex is centred on helix αC and the loop preceding this helix ( αB region; Figure 1E ) . The interaction between Glu118 from the αC-helix and Arg100 from the VAIK ( VTVR in STRADα ) motif ( analogous to the Glu51-Lys33 interaction in CDK2 ) is maintained , albeit via two water molecules ( Figure 1B ) . Another example in which this type of interaction is involved in regulating the activity of protein kinases is the ligand-induced dimerisation of the EGFR family of tyrosine kinases ( Figure 4C ) . Although this type of dimer has not been observed in solution , crystal structures and biochemical data demonstrate the importance of dimer formation that involves the intermolecular interaction of the EGFR αC-helix on one monomer and the C-lobe on the other monomer ( Figure 4C ) [31] . A comparison between the structure of active , dimeric EGFR kinase with the monomeric form reveals the role of dimerisation for keeping the EGFR kinase in the closed and active conformation . Similarly , the structure of STRADα in complex with MO25α resembles the closed conformation of both CDK2 and EGFR kinase , with its activation loop and αC-helix positioned in an orientation that is typical of active protein kinases ( Figure 4D and 4E ) . Such regulatory mechanism may also explain why some members of the EGFR family of kinases that lack kinase activity and are classified as pseudokinases ( Her3 ) are still able to exert their function [31] , despite their “inactivatory” substitutions , similar to what has been observed for STRADα ( Figure 1C ) . The interactions in the EGFR homodimer , the CDK2/cyclin A heterodimer , and the STRADα/MO25α complex are similar only in general topological terms . However , it appears that the mechanism of protein kinase interaction via helix αC with their activity modulators is wider than previously thought , and not exclusive to the CDK family of kinases . Indeed , there are many examples of how protein kinases are stabilised in an active conformation via helix αC . These include members of the MAP kinase family [32] , the AGC family of kinases ( [33]–[35] , and several tyrosine kinases ( [36] ) . Although in these examples the αC-helix is stabilised by flanking N- or C-terminal sequences/domains present in the same polypeptide chain , the mechanisms of allosteric activation are similar . Although MO25α appears to induce a STRADα active conformation similar to CDK2/cyclin A , the effect of this “active conformation” cannot be measured through ATPase/kinase activity due to STRADα being a pseudokinase . Instead , we investigated how affinity of ATP for STRADα was modulated by its interaction with MO25α . We used the fluorescent ATP analogue 2′ , 3′-O-2 , 4 , 6-trinitrophenyl-ATP ( TNP-ATP ) , whose fluorescence emission is enhanced upon its titration with ATP-binding proteins/enzymes [37] , a feature that has previously been exploited to measure equilibrium binding constants of kinases for ATP [4] . Using this approach , the Kd of STRADα for TNP-ATP in the absence of MO25α was determined to be 1 . 1 µM ( Figure 5A , 5B , and 5E ) . Kd values of STRADα for ATP and ADP were also assessed by their ability to displace bound TNP-ATP and found to be 2–3 µM ( Figure 5C , 5D , and 5E ) . Strikingly , addition of an equimolar amount of MO25α to STRADα enhanced binding of TNP-ATP by an order of magnitude ( Figure 5A , 5B , and 5E ) and TNP-ATP displacement by two orders of magnitude ( Figure 5C , 5D , and 5E ) , indicating significantly stronger affinity compared to the interaction of ATP as a substrate to active kinases . In contrast , the binding of STRADα to TNP-ATP was not enhanced by addition of the MO25α ( R227A/M260A ) mutant that is unable to bind STRADα ( Figure 5A and 5B ) . The lack of a Mg2+ binding motif on STRADα suggests that Mg2+ should not contribute to the STRADα-ATP interaction . Indeed , Mg2+ did not affect binding of STRADα to TNP-ATP or displacement of TNP-ATP by ATP or ADP ( Figures 5 and S6 ) . This is in contrast with the CASK “pseudokinase , ” where Mg2+ reportedly inhibits ATP binding and hence kinase activity [4] . It should be noted that although STRADα does not appear to require Mg2+ ions to bind ATP , most cellular ATP is complexed to Mg2+ ions . Although there is no space for Mg2+ to bind in the canonical protein kinase mode through the DFG motif , Mg2+ ions could reside in the solvent-exposed region of the phosphate moiety , replacing one of the ordered water molecules . Alternatively , it is possible that conformational changes in the structure could accommodate Mg2+ without affecting the ability of STRADα to bind MO25α ( see below ) . As mentioned previously , the canonical Mg2+ coordinating residues appear to have been substituted through evolution with positively charged residues ( Arg240 and H200 ) , thus making redundant the role of Mg2+ ions . To further investigate the functional consequences of ATP binding to STRADα , we employed quantitative SPR measurements to evaluate how ATP influenced affinity of STRADα for MO25α ( Figures 6 and S7 ) . In the absence of ATP , the binding of STRADα for MO25α was fitted to a single-site binding equation ( Figure 6A and 6E ) . From measuring the rate constants for association and dissociation ( Figure S7 and Table S1 ) , the dissociation constant Kd was calculated as 3 . 8 µM ( Figure 6A and 6E ) . However , in the presence of ATP , binding could be fitted to a two-site binding equation ( Hill slope of 0 . 4 , Figure 6A and 6E ) . The second binding constant ( Kd2 ) was measured as 12 nM , over two orders of magnitude higher than Kd1 calculated as 2 . 5 µM ( Figure 6A and 6E ) . MgATP enhanced binding of STRADα to MO25α , to a similar extent as ATP ( Figure 6A ) . These results indicate that binding of ATP to STRADα leads to a high-affinity MO25α interaction site being exposed . Mutation of Met260 in the WEF binding pocket of MO25α did not significantly affect binding of MO25α to STRADα , nor did it influence the effect of ATP at enhancing interaction ( Figure 6B and 6E ) . It should be noted that this observation contrasts with the data obtained from coexpression studies in 293 cells ( Figure 3A ) and previous studies [16] , in which mutation of Met260 inhibits MO25α binding to STRADα , suggesting that the WEF pocket is required for cellular complex assembly of MO25α and STRADα . Mutation of Arg227 , in the newly identified concave site of MO25α , which interacts with the αB site of STRADα , virtually abolished binding of STRADα observed by SPR in the absence of ATP . In the presence of ATP or MgATP , no two-site binding of MO25α ( R227A ) to STRADα was detected , displaying only low micromolar binding with a single site ( Figure 6C and 6E ) . A double MO25α ( R227A/M260A ) mutant failed to interact with STRADα even in the presence of ATP ( Figure 6D and 6E ) . These results indicate that the key STRADα high-affinity binding site on MO25α lies on the concave surface and is only recognized by STRADα in the presence of ATP . Together with the finding that MO25α also enhances affinity of STRADα for ATP ( Figure 5 ) , this suggests that the interaction of ATP and MO25α to STRADα is cooperative . A similar synergistic mechanism is observed for the PKA catalytic subunit where a nucleotide analog was shown to stabilise a complex with the PKI inhibitory peptide [38] . However , in the case of PKA/PKI interaction the γ-phosphate cannot be transferred because there is no acceptor , whereas in case of STRAD , it cannot be transferred because of the lack of a base catalyst . Having established that ATP increases the affinity of STRADα-MO25α interaction , we next explored whether ATP binding to STRADα also affects assembly and activity of the LKB1 heterotrimeric complex . Using the STRADα-ATP structure , a number of STRADα mutants were designed to disrupt binding of the adenine or phosphate moieties of ATP ( Figure 7A ) . Four of these were indeed unable to interact with TNP-ATP in the presence or absence of MO25α ( Figure 7B ) . Interestingly , these mutants also affected association with LKB1 when coexpressed in 293 cells ( Figure 7C ) , suggesting that binding of ATP to STRADα , in the absence of MO25α , enhances the ability of STRADα to interact with LKB1 . However , these mutants were capable of forming complexes with LKB1 when coexpressed with LKB1 and MO25α ( Figure 7D ) , that retained catalytic activity as measured by activation of AMPK ( Figure 7D ) . It is possible that binding of MO25α to these STRADα mutants compensates for their inability to bind ATP , by inducing a closed “active-like” conformation of STRADα , capable of binding and activating LKB1 . To explore this idea , we generated mutants of STRADα incapable of binding to both ATP and MO25α . Strikingly , we found that these combined STRADα mutants lost their ability to activate LKB1 , despite still being capable of forming a heterotrimeric complex ( Figure 7E ) . Taken together , these observations suggest that the closed “active-like” conformation of STRADα is maintained through binding to ATP and/or MO25α , and is required for activation of LKB1 . Mutations that prevent STRADα from binding to ATP or MO25α do not affect activation of LKB1 ( Figures 3 and 7 ) , suggesting that ATP binding to STRADα can compensate for loss of MO25α interaction and vice versa . However , loss of both ATP and MO25α binding prevents STRADα from activating LKB1 . Such mutations may leave STRADα in the open “inactive-like” conformation incapable of activating LKB1 . We have tried unsuccessfully to crystallise STRADα in the absence of MO25α in order to demonstrate this . Binding of ATP to several kinases , including the EGF receptor tyrosine kinase , promotes the closed , active conformation of these enzymes . Moreover , as discussed above , binding of cyclin to CDK2 is reminiscent of the interaction of STRADα with MO25α , and interaction of cyclin A is well known to promote the closed active conformation of CDK2 [30] . The PMSE-causing mutation in humans results in a STRADα truncation at residue 251 , thus removing the last 180 amino acids [14] . Inspection of the STRADα structure reveals that this mutation would delete almost half of the C-terminal lobe of the pseudokinase domain , beginning with structurally vital components such as helix αF ( Figure 8A ) . This could destabilize the STRADα protein , as helix αF forms numerous hydrophobic interactions within the C-lobe of the pseudokinase domain , which would become solvent exposed in the PMSE mutant . We attempted to express the PMSE-STRADα ( residues 1–251 ) mutant in 293 cells and found that it was expressed at significantly lower levels than full-length STRADα ( Figure 8B ) , consistent with this fragment being unstable . Moreover , STRADα ( 1–251 ) failed to interact with or activate LKB1 ( Figure 8B ) . These results confirm that the STRADα mutation found in PMSE patients represents a loss-of-function mutation that would be unable to stimulate the LKB1 pathway . This could account for the elevated mTOR pathway activity that was observed in neuronal cells derived from PMSE patients [14] . We have described the first structure of the STRADα pseudokinase and its interaction with MO25α , a heterodimeric interaction within the heterotrimer LKB1 tumour suppressor complex . A key discovery is the identification of an unexpected extensive interaction between STRADα and the concave surface of MO25α , previously proposed to harbour a ligand binding site [16] . Armadillo repeat proteins that are structurally related to MO25α , such as PUM1 [19] , β-catenin [17] , and importin-α [18] , also bind their macromolecular partners along their concave surface . In general topological terms , the STRADα/MO25α complex resembles the interaction between CDK2 and cyclin A , and the EGFR/EGFR dimer , and provides another example of protein kinase regulatory mechanism via helix αC . Our data show that , despite lacking most essential catalytic residues , STRADα has maintained its ability to adopt a closed active-like conformation , which binds ATP and possesses an ordered activation loop similar to active protein kinases . This closed conformation is stabilized through binding of ATP and/or MO25α . Moreover , binding of MO25α to STRADα markedly enhances affinity for ATP , and binding of ATP to STRADα stimulates interaction with MO25α . Our findings support a model in which binding of either MO25α or ATP is sufficient to enable STRADα to activate LKB1 . Consistent with this , mutant forms of STRADα that are incapable of binding both ATP and MO25α can no longer activate LKB1 , whereas mutant forms of STRADα that retain the ability to bind either ATP or MO25α still activate LKB1 . Thus , the closed active-like conformation , rather than catalytic phosphoryl transfer activity , is likely to be the key to the mechanism by which STRADα activates the LKB1 tumour suppressor . A model of how STRADα/MO25α might interact and activate LKB1 based on known mutagenesis and structural data is presented in Figure 9 . Future work may establish other examples of pseudokinases that , like STRADα , regulate signal transduction networks through their conformational state alone . Very recent reports have described the structures of VRK3 [39] and ROP2 [40] pseudokinases , both incapable of binding ATP . Both studies support the notion put forward in this paper that pseudokinases may function by means of conformational state rather than catalytic activity , although in an ATP-independent manner .
Restriction enzyme digests , DNA ligations , and other recombinant DNA procedures were performed using standard protocols . All mutagenesis were performed using the QuickChange site-directed mutagenesis method ( Stratagene ) with the KOD polymerase ( Novagen ) . DNA constructs used for transfection were purified from E . coli DH5α using Qiagen Plasmid kits according to the manufacturer's protocol . All DNA constructs were verified by DNA sequencing , which was performed by the Sequencing Service , College of Life Sciences , University of Dundee , United Kingdom , using DYEnamic ET terminator chemistry ( Amersham Biosciences ) on Applied Biosystems automated DNA sequencers . Lysis buffer used for HEK 293 cells was 50 mM Tris-HCl ( pH 7 . 5 ) , 1 mM EGTA , 1 mM EDTA , 1% ( w/v ) Nonidet P40 ( substitute ) , 1 mM sodium orthovanadate , 50 mM sodium fluoride , 5 mM sodium pyrophosphate , 0 . 27 M sucrose , 0 . 1% ( v/v ) 2-mercaptoethanol , 1 mM benzamidine , and 0 . 1 mM PMSF . Buffer A was 50 mM Tris-HCl ( pH 7 . 5 ) , 0 . 1 mM EGTA , and 0 . 1% ( v/v ) 2-mercaptoethanol . SDS sample buffer contained 50 mM Tris-HCl ( pH 6 . 8 ) , 2% ( w/v ) SDS , 10% ( v/v ) glycerol , 0 . 005% ( w/v ) bromophenol blue , and 1% ( v/v ) 2-mercaptoethanol . TBS-T buffer was Tris-HCl ( pH 7 . 5 ) , 0 . 15 M NaCl , and 0 . 5% ( v/v ) Tween . All protein concentrations were determined using the Bradford reagent ( Bio-Rad ) and by measuring the absorbance at 595 nm , unless indicated otherwise . A bicistronic expression system was used to coexpress and purify the STRADα/MO25α complex in E . coli . Expression vectors were kindly donated by Dr . Roger Williams ( University of Cambridge , United Kingdom ) . The cloning procedure was followed as described in [41] . Briefly , both STRADα and MO25α genes were subcloned as separate cassettes from the pOPT single vectors into a pOPCH polycistronic vector . Full-length MO25α ( residues 1–341 ) was subcloned from a pOPT ( no tag ) vector as an NdeI/BamH1 insert . STRADα ( residues 59–431 ) with an N-terminal 6-His tag followed by a Tobacco Etch Virus ( TEV ) protease site ( sequence MAHHHHHHMENLYFQG ) was subcloned from a POPTH vector as a BspE1/Mlu1 insert . For more information on the expression and purification of STRADα for activity assays , see Text S1 . N-terminally 6-His-tagged STRADα was coexpressed with untagged full-length MO25α in E . coli BL21 ( DE3 ) pLysS cells . Cells were grown in Luria Bertani medium to A600 = 0 . 7 at 37°C , before protein expression was induced by the addition of 250 µM isopropyl-β-d-thiogalactopyranoside ( IPTG ) and incubated for a further 16 h at 26°C . Cells were harvested by centrifugation for 30 min at 3 , 500g and resuspended in ice-cold lysis buffer ( 50 mM Tris-HCl [pH 7 . 8] , 50 mM NaCl , 10% glycerol , 20 mM imidazole , 1 mM benzamidine , 0 . 2 mM EGTA , 0 . 2 mM EDTA , 0 . 1 mM PMSF , 0 . 075% ( v/v ) β-mercaptoethanol , 0 . 5 mg/ml lysozyme , and 0 . 3 mg/ml DNAse-I . Cells were lysed using a French Press cell disrupter ( 18 , 000 psi ) , and the lysate was cleared by centrifugation at 26 , 000g for 30 min . The supernatant was then passed through a 0 . 22-µm filter before loading onto a 5-ml HiTrap IMAC HP column ( GE Healthcare ) previously charged with Ni2+ . The column was then washed with ten volumes of wash buffer ( lysis buffer without lysozyme , DNAse-I , and PMSF ) , and the STRADα/MO25α complex was eluted by applying a gradient of 20–300 mM imidazole in wash buffer . The sample was then concentrated to 3 ml and loaded onto a Superdex 75 26/60 gel filtration column , pre-equilibrated in 25 mM Tris ( pH 7 . 8 ) and 1 mM DTT . For the methylated protein complex , the sample was dialyzed into 25 mM Tris-HCl ( pH 7 . 5 ) , 50 mM NaCl , 10% glycerol , 1 mM benzamidine , and 0 . 075% ( v/v ) β-mercaptoethanol after imidazole elution , and subjected to lysine methylation using formaldehyde and dimethylamine-borane complex , as described elsewhere [21] . The methylated STRADα/MO25α complex was then passed through a desalting column prior to loading onto a gel filtration column as explained above . The binary complex eluted as a single peak , and its purity was assessed by SDS-PAGE . The STRADα/MO25α complex was concentrated to 7 . 5 mg/ml , followed by addition of ATP to a final concentration of 10 mM and MgCl2 ( final concentration of 1 mM ) . The sitting drop vapour diffusion method was used to grow crystals by mixing 1 µl of protein solution , 1 µl of mother liquor . For the unmethylated complex , the optimised mother liquor consisted of 20 mM Li2SO4 , 50 mM sodium citrate ( pH 5 . 6 ) , 6% ( v/v ) PEG4000 . For the methylated complex , the mother liquor was composed of 0 . 1 M MES ( pH 6 . 4 ) , 10% ( v/v ) PEG8000 . For both conditions , 0 . 25 µl of 1 M NDSB-256 was added to the crystallisation drop . Rod-shaped crystals of the unmethylated complex appeared after 3 h and grew to 0 . 05 mm ( maximum dimension ) after 24 h . The methylated sample yielded bigger crystals that appeared after 24 h and grew to a maximum length of 0 . 5 mm after 3 d . Crystals were flash frozen in liquid nitrogen after cryoprotection with mother liquor containing 20% ( v/v ) glycerol ( unmethylated ) and 25% ( v/v ) PEG8000 and 10% ( v/v ) PEG300 ( methylated ) . Data were collected at 100 K on stations ID14-3 , ID14-4 , and ID23-2 at the European Synchrotron Radiation Facility ( ESRF ) and processed using the MOSFLM and SCALA programs from the CCP4 package [42] ( Table 1 ) . The structures of the unmethylated/methylated complexes were solved by a combination of molecular replacement with MOLREP [43] and real-space searches with FFFEAR [44] . An initial molecular replacement run was carried out with MOLREP using the 1 . 85 Å structure ( Protein Data Bank ID [PDB ID] 1UPK ) of MO25α [16] as a search model . Using the resulting phases , the STRADα molecule was then located by performing a real-space search with FFFEAR [44] using the 2 . 1 Å structure ( PDB ID 1U5R ) of TAO2 [23] . Thus , a solution with one complex in the asymmetric unit was found , and the structure was refined by alternating rounds of refinement with REFMAC5 [45] ( including TLS refinement during the last macrocycles ) and manual model building with the program COOT [46] . For the methylated complex , this resulted in a final model with an R-factor of 0 . 206 ( Rfree = 0 . 254 ) that was validated using PROCHECK [47] and MOLPROBITY [48] ( Table 1 ) . STRADα residues 292–347 , 383–385 , and 402–424 , and MO25α residues 337–341 were not associated with clear electron density and were not included in the model . Figures were prepared using the PyMOL molecular graphics system available at http://www . pymol . org [49] . Secondary structure was analysed using DSSP [50] and sequence alignments were performed using MUSCLE [51] , which were edited and displayed using the program ALINE developed by Charlie Bond and Alexander Schüttelkopf . Two hundred ninety-three cells were cultured on 10-cm diameter dishes in 10 ml of DMEM supplemented with 10% ( v/v ) fetal bovine serum , 2 mM l-glutamine , 100 U/ml penicillin , and 0 . 1 mg/ml streptomycin . For transfection experiments , 3–9 µg of DNA were mixed with 20 µl of 1 mg/ml polyethylenimine ( Polysciences ) in 1 ml of plain DMEM for each dish; the mixture was left to stand for 30 min and added onto the cells . Cells were lysed 36 h posttransfection in 1 ml of ice-cold lysis buffer per dish . The cell lysates were clarified by centrifugation at 20 , 000g for 15 min at 4°C , and the supernatants divided into aliquots , frozen in liquid nitrogen , and stored at −20°C . 10-cm diameter dishes of 293 cells were transiently transfected with 3 µg of the pEBG-2Tconstructs together with 3 µg of the indicated pCMV5 constructs as described above . Cells were harvested and lysed 36-h posttransfection , and the clarified lysates were incubated for 1 h on a rotating platform with glutathione-Sepharose ( GE Healthcare; 20 µl/dish of lysate ) previously equilibrated in lysis buffer . The beads were washed twice with lysis buffer containing 150 mM NaCl and twice with 50 mM Tris HCl , pH 7 . 5 . For immunoblotting analysis , the beads were resuspended in SDS sample buffer after this step and the samples immunoblotted as described above . For protein kinase assays and gel electrophoresis , the beads were washed twice more with Buffer A , and the proteins were eluted from the resin by incubation with the same buffer containing 270 mM sucrose and 20 mM of reduced glutathione . The beads were then removed by filtration through a 0 . 44-µm filter , and the eluate was divided into aliquots and stored at −80°C . The activity of recombinant LKB1/STRADα/MO25α complexes was assayed towards the LKBtide peptide substrate . All assays were performed by using 0 . 35 µg of recombinant proteins expressed and purified from HEK293 cells as described above . Phosphotransferase activity towards the LKBtide peptide ( SNLYHQGKFLQTFCGSPLYRRR ) [52] was measured in a total assay volume of 50 µl consisting of 50 mM Tris-HCl ( pH 7 . 5 ) , 0 . 1 mM EGTA , 0 . 1% ( v/v ) 2-mercaptoethanol , 10 mM magnesium acetate , 0 . 1 mM [γ-32P]ATP ( 200 cpm/pmol ) , and 0 . 2 mM LKBtide peptide . The assays were carried out at 30°C and were terminated after 15 min by applying 40 µl of the reaction mixture onto P81 membranes . These were washed in phosphoric acid , and the incorporated radioactivity was measured by scintillation counting as described previously for MAP kinase [53] . One unit ( U ) of activity represents the incorporation to the substrate of 1 nmol of γ-32P per minute . The AMPK heterotrimeric complex was purified from E . coli , and the AMPK activity was measured following its phosphorylation with LKB1 as reported by Lizcano et al . [52]; 10 µg of AMPK complex ( α1β2γ1 subunits ) was incubated with or without 0 . 3 ng of wild-type or mutant LKB1/STRADα/MO25α complex in Buffer A containing 5 mM magnesium acetate and 0 . 1 mM cold ATP , in a final volume of 20 µl . After incubation at 30°C for 30 min , the AMPK kinase activity was determined by adding 30 µl of 5 mM magnesium acetate , 0 . 1 mM [γ-32P]ATP ( 300 cpm/pmol ) , and 0 . 2 mM AMARA peptide ( AMARAASAAALARRR ) [54] as substrate . After incubation for 20 min at 30°C , incorporation of γ-32P into the peptide substrate was determined by applying the reaction mixture onto P81 phosphocellulose paper and scintillation counting as described in the previous section . One unit ( U ) of activity represents the incorporation to the substrate of 1 nmol of γ-32P per minute . The indicated amounts of cell lysates or purified proteins were subjected to SDS-PAGE and transferred to nitrocellulose membranes . The membranes were blocked for 1 h in TBS-T buffer containing 5% ( w/v ) skimmed milk . The anti-GST , anti-Flag , and anti-Myc antibodies ( Sigma ) were diluted 1 , 000-fold before the membranes were immunoblotted in the same buffer containing the indicated antibodies , for 16 h at 4°C . Membranes were then washed six times with TBS-T buffer and incubated with the appropriate horseradish peroxidase-conjugated secondary antibodies ( Pierce ) in TBS-T buffer containing 10% ( w/v ) skimmed milk . After repeating the washing steps , detection was performed using the enhanced chemiluminescence reagent ( Amersham Pharmacia Biotech ) , and the films were developed using a film automatic processor ( SRX-101; Konica Minolta Medical ) . For nucleotide binding experiments , wild-type and mutant forms of STRADα ( residues 54–431 ) and MO25α ( residues 1–341 ) were expressed individually as GST fusion proteins in E . coli . Cells were grown in Luria Bertani medium to A600 = 0 . 7 at 37°C , and protein expression was induced by the addition of 250 µM IPTG and incubated for a further 16 h at 26°C . Cells were harvested by centrifugation for 30 min at 3 , 500g and resuspended in ice-cold wash buffer ( 50 mM Tris-HCl ( pH 7 . 8 ) , 150 mM NaCl , 5% ( v/v ) glycerol , 1 mM benzamidine , 1 mM EGTA , 1 mM EDTA , 0 . 1 mM PMSF , and 0 . 01% ( v/v ) β-mercaptoethanol , supplemented with 0 . 5 mg/ml lysozyme and 0 . 3 mg/ml DNAse-I . Cells were lysed by sonication ( 10 × 10 s pulses ) and clarified lysates ( by centrifugation at 26 , 000g ) were incubated for 1 h on a rotating platform with glutathione-Sepharose ( GE Healthcare; 0 . 5 ml/l of culture ) pre-equilibrated in wash buffer . The beads were then washed with ten column volumes of wash buffer and a further 50 column volumes of high-salt wash buffer containing 500 mM NaCl . Beads were re-equilibrated in ten column volumes of wash buffer , and the proteins were eluted by incubating with PreScission protease for 16 h . Protein eluates were dialysed for 16 h against 5 l of assay buffer containing 50 mM Tris-HCl ( pH 7 . 8 ) , 50 mM NaCl , 270 mM sucrose , and 1 mM DTT , concentrated to 7 mg/ml , divided into aliquots , and stored at −80°C . For SPR measurements , wild-type and mutant forms of MO25α were expressed and purified as above . His-STRADα ( residues 59–431 ) was isolated in complex with MO25α as described for crystallisation . After gel filtration ( GF ) in GF buffer containing 50 mM Tris-HCl ( pH 7 . 8 ) , 50 mM NaCl , 270 mM sucrose , and 0 . 075% ( v/v ) β-mercaptoethanol , the STRADα/MO25α complex ( 20 mg ) was resuspended in 20 ml of binding buffer ( BB ) , consisting of GF buffer with increased NaCl concentration ( 300 mM ) . This sample was passed through 2 ml of Ni2+-agarose beads ( Invitrogen ) , equilibrated in BB , and the beads were washed with 50 column volumes of BB containing 500 mM NaCl and were re-equilibrated with ten column volumes of BB . His-STRADα was eluted in binding buffer supplemented with 150 mM imidazole . The eluted His-STRADα sample was equally divided and dialyzed against the assay buffer mentioned above . Untagged STRADα was obtained by incubation with TEV protease for 16 h at 4°C . Uncleaved STRADα and the TEV protease were removed by passing the postcleavage sample through Ni2+-agarose beads . His-STRADα and untagged STRADα were finally dialyzed into assay buffer , concentrated , and stored as above . Protein concentrations were determined by measuring the absorbance of the purified proteins at 280 nm in assay buffer . Fluorescent measurements of TNP-ATP ( Molecular Probes ) , were obtained at 25°C in assay buffer ( with the addition of 0 . 5–1 . 0 mM MgCl2 where indicated ) using 1-cm pathlength cuvettes in a VARIAN Cary Eclipse Fluorescence spectrophotometer ( Varian ) . Fluorescence was recorded using a 410-nm/540-nm excitation/emission wavelengths from 500 to 600 nm . In all cases , signal from the TNP-ATP buffer control was subtracted as background . For all binding studies , STRADα and STRADα mutants were assayed at 2 µM . In cases where STRADα/MO25α complexes were assayed , wild-type or mutant MO25α ( 2 µM ) were preincubated for at least 2 h at 4°C prior to a fluorescence binding experiment . For saturation binding experiments , concentrated stocks of TNP-ATP were added stepwise , covering a range of concentrations from 0 . 05 to 30 µM . For displacement experiments , the concentration of TNP-ATP was fixed at 5 µM , and ATP or ADP was titrated in , covering a range of concentrations from 0 . 05 to 500 µM . In all assays , concentrated stocks of nucleotides were added to 1 ml of reaction mixture in steps of 0 . 5 to 1 . 0 µl , ensuring that the total added volume did not exceed 1% of the total volume of the reaction . All data were analysed using GraphPad-PRISM software ( http://www . graphpad . com ) . To calculate the Kd values for TNP-ATP , data from saturation binding experiments were fitted to the following quadratic equation suitable for tight binding interactions with ligand depletion [55]:where [RL] equals the concentration of receptor/ligand complex , calculated as the fractional occupancy ( Fx/Fmax ) × [R]; [R] equals the total binding capacity , fixed at 1 . 5 µM; and [L] equals the concentration of added TNP-ATP . In the displacement studies , equilibrium constant values for ATP and ADP were calculated by first determining the logEC50 value , using a standard dose-response equation: Fx/Fmax = minimum+ ( maximum−minimum ) / ( 1+10 ( [N]−logEC50 ) ) , where [N] equals the concentration of added nucleotide , and Fx/Fmax represents the fractional occupancy . Equilibrium constants for the competing ATP and ADP ( KdN ) , were fitted using the equation: logEC50 = log ( 10logKdN × ( 1+[TNP−ATP]/KdTNP−ATP ) ) . SPR measurements were performed using a BIAcore T100 instrument . Wild-type and mutant forms of MO25α were immobilized on a CM5 sensor chip using standard amine-coupling chemistry , and 10 mM HBS ( pH 7 . 4 ) was used as the running buffer . The carboxymethyl dextran surface was activated with a 7-min injection of a 1∶1 ratio of 0 . 4 M 1-ethyl-3- ( 3-dimethylaminopropyl ) carbodiimide hydrochloride ( EDC ) /0 . 1 M N-hydroxy succinimide ( NHS ) . MO25α ( 5–7 µM ) was coupled to the surface with a 1-min injection of protein diluted in 10 mM sodium acetate ( pH 5 . 5 ) . Remaining activated groups were blocked with a 7-min injection of 1 M ethanolamine ( pH 8 . 5 ) . MO25α was immobilised on three flow cells of a CM5 chip at densities 1 , 700–2 , 500 RU performed at 25°C , leaving one flow cell as a reference to subtract any possible nonspecific binding . STRADα was prepared in running buffer containing 50 mM Tris ( pH 7 . 8 ) , 50 mM NaCl , 270 mM sucrose , 1 mM DTT , 0 . 005% P20 , and 0 . 1 mg/ml BSA in the presence/absence of 100 µM ATP and 1 mM MgCl2 , and injected over all four surfaces at nine concentrations of a 3-fold concentration series ( 5 µM to 0 . 3 nM ) . Each concentration was injected in duplicate over all surfaces . Association was measured for 60 s at a flow rate of 50 µl/min , and dissociation was measured for 3 min . STRADα dissociated completely from the MO25α surfaces , thus eliminating the need for a regeneration step . Data were analysed using Scrubber 2 ( BioLogic Software ) and CLAMP software . Data were double referenced to the reference surface to subtract any possible nonspecific binding and to the blank buffer injections to subtract drift of the target from surface . Data were fitted to a 1∶1 or 2∶1 binding site model where appropriate . Kinetic association ( ka ) and dissociation rate ( kd ) constants were separately determined from the BIAcore sensorgrams , and equilibrium dissociation constants ( Kd ) were calculated as: Kd1 = kd1/ka1 and Kd2 = kd2/ka2 . Equilibrium constants were also independently calculated from a saturation binding curve , by fitting the measured response ( R ) from specific binding to the following equation: R = ( Rmax1[STRAD]/ ( [STRAD]+Kd1 ) ) + ( Rmax2[STRAD]/ ( [STRAD]+Kd2 ) ) , where Rmax1 and Rmax2 are the relative maximal changes in response for sites 1 and 2 , respectively , and Kd1 and Kd2 are the equilibrium dissociation constants for sites 1 and 2 , respectively . Dose-response curves for calculating the Hill slope ( H ) of the data were fitted with the following equation: R = minimum+ ( maximum−minimum ) / ( 1+10 ( ( logEC50−[STRAD] ) ×H ) ) using GraphPad-PRISM software . Coordinates and observed structure factor amplitudes have been deposited at the Worldwide Protein Data Bank ( wwPDB , http://www . wwpdb . org/ ) , with accession code 3GNI .
|
There are 518 human protein kinases that are responsible for orchestrating the phosphorylation-dependant signal transduction events that regulate almost all cellular processes . Curiously , approximately 10% of protein kinases lack one or more catalytic residues , and these kinases have been termed pseudokinases . It has been proposed that some pseudokinases act as scaffolds , bringing together proteins involved in signalling networks . Here , we report the structure of the pseudokinase STRADα in complex with the adaptor protein MO25α; together these two proteins regulate the LKB1 tumour suppressor kinase . Despite lacking several key catalytic residues , STRADα binds ATP and adopts an active conformation typical of catalytically competent kinases . The affinity of STRADα for ATP is enhanced by MO25α and vice versa . We go on to demonstrate through mutagenesis studies that binding to both ATP and MO25α is essential for the activation of LKB1 . Our data suggest that STRADα exerts its functions through an active conformation , not through actual catalytic activity , thus raising the possibility that pseudokinases regulate signalling networks by adopting different structural conformations .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"biology/cell",
"signaling",
"biochemistry/cell",
"signaling",
"and",
"trafficking",
"structures",
"biochemistry/biomacromolecule-ligand",
"interactions",
"biochemistry/macromolecular",
"assemblies",
"and",
"machines",
"biochemistry/structural",
"genomics",
"molecular",
"biology"
] |
2009
|
ATP and MO25α Regulate the Conformational State of the STRADα Pseudokinase and Activation of the LKB1 Tumour Suppressor
|
Wolbachia endosymbionts carried by filarial nematodes give rise to the neglected diseases African river blindness and lymphatic filariasis afflicting millions worldwide . Here we identify new Wolbachia-disrupting compounds by conducting high-throughput cell-based chemical screens using a Wolbachia-infected , fluorescently labeled Drosophila cell line . This screen yielded several Wolbachia-disrupting compounds including three that resembled Albendazole , a widely used anthelmintic drug that targets nematode microtubules . Follow-up studies demonstrate that a common Albendazole metabolite , Albendazole sulfone , reduces intracellular Wolbachia titer both in Drosophila melanogaster and Brugia malayi , the nematode responsible for lymphatic filariasis . Significantly , Albendazole sulfone does not disrupt Drosophila microtubule organization , suggesting that this compound reduces titer through direct targeting of Wolbachia . Accordingly , both DNA staining and FtsZ immunofluorescence demonstrates that Albendazole sulfone treatment induces Wolbachia elongation , a phenotype indicative of binary fission defects . This suggests that the efficacy of Albendazole in treating filarial nematode-based diseases is attributable to dual targeting of nematode microtubules and their Wolbachia endosymbionts .
Wolbachia are intracellular maternally transmitted bacteria present in the majority of all insect species as well as some mites , crustaceans and filarial nematodes [1] , [2] . Wolbachia were initially studied in insects because they induce unconventional reproductive phenotypes including sperm-egg cytoplasmic incompatibility , feminization of males , male-killing , and parthenogenesis [3] , [4] . Wolbachia are essential endosymbionts of some filarial nematodes and recent studies demonstrated that they are the causative agent of African river blindness and also contribute to lymphatic filariasis [5] , [6] . One sixth of the world population is at risk of infection by Wuchereria bancrofti , Brugia timori and Brugia malayi , the filarial nematode species that cause lymphatic filariasis [7] . Wolbachia released from filarial nematodes into the human body trigger an inflammatory reaction that underlies the lymphedema and corneal occlusion associated with these neglected diseases [8] , [9] , [10] , [11] , [12] , [13] , [14] , [15] , [16] , [17] , [18] . Lymphatic filariasis and African river blindness have traditionally been treated through the administration of three drugs , singly or in combination: diethylcarbamazine ( DEC ) , ivermectin ( IVM ) and albendazole ( ALB ) . These drugs target the filarial nematodes associated with these diseases , namely Onchocerca volvulus , B . timori , B . malayi , and W . bancrofti [5] , [6] , [19] . DEC disrupts the nematodes by targeting the arachidonic acid metabolic pathway in the host [20] . IVM disrupts glutamate-gated chloride channels in the nematode that control release of excretory/secretory vesicles that would normally suppress the immune response [21] , [22] . ALB is a benzimidazole used to disrupt the nematode microtubule cytoskeleton [23] . Orally administered ALB is rapidly metabolized by in the intestinal mucosa and liver into albendazole sulfoxide ( ALB-SO ) and albendazole sulfone ( ALB-SO2 ) [24] , [25] . ALB-SO is normally considered to be the “active , ” form of Albendazole against systemic parasites , while ALB-SO2 is considered to be an inactive form of the drug [26] . All three drugs exhibit microfilaricidal effects [19] . The macrofilaricidal effects of ALB are not clear , though specific formulations induce worm sterility in animal models [27] . In addition , a number of clinical trials demonstrate that ALB when used in combination with DEC or IVM is macrofilaricidal [28] , [29] , [30] , [31] , [32] . Wolbachia are obligate symbionts of filarial nematodes required for normal embryogenesis , larval development and perhaps most significantly adult survival [33] , [34] , [35] , [36] , [37] , [38] , [39] , [40] , [41] . A recent study demonstrated that loss of Wolbachia in the adult results in high levels of apoptosis throughout the nematode [37] . Studies have also found that much of the pathology associated with filarial nematode diseases is due to induction of innate and adaptive host immune responses upon release of Wolbachia from their nematode hosts [8] , [9] , [10] , [11] , [12] , [13] , [14] , [15] , [16] , [17] , [18] . These discoveries suggest that compounds directly targeting Wolbachia may be a powerful alternative to the more traditional approaches for treating these diseases . The major advantage of this approach is that it targets adults as well as microfilaria and the Wolbachia will be eliminated prior to death of the nematode , reducing the destructive effects of the human immune response . In addition , loss of Wolbachia leads to a slow death of the adults , providing time for the infected individual to clear the dead nematodes without deleterious side effects [35] , [42] , [43] , [44] , [45] , [46] , [47] . Finally , antihelminthic drugs such as ivermectin , administered to patients co-infected with Loa Loa nematodes , can potentially trigger lethal encephalitis [48] . Loa loa does not require or maintain Wolbachia and thus will not be affected by anti-Wolbachia therapies , thereby avoiding these deleterious side effects [49] . The promise of the anti-Wolbachia based therapies in combating lymphatic filariasis has been demonstrated in clinical trials in which daily doses of doxycycline ( DOX ) for 4 weeks resulted in nematode sterility and death [6] . In addition , the pathologies associated with the infection , lymphedema and hydrocele , were dramatically reduced [8] , [9] . These studies also revealed that a three-week course of DOX was insufficient to produce significant mortality of the adult nematodes , highlighting the need to identify more potent anti-Wolbachia compounds [50] . To this end , we generated a Wolbachia-infected Drosophila cell line and conducted an automated , cell-based screen to identify lead compounds that reduced intracellular Wolbachia infection . This screen of two libraries totaling 4926 compounds yielded 40 anti-Wolbachia compounds , including several that structurally resembled ALB . Our follow-up testing indicated that ALB-SO2 directly targets Brugia by depolymerizing its microtubules . Here we demonstrate that ALB-SO2 also disrupts Brugia Wolbachia independently of its effects on the Brugia cytoskeleton . Furthermore , ALB-SO2 treatment of Brugia led to visibly elongated Wolbachia morphology indicative of a binary fission failure , consistent with a direct impact of ALB-SO2 upon B . malayi Wolbachia .
To identify compounds that affect intracellular Wolbachia titer , we generated new Drosophila tissue culture cells constitutively infected with Wolbachia [51] ( see Methods ) . There is currently no stable nematode tissue culture line , nor any type of cell line constitutively infected with Wolbachia derived from filarial nematodes . Wolbachia-infected insect cells provide an opportunity to identify drugs that disrupt Wolbachia through conserved molecular pathways . The cell line used for this study , JW18 , is particularly amenable to high throughput screening as Wolbachia are maintained in approximately 90% or more of the host cell population ( n = 1053 cells scored ) . The Wolbachia load in 6 . 7% of infected cells ranges from 1–46 bacteria , while Wolbachia load in the other 93% of infected cells is obscured by crowding of the bacteria ( n = 205 cells scored ) . The mitotic index of JW18 and tetracycline-cured JW18 cells , henceforth referred to as JW18TET , was 0 . 27% and 0 . 68% respectively , which are not significantly different according to Chi square tests ( n = 1876 and 2339 ) . Furthermore , no significant difference was observed in the frequency of binucleate cells between JW18 ( 9 . 1% , n = 873 ) and JW18TET cells ( 11% , n = 1081 ) . Thus , Wolbachia do not exert an obvious influence on the regulatory or structural mechanisms governing progression of the cell cycle in the JW18 cell line . However this analysis does not preclude more subtle cycle cell effects . To test whether Wolbachia exhibit normal interactions with the host cytoskeleton , we took advantage of the fact that this cell line constitutively expresses a GFP-Jupiter fusion protein that binds to and labels microtubules [52] ( Figure 1 ) . During interphase , Wolbachia are closely associated with Jupiter-GFP-labeled microtubules ( Figure 1A–B , Video S1 ) . Live imaging demonstrates that Wolbachia move processively along those interphase microtubules ( Video S2 ) , consistent with earlier reports of Wolbachia-microtubule interactions [53] , [54] , [55] , [56] . During mitosis , Wolbachia were asymmetrically distributed throughout the cytoplasm in 82% of cells ( n = 56 , Figure 1C–F , Table S1 ) , reminiscent of Wolbachia localization patterns observed in embryonic and larval neuroblasts [56] . These data indicate that Wolbachia distribution in the JW18 cell line is consistent with that of intact Drosophila tissues . High-throughput cell-based screens using automated microscopy have proven effective in identifying new compounds targeting specific biological processes [57] , [58] . Here we used the JW18 cell line in a 384 well format to screen 2000 compounds from the Spectrum Collection and 2926 compounds from the National Cancer Institute for reduction of Wolbachia titer . These libraries include structurally and functionally diverse synthetic compounds , FDA-approved drugs , compounds with biological activity and a set of natural products . After a 5-day incubation period , the cells were fixed , stained , and imaged using automated robotics ( Figure S1 , Materials and methods ) . Customized software was used to analyze the images and score the percentage of Wolbachia-infected cells in each well . Treatment wells showing significant reduction of Wolbachia-infected cells in at least 2 of 3 replicates as compared to the JW18 control and the entire cell population in general were scored as preliminary hits . Compounds known to be generally hazardous were excluded , resulting in a finalized hit list . 18 compounds from the Spectrum library , and 22 compounds from the NCI library exhibited consistent , potent anti-Wolbachia activity in the screen ( Figure 2 , Table S2 ) . A number of these compounds are already known to exert antimicrobial activity , consistent with expectations from the assay . For example , one of the hits CID484401 ( totarol acetate ) is a known inhibitor of the essential bacterial division protein , FtsZ [59] . CID42640 is a DNA damaging agent that has been shown to inhibit Mycobacterium tuberculosis [60] ) . Significantly , the hit CID313612 is structurally similar to CID42640 . CID16524 ( pyronin B ) is a quaternary ammonium compound , many of which serve as the antimicrobial agents in commercial disinfectants [61] . An additional hit compound , CID6364517 has also been shown in a prior screen to exert antibacterial activity against Streptococcus pyrogenes ( Pubchem BioAssay AID1900 and AID1915 , conducted by the Broad Institute ) . Furthermore , a number of chemotherapy drugs were identified by the screen: CID30323 ( daunorubicin ) , CID31703 ( doxorubicin ) , CID65348 ( epirubicin ) and CID4212 ( mitoxantrone ) were identified as hits , along with the derivitives CID5351490 ( cinerubin B ) and CID27590 ( mitomycin B ) . Many of these drugs were originally isolated from bacteria in nature . It is presumed that these compounds are used to gain a competitive advantage over neighboring bacteria . As such , this class of drugs is referred to as to “anticancer antibiotics” [62] , [63] , [64] , [65] . Doxycycline ( DOX ) did not come up as a hit in this screen . This was unexpected as DOX has proven to be an effective anti-Wolbachia reagent in lab and clinical settings [6] . However , treating the same Wolbachia strain with DOX in the context of Drosophila oogenesis revealed marked Wolbachia susceptibility to DOX ( Figure 3A–C ) . This suggests that the mechanism accounting for this difference in DOX efficacy is dependent upon properties of the host cell rather than the Wolbachia strain . It may be the efflux pumps in the JW18 cell line are particularly effective at expelling DOX . Alternatively , the bacteriostatic effects of DOX may not readily be detected in our assay because of the relative growth rates of Wolbachia and the JW18 host cells . Regardless of the mechanism , it is not general , as this screen yielded a number of new potent anti-Wolbachia compounds . To further investigate the hit compounds revealed by the screen , we first examined their structures . This revealed that the benzimidazole CID5382764 and the benzthiazoles CID5458770 and CID5351210 share structural similarity to ALB , the widely used , FDA-approved anthelmintic drug known for disrupting helminthic microtubules ( Figure 4 ) . The hit compounds were also assessed for cytotoxicity by consulting prior screens run by the NCI against human cancer cell lines and measuring the impact of the compounds on cell proliferation in our own assay . This indicated that 11 of 22 hits from the NCI library and 7 of 18 hits from the Spectrum library exhibit unacceptable toxicity levels , while the remaining 22 hit compounds exhibited low or unknown cytotoxicity ( Figure 2 , Table S2 ) . By this measure , two of the ALB-like compounds , CID5458770 and CID5382764 , were indicated to be non-toxic ( Figure 2 ) . Drosophila has proven a particularly effective model system for in vivo compound testing of a wide array of biological processes [66] . Here we used the Drosophila oocyte as a secondary screen for retesting the ALB-like hits identified in the cell-based screen . This system provides the advantages that the Wolbachia strain is the same as in the JW18 cell line , and Wolbachia titer in the oocyte is known to greatly increase between stage 3 and 10A , approximately a 40-hour period [54] , [67] , [68] . Starved adult Drosophila were fed yeast paste containing compounds at a concentration of 100 uM for 24 hours . Stage 10A oocytes were fixed and labeled , followed by imaging and quantification of their Wolbachia as described in the Materials and Methods . Average titer measurements from single focal planes have been shown to be representative for comparing Wolbachia titer between different conditions [67] . DMSO control oocytes exhibited 268+/−14 . 3 Wolbachia within a single oocyte focal plane ( Figure 3A–C ) . Oocytes treated with DOX exhibited 141+/−12 . 4 Wolbachia ( p< . 001 ) . Treatments with CID5458770 yielded oocytes displaying 174+/−12 . 7 Wolbachia ( p< . 001 ) . Furthermore , CID5382764-treated oocytes exhibited 161+/−15 . 9 Wolbachia ( p< . 001 , Figure 3 ) . Thus , both of the non-toxic , ALB-like hits identified by the cell screen deplete Wolbachia titer in Drosophila oogenesis similarly to DOX . These results motivated us to determine whether ALB and its common metabolites , ALB-SO and ALB-SO2 ( Figure 4 ) , also affect Wolbachia titer in vivo . Wolbachia counts from single oocyte focal planes indicated that ALB and ALB-SO treatments did not significantly affect Wolbachia titer , with 216+/−22 . 5 and 299+/−30 . 0 Wolbachia detected per oocyte , respectively ( Figure 3C ) . However , ALB-SO2-treated oocytes exhibited markedly less Wolbachia than the control , with 167+/−14 . 8 Wolbachia evident per oocyte ( p< . 001 , Figure 3 ) . This indicates that the ALB-SO2 metabolite exerts anti-Wolbachia effects in Drosophila . To investigate whether the anti-Wolbachia effect of ALB-SO2 applies to a disease model , we treated adult B . malayi nematodes in vitro . After a 3-day incubation period , Wolbachia were imaged in the distal tip region of the Brugia ovary referred to as the “mitotic proliferation zone” due to enriched replication of host nuclei in this tissue [69] . Wolbachia densely populate this region of the distal ovary in DMSO controls ( Figure 3D ) . By contrast , Wolbachia titer is visibly depleted in worms treated with either ALB-SO2 or DOX ( Figure 3E–F ) . Quantitation of Wolbachia further supports this observation , with DMSO-treated Brugia exhibiting 5 . 4 Wolbachia on average per host nucleus ( n = 419 bacteria scored ) . By contrast , ALB-SO2-treated worms carried significantly fewer bacteria , exhibiting 1 . 3+/−0 . 45 Wolbachia per host nucleus ( n = 263 ) ( p< . 001 ) . This depletion was similar to DOX-treated worms , which exhibited 1 . 1+/−0 . 19 Wolbachia per host nucleus ( n = 283 ) ( p< . 001 ) . This indicates that the Wolbachia-depleting impact of ALB-SO2 extends to B . malayi . A consistent Wolbachia-disrupting effect for ALB-SO2 raises the question of the mechanism of action of this metabolite . ALB and other benzimidazoles are known to bind to beta-tubulin and disrupt microtubule polymerization [70] , [71] , [72] , [73] . Previous studies from Drosophila have demonstrated that Wolbachia titer is affected by host microtubules [54] , [67] . This raises the possibility that the reduction of Wolbachia titer upon exposure to ALB-SO2 is due to an impact of this compound on microtubule organization . Prior mutant studies identified key amino acids within beta tubulin that are important for the microtubule-disrupting impact of benzimidazoles . Residues N165 and Y200 are thought to form a hydrogen bond that restricts accessibility to a benzimidazole binding site [74] . Interestingly , most of the beta tubulin homologs in Drosophila encode the residues that correspond to benzimidazole resistance ( Figure S2 ) . This suggests that the Drosophila microtubule cytoskeleton should be unaffected by ALB-SO2 treatment . To test whether ALB-SO2 affects microtubule organization in Drosophila oogenesis , a combination of approaches was used . Intracellular Wolbachia localization was examined , taking advantage of the prior finding that Wolbachia concentrate at the oocyte posterior cortex in a microtubule-dependent manner [55] , [75] . In this study , posterior Wolbachia localization was detected in 95% of DMSO controls and 93% of ALB-SO2-treated oocytes ( n = 56 and 46 , respectively , Figure 5G–H ) . This differed significantly from oocytes treated with the microtubule-disrupting drug , colchicine , where only 21% exhibited posterior Wolbachia localization ( p< . 001 , n = 13 , Figure 5I ) [55] . The Drosophila oocyte cytoskeleton was also directly examined by immunolabeling microtubules [76] . Stage 10B oocytes are known to undergo large-scale , microtubule-dependent cytoplasmic streaming , coincident with formation of microtubule bundles . As cytoplasmic streaming is a highly dynamic process , this bundling varies somewhat between oocytes , and changes within individual oocytes over time [76] , [77] . DMSO controls and oocytes treated with ALB-SO2 exhibited microtubule bundling at stage 10B , while the cytoplasm of colchicine-treated stage 10B oocytes was devoid of filamentous structure ( Figure 5J–L ) [77] . Thus , ALB-SO2 did not affect the overall orientation or structure of the oocyte microtubule cytoskeleton in Drosophila , though this compound dramatically decreases Wolbachia titer ( Figure 3 ) . This indicates that ALB-SO2 reduces Wolbachia titer through a microtubule-independent mechanism in Drosophila . The findings above raise the question of whether ALB-SO2 disrupts Wolbachia titer independently of microtubules in B . malayi , analogous to Drosophila . The beta-tubulin genes of B . malayi carry the amino acid changes of N165S and Y200F that are associated with susceptibility to benzimidazoles like ALB-SO2 ( Figure S2 ) . To assess the impact of ALB-SO2 on B . malayi microtubules in vivo , whole-mount immunostaining was performed . Examination of Brugia hypodermal chords revealed a dense meshwork of microtubules in DMSO-treated controls ( Figure 5A–C ) . In contrast , treatment with ALB-SO2 disrupted the Brugia microtubule cytoskeleton , although linear remnants remained visible throughout the hypodermal chord ( Figure 5D–F ) . This demonstrates that ALB-SO2 disrupts much of the host microtubule cytoskeleton in B . malayi . To next determine whether the titer of Brugia Wolbachia relies upon host microtubules , we treated B . malayi with colchicine . To verify that colchicine disrupts microtubules in B . malayi , immunostaining was performed in the hypodermal chords of colchicine-treated worms as above . This analysis revealed that microtubule structure and organization was largely eliminated by colchicine treatment . To then assess the impact of colchicine on Wolbachia titer , Wolbachia were imaged in the distal ovary . Interestingly , colchicine-treated worms exhibited an average of 5 . 1+/−0 . 93 Wolbachia per host nucleus ( n = 193 bacteria scored ) , a value that is not significantly different from its DMSO controls by Chi square test ( Figure 3G ) . This indicates that microtubule disruption has little impact on Wolbachia titer in B . malayi . This suggests that ALB-SO2 is also unlikely to suppress Wolbachia titer through its microtubule-disrupting effects , and alternatively implicates a microtubule-independent mechanism for ALB-SO2 suppression of Wolbachia titer in B . malayi . To further pursue the mechanism by which ALB-SO2 disrupts Wolbachia , we examined its impact on Wolbachia morphology in Brugia tissue . In the mitotic proliferation zone of the distal ovary , ALB-SO2 treatment corresponded to elongation of Wolbachia nucleoids as compared DMSO controls ( Figure 6C–F ) . Wolbachia were also immunostained with an antibody raised against Wolbachia FtsZ , which crisply defines the boundaries of the Wolbachia cytoplasm [78] . The FtsZ staining also revealed an elongated Wolbachia morphology in ALB-SO2-treated Brugia relative to the DMSO control ( Figure 6G–H ) . Furthermore , quantification of Wolbachia length indicated that only 2 . 5% of Wolbachia in the DMSO control exceeded 2 . 5 µm in length , whereas 17% of Wolbachia in ALB-SO2 treated worms exceeded this length ( n = 201 and 126 , respectively ) ( p< . 001 ) ( Figure 6I ) . This demonstrates that ALB-SO2 induces Wolbachia elongation in the Brugia ovary . Wolbachia were also examined in the hypodermal chords , which are syncytial somatic tissues that run along the entire length of the Brugia body wall . Comparing Wolbachia in DMSO and ALB-SO2-treated hypodermal chords revealed little difference in bacterial morphology ( Figure 6A–B ) . This indicates that Wolbachia in the hypodermal chord do not respond to ALB-SO2 , in contrast to Wolbachia in the distal ovary of B . malayi . A role for ALB-SO2 in inducing Wolbachia elongation could be due to an indirect effect from disrupting host microtubules , or through an alternative mechanism . To distinguish between these possibilities , B . malayi ovaries were treated with colchicine and assessed for Wolbachia length . Examination of DMSO vs . colchicine-treated Brugia reveals no obvious differences in Wolbachia morphology between conditions ( Figure 6J–K ) . Wolbachia length was also quantified , and neither DMSO nor colchicine-treated tissues harbored Wolbachia exceeding 2 . 5 µm in length ( Figure 6I ) . This differs significantly from ALB-SO2-treated Wolbachia , which frequently exceeded 2 . 5 µm in length ( p< . 001 ) ( Figure 6I ) . Thus , overall microtubule disruption in Brugia did not change Wolbachia morphology as was seen for ALB-SO2 treatment . This indicates that ALB-SO2 induces Wolbachia elongation independently of host microtubules in Brugia malayi .
Here we report the first high-throughput cell-based screen using automated microscopy to identify anti-Wolbachia compounds . The cell line used in this screen was established from primary culture preparations from Wolbachia-infected Drosophila embryos bearing the GFP-tagged microtubule binding protein Jupiter ( Figure 1 ) . Wolbachia have been stably maintained and transmitted in this line for years . Live analysis of interphase JW18 cells reveals that Wolbachia closely associate with and move along microtubules ( Video S1 ) , consistent with previous studies showing that Wolbachia transport and positioning are microtubule-dependent [54] , [55] , [56] . Although there is variability from generation to generation , approximately 90% of JW18 cells are infected with variable amounts of bacteria . This is consistent with previous studies of Wolbachia-infected Drosophila , Spodoptera frugiperda and Aedes albopictus cell lines showing widely variable bacteria loads per cell [79] , [80] , [81] , [82] , [83] , [84] . Perhaps this variation in Wolbachia load is due in part to asymmetric partitioning of Wolbachia in mitotic JW18 cells ( Figure 1 , Table S1 ) . This same Wolbachia strain has previously been reported to segregate asymmetrically in mitotic embryonic Drosophila neuroblasts and Aa23 tissue culture cells [56] , [82] . In addition , variants have been identified in D . simulans in which asymmetric segregation of Wolbachia may occur in the germline stem cells of the oocyte [85] . This contrasts with other evidence that Wolbachia partition evenly during mitosis in early Drosophila embryogenesis and in A . albopictus C7-10R tissue culture cells [56] , [83] . Future work is needed to address the underlying mechanisms that govern segregation of Wolbachia in mitotic cells . Our automated , image-based screens of 4926 compounds from the NCI and Spectrum libraries revealed 40 compounds that depleted Wolbachia titer . Of these , 22 were classified as either non-toxic or of unknown toxicity ( Figure 2 , Table S2 ) . Among the more characterized hit compounds , we identified totarol acetate , an inhibitor of the key bacterial division protein , FtsZ . This finding is consistent with a previous report describing FtsZ as a potential target for anti-Wolbachia therapy [86] . We also identified pararosaniline pamoate and pyrvinium pamoate as Wolbachia-depleting hits . These drugs have been employed as antihelmintics to treat patients infected with roundworms like Enterobius vermicularis and the trematode Schistosoma japonicum [87] , [88] . Our result raises an additional possibility that these hosts rely upon endosymbiotic bacteria to support their viability , and an anti-bacterial effect of these drugs thus indirectly compromises the host . No other Secernentea outside of the Filariodea have been reported to carry Wolbachia to date , though the Rhabditida order is the only one to have been systematically tested [89] . However , the Trematodes Acanthatrium oregonense and Nanophyetus salmincola have been reported to harbor the mutualistic bacteria Neorickettsia risticii and N . helminthoseca , respectively [90] . Furthermore , estradiol was also identified by our screen as a Wolbachia-depleting compound . This finding may help to explain why women exhibit symptoms of lymphatic filariasis less often than men in endemic areas , where 10% or fewer of females exhibit symptoms as compared to 10–50% of men [91] . Since Wolbachia can induce a TLR-2 and TLR-6-mediated inflammatory response analogous to that observed in lymphatic filariasis patients , the literature invokes a contribution of Wolbachia to the pathology of lymphatic filariasis [8] , [9] , [10] , [15] , [16] , [17] . Perhaps higher estradiol levels in female patients help to suppress Wolbachia load , thereby preventing TLR-2/6 induction and development of lymphatic filariasis symptoms . An examination of the largely uncharacterized hits from the JW18 screen revealed that CID5458770 , CID5382764 and CID5351210 structurally resemble ALB ( Figure 4 ) . Follow-up studies revealed that the non-toxic ALB-like compounds CID5458770 and CID5382764 exhibited Wolbachia-depleting activity in Drosophila oogenesis as well . This raised the possibility that ALB or its metabolites may have a similar activity in vivo . These studies surprisingly revealed that the ALB-SO2 metabolite depleted intracellular Wolbachia titer in Drosophila oogenesis , while ALB and ALB-SO did not ( Figure 3 ) . Furthermore , ALB-SO2 disrupts Wolbachia titer and morphology and alters host microtubules in B . malayi ( Figure 3 , 5 , 6 ) . These findings contrast with routine descriptions in the literature of ALB-SO2 as an inactive metabolite . It appears that ALB-SO2 inactivity has been interpreted from the lower abundance of the ALB-SO2 metabolite in human serum and urine relative to ALB-SO [24] , [25] , [26] as well as the relatively weaker ability of ALB-SO2 to compete against unlabeled benzimidazoles for binding to Haemonchus contortus tubulin in vitro [72] . However , the sequence of Brugia beta tubulin predicts susceptibility to benzimidazoles like ALB-SO2 ( Figure 4 , Figure S2 ) , and evidence from other non-nematode systems supports a role for ALB-SO2 as an active metabolite . An in vitro competitive binding assay using colchicine showed that ALB-SO2 is slightly better at disrupting polymerization of Ascaris suum tubulin than either ALB or ALB-SO [92] . Studies from the tapeworm Echinococcus multilocularis showed ALB-SO2 to be similarly effective to ALB-SO in inducing structural defects and worm lethality . Other studies of the microsporidian parasites Encephalitozoon cuniculi , E . hellem and E . intestinalis indicated that ALB-SO2 is at least 5 times more effective at inhibiting growth than either ALB or ALB-SO [93] . Building upon those prior findings , this study is the first we are aware of to definitively show an active role for the ALB-SO2 metabolite in filarial nematodes . A surprising outcome from this study was that ALB-SO2 disrupts Wolbachia independently of its effects on host microtubules . Previous studies have shown that host microtubules and microtubule-based motor proteins facilitate replication of Salmonella and Chlamydia [94] , [95] , [96] , [97] , [98] , [99] , [100] , [101] . Prior work in Drosophila indicates that normal levels of Wolbachia rely in part upon intact host microtubules as well [54] , [67] . This study confirms that Brugia microtubules are vulnerable to ALB-SO2 in vivo , raising the possibility that ALB-SO2 acts indirectly upon Wolbachia through the host cytoskeleton . However , colchicine treatments that eliminate Brugia microtubules had no significant impact on Wolbachia titer or elongation . Furthermore , ALB-SO2 suppressed Wolbachia titer in Drosophila even though there was no detectable impact on host microtubules in this system . Thus , ALB-SO2 does not disrupt Wolbachia through an influence on the host microtubule cytoskeleton . The elongation of Wolbachia induced by ALB-SO2 suggests a possible mechanism of action for this compound . It has been widely documented that extensive bacterial elongation , referred to as filamentation , ensues when binary fission is disrupted , due to continued growth of the bacteria despite the failure of abscission [73] , [102] . Thus , Wolbachia elongation in ALB-SO2-treated ovaries suggests that ALB-SO2 is preventing abscission of growing and replicating Wolbachia . By contrast , ALB-SO2 treatment had no visible impact on Wolbachia length in Brugia hypodermal chords . This suggests that the Wolbachia bacteria residing within the hypodermal chord are in a non-growing steady state as compared to the proliferating Wolbachia population in the distal region of the B . malayi ovary . If confirmed , this implies that bacteriostatic drugs may not be effective at clearing Wolbachia from the hypodermal chords . ALB-SO2 could disrupt Wolbachia binary fission by targeting a number of different factors . Proper binary fission relies upon assembly of FtsZ filaments at the future fission site , which then leads to recruitment of numerous other factors that help stabilize and constrict the division septum [103] . ALB has been shown to suppress titer and induce filamentation of Mycobacterium tuberculosis [104] , and numerous studies demonstrate that ALB-like compounds target and disrupt FtsZ in Staphylococci , Escherichia coli and M . tuberculosis [73] , [104] , [105] , [106] , [107] , [108] , [109] , [110] , [111] , [112] , [113] , [114] , [115] . A role for ALB-SO2 in disrupting Wolbachia FtsZ function would be consistent with this body of work . However , given staining variability and resolution limits of the Brugia system , it is not currently possible to distinguish whether FtsZ concentration or distribution in vivo is significantly different between DMSO and ALB-SO2-treated conditions . An alternative possibility is that ALB-SO2 disrupts binary fission without directly targeting FtsZ . For example , some ALB-like compounds can intercalate into DNA and thus serve as potential DNA damaging agents [73] . DNA damage is known to induce the FtsZ-inhibitor SulA , leading to bacterial filamentation [116] , [117] . It is further possible that ALB-SO2 targets other as-yet unrecognized factors that are key to initiation and execution of Wolbachia binary fission . This study has redefined the mechanism by which ALB acts against filarial nematodes . Building upon work by others , the data of this study suggest that ALB administered to humans disrupts B . malayi by a two-fold impact . First , ALB and its metabolites disrupt the filarial microtubule cytoskeleton , leading to rapid death of microfilariae . Second , our study indicates that the ALB-SO2 metabolite is directly targeting Wolbachia . Certain formulations of ALB were previously shown to lead to Brugia sterility in animal models , but the mechanism underlying this effect was unclear [27] . It is possible that microtubule disruption by ALB and its metabolites directly compromises the structure of the adult germline in Brugia . However , DOX-based elimination of Wolbachia from the female germline has also been shown to induce apoptosis in germline cells [37] . Perhaps ALB-SO2-mediated disruption of Wolbachia contributes further to destruction of the Brugia germline . ALB-SO2 may serve as a valuable asset in the campaign against Wolbachia-based disease . Used in tandem with conventional antibiotics , ALB-SO2 may shorten the effective treatment time of anti-Wolbachia treatments from the current 4 to 6 weeks [5] , [6] . It is likely that more potent derivatives of ALB-SO2 may be discovered with enhanced the specificity for Wolbachia . Identification of the ALB-SO2 target will enable optimization of the compound to provide a possible alternative route for future treatment of African river blindness and lymphatic filariasis .
The JW18 cell line bearing the Jupiter-GFP transgene was generated according to the method described in Karpova et al , 2006 . 1 to 15 hour old embryos derived from Wolbachia-infected flies carrying a Jupiter-GFP transgene were homogenized and plated in flasks . During the next six months of maintenance , five of the initial twenty seed flasks converted into immortal tissue culture lines . The JW18 cell line was selected for further pursuit due to its planar growth pattern and stable abundant Wolbachia infection . Cells were maintained at 25–26°C in Sang and Shields media containing 10% fetal bovine serum , split weekly at a 1:3 dilution . A cured version of the JW18 line , referred to as JW18TET , was generated by treating the cells with tetracycline [34] , [118] , [119] . A Chi-square test was used to test for differences in the frequency of mitosis between JW18 and JW18TET cells . Several small chemical libraries were used in this study . The Spectrum Collection ( MicroSource Discovery Systems , Inc ) of 2000 compounds contains 1000 drugs with known pharmacological properties , 600 natural products , and 400 other bioreactive compounds . The NCI Diversity Set I contains 1990 compounds selected for structural uniqueness and fewer than 5 rotatable bonds . The NCI Mechanistic Set contains 879 compounds , selected for their diversity of impact in human tumor cell lines screened by the NCI . The NCI Challenge Set contains 57 compounds that stood out in the NCI tumor cell screens because of the unusual patterns of cell lethality and resistance induced by these compounds . The stock concentration of the compounds in these libraries was 10 mM . Furthermore , these libraries were reformatted into 384-well plates by the UCSC screening center . Specifically , compounds were placed into columns 3–22 of the plates , leaving columns 1–2 and 23–24 vacant to allow for untreated control wells . Cells were plated in 384-well , clear bottom plates ( Griener Bio-one ) pre-coated with 0 . 5 mg/mL Concanavalin A . JW18 cells were added to 22 columns , and JW18TET cells were added to the remaining 2 columns at a concentration of 6500 cells per well . After the cells adhered to the plates for 4–6 hours , compounds were transferred into 20 columns of JW18 cells in the center of the plate using a Janus MDT pin tool . The final concentration of compound was 100 uM per well . All treatments were distributed into 3 plate replicates . After a 5-day incubation with the compounds at 25°C , the cells were prepared for imaging . Cells were fixed for 20 minutes in 4% formaldehyde and rinsed with PBS using an automated BioTek liquid handler . All staining solutions were administered using a Multidrop robot , with extensive rinsing between treatments . Mouse anti-histone ( MAB052 , Millipore ) and goat anti-mouse Alexa 594 ( Invitrogen ) was diluted to 1∶1250 in PBS/0 . 1% Triton . A stock solution saturated with DAPI was used at a final concentration of 1∶40 . After staining , PBS+sodium azide was added to all wells of the plates . Stained plates were imaged using an ImageXpress Micro system ( Molecular Devices , Sunnyvale CA ) . 10 images were acquired per well at 40× magnification . These images were analyzed using customized analysis software provided by Molecular Devices . The software routine masks any areas where clumps of cells are detected , based upon intensity of the Jupiter-GFP . The boundaries of the remaining cells and their nuclei are recognized based upon the Jupiter-GFP and anti-histone stains . A mask is applied to the nuclei , thus obscuring the histone and DAPI signal from those areas . A threshold for DAPI fluorescence detection is set to detect as much Wolbachia as possible in JW18 control cells while minimizing detection of background DAPI signal in JW18TET control cells . The remaining cytoplasmic DAPI , specifically labeling Wolbachia , is scored in individual cells to determine whether each cell is infected with Wolbachia . The cutoff value distinguishing “infected” from “uninfected” cells is 4000–5000 cytoplasmic DAPI fluorescence units per cell . This is a stringent limit , as values from 10 randomly selected 384-well plates indicate that the untreated JW18TET cells exhibit an average of 2383 +/−78 . 30 cytoplasmic DAPI fluorescence units per cell as compared to untreated JW18 cells , which exhibit an average of 26146 +/−880 . 4 DAPI fluorescence units per cell . A spreadsheet from the data analysis software indicates the quantity of Wolbachia-infected cells versus total cells measured in each well . An average and standard deviation were calculated to assess the frequency of Wolbachia infection in JW18 and JW18TET control cells , using the descriptive statistics function in SPSS ( IBM ) . These values were used to calculate a Z′ factor for each plate [120] . The Z′ factor represents 1 – ( the sum of the standard deviations for each control divided by the absolute value of the difference between mean values for each control . ) Z′ factors regarded as acceptable by the field range from 0–1 . Our Z′ factors range from 0 . 2 to 0 . 65 per plate , verifying that the infected and uninfected controls were clearly distinguishable by the assay . To identify preliminary “hit” compounds that substantially reduce intracellular Wolbachia titer , an initial hit range was calculated to lie between the JW18 average infection frequency − 3 standard deviations , and the average JW18TET infection frequency + 3 standard deviations [120] . To enable comparison of hits identified on different treatment plates , the scaling of this initial hit range was next reset to span from 0–1 , thus applying a uniform , normalized hit range to all plates . To further increase the stringency for identifying hits , we also calculated an average infection frequency for all JW18 cells on the plate ( treated or not ) , as most treatment wells are expected to be indistinguishable from untreated controls . Wells that lay within 3 standard deviations of the mean were removed from the hit list . Hit wells identified in only 1 of 3 replicates were also removed from the hit list . From the remaining hit candidates , we further identified compounds designated as hazardous by the National Cancer Institute , which includes alkylating agents , corrosives , carcinogens , explosives , flammables , oxidizers , poisons and known toxins . Any hits falling into these hazard classes were removed from further consideration . The hit compounds not excluded by these stringent criteria were designated as finalized hits . The cytotoxicity of each hit from the Spectrum and NCI library screens was determined using a stepwise classification process . The Pubchem Bioassay Database was mined to assess the impact of our hits in prior mammalian tumor cell cytotoxicity screens conducted by the NCI . The majority of our hit compounds have already been tested for cytotoxicity in 45–115 screens conducted by the NCI . For those hits , if over 50% of the NCI screens indicated the drug to be cytotoxic , we designated that hit as “toxic” in our listing . If 10%–50% of the NCI screens indicated cytotoxic properties for that compound , we classified it as “moderately toxic . ” If less than 10% of NCI screens indicated toxicity , we classified it “non-toxic . ” A subset of our hit compounds have been run in 2 or fewer prior NCI screens . For those , hits , a preliminary cytotoxicity designation was assigned based upon the cell density reported by our screen . The average cell density per well was first calculated for each set of treatments , and treatments that failed to shift the cell density of a single well more than +/−33% from the plate average were designated as “non-toxic . ” Treatments that increased the average cell density to more that 33% over the average density were classified as “toxicity unclear” , while treatments that reduced average cell density to 33%–66% below the average cell density were classified as “moderately toxic . ” ALB ( Sigma ) , ALB-SO and ALB-SO2 ( Santa Cruz Biotechnology ) were dissolved into DMSO to create 10 mM stock solutions . The flies used for this study , w; Sp/Cyo; Sb/TM6B , were reared on fly food consisting of 0 . 5% agar , 7% molasses , 6% cornmeal , and 0 . 8% killed yeast . Newly eclosed flies were collected , reared for 3 days , starved one day , and then fed compounds of interest for one day . For titer assessment experiments , compounds were diluted to a final concentration of 100 uM in yeast paste . Equivalent amounts of carrier DMSO diluted into these nutrient sources were used as a control . Brugia microfilariae were provided by TRS ( Athens , Georgia ) . These were resuspended in 2 mL tissue culture media containing 50 uM of each compound of interest , except for colchicine , which was administered at 20 uM . Control worms were provided an equivalent dilution of carrier DMSO alone . Microfilariae were incubated with the compounds for 1 day at 37 C with 7 . 5% CO2 . Adult Brugia were incubated for 3 days . Tissue culture chamber slides were coated with 0 . 5 mg/mL Concanavalin A , followed by addition of JW18 cells . After a 24-hour incubation at 25°C , cells were exposed to Syto-11 ( Molecular Probes ) for one minute at a dilution of 1∶50 , 000 . Cells were imaged on a Leica SP2 confocal microscope at 100× magnification and 2 . 75× optical zoom using FITC filters . Images were acquired at 5-second intervals for up to 10 minutes . A combination of fixation and staining methods was used . Propidium iodide and microtubule staining of Drosophila ovarian tissue was also done using established methods [54] , [55] , [76] , [77] . Brugia staining was performed as previously described [121] . Briefly , worms were sectioned and fixed in PFA 3 . 2% for 10 minutes , rinsed in PBS+0 . 1% Triton-X100 , and incubated overnight in RNAseA ( 10 mg/mL ) in PBST , prior a 30 second incubation in a propidium iodide solution ( 1 mg/mL diluted100X in PBST ) . Worm fragments were washed for 1 minute in PBST and mounted in DAPI Vectashield mounting medium ( Vector Labs ) . For FtsZ staining , worm fragments were incubated with rabbit anti-FtsZ [78] in PBST after a 1∶500 dilution , after RNAse treatment , washed three times for 10 minutes , before adding a CY5-conjugated anti-rabbit secondary antibody , followed by 3 washes of 10 minutes in PBST and mounting in DAPI Vectashield mounting medium . Microtubules were stained using the monoclonal antibody DM1α ( Sigma ) at 1∶250 and an Alexa488-conjugated goat anti-mouse secondary antibody ( Molecular Probes ) was used at 1∶250 . Mouse anti-histone H1 ( Millipore MAB052 ) was used at 1∶500 , and rabbit anti-phospho-histone H3 ( Millipore ) was used at 1∶1000 . Data collection was conducted as previously . All tissues were imaged using Leica SP2 and Leica SP5 confocal microscopes . Wolbachia were quantified in single focal planes of stage 10A Drosophila oocytes using established methods [67] . To measure Wolbachia length in the Brugia ovaries , 11 images representing a 2 micrometer-thick Z-stack were merged to make a single image , followed by assessment of bacterial length using the Leica SP2 line quantification function . Average Wolbachia titer values associated with drug treatments were normalized against their respective DMSO controls as previously to ensure comparability between experiments [67] . Statistical analysis was conducted using established methods . The ANOVA function in SPSS was used to evaluate Wolbachia titer differences in Drosophila oogenesis and differences in Wolbachia length in B . malayi . Chi-square tests were used to compare the frequency of Wolbachia posterior localization in Drosophila oogenesis and to evaluate ratios of Wolbachia per host nucleus in the Brugia ovary .
|
Wolbachia-based neglected diseases currently threaten one sixth of the world population . Millions of people are infected with filarial nematodes that rely upon endosymbiotic Wolbachia bacteria for their survival . These Wolbachia ultimately induce an immune response that gives rise to African river blindness or lymphatic filariasis . Thus , targeting Wolbachia will prevent induction of disease symptoms while also eliminating the filarial nematode infection . To identify new , fast-acting anti-Wolbachia drugs , we tested candidate compounds in Wolbachia-infected insect cells using automated robotics . Some of the anti-Wolbachia compounds we discovered closely resembled Albendazole , an FDA-approved drug already used worldwide to combat filarial nematode infections . We found that a common Albendazole metabolite strongly suppresses Wolbachia density in fruit flies and filarial nematodes by disrupting Wolbachia growth . These findings suggest that Albendazole is effective in treating filarial nematode-based diseases because it independently targets both the nematode and its essential Wolbachia endosymbionts . This has immediate implications for treating lymphatic filariasis and African river blindness .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"drugs",
"and",
"devices",
"biology",
"microbiology",
"molecular",
"cell",
"biology"
] |
2012
|
A Cell-Based Screen Reveals that the Albendazole Metabolite, Albendazole Sulfone, Targets Wolbachia
|
In the sea urchin embryo , specification of the dorsal-ventral axis critically relies on the spatially restricted expression of nodal in the presumptive ventral ectoderm . The ventral restriction of nodal expression requires the activity of the maternal TGF-β ligand Panda but the mechanism by which Panda restricts nodal expression is unknown . Similarly , what initiates expression of nodal in the ectoderm and what are the mechanisms that link patterning along the primary and secondary axes is not well understood . We report that in Paracentrotus lividus , the activity of the maternally expressed ETS-domain transcription factor Yan/Tel is essential for the spatial restriction of nodal . Inhibiting translation of maternal yan/tel mRNA disrupted dorsal-ventral patterning in all germ layers by causing a massive ectopic expression of nodal starting from cleavage stages , mimicking the phenotype caused by inactivation of the maternal Nodal antagonist Panda . We show that like in the fly or in vertebrates , the activity of sea urchin Yan/Tel is regulated by phosphorylation by MAP kinases . However , unlike in the fly or in vertebrates , phosphorylation by GSK3 plays a central role in the regulation Yan/Tel stability in the sea urchin . We show that GSK3 phosphorylates Yan/Tel in vitro at two different sites including a β-TRCP ubiquitin ligase degradation motif and a C-terminal Ser/Thr rich cluster and that phosphorylation of Yan/Tel by GSK3 triggers its degradation by a β-TRCP/proteasome pathway . Finally , we show that , Yan is epistatic to Panda and that the activity of Yan/Tel is required downstream of Panda to restrict nodal expression . Our results identify Yan/Tel as a central regulator of the spatial expression of nodal in Paracentrotus lividus and uncover a key interaction between the gene regulatory networks responsible for patterning the embryo along the dorsal-ventral and animal-vegetal axes .
In the sea urchin embryo , specification of the dorsal-ventral ( D/V ) axis critically relies on zygotic expression of the gene encoding the TGF-β family member Nodal in the presumptive ventral ectoderm [1] . Nodal signaling promotes specification of the ventral ectoderm and triggers expression of BMP2/4 , which acts as a long-range morphogen that specifies and patterns the dorsal region [2] . Sea urchin embryos lacking Nodal function develop with a strongly radialized phenotype and fail to express both ventral and dorsal marker genes . However , injection of nodal mRNA into a single cell can rescue a complete dorsal-ventral axis in these nodal morpholino injected embryos [1] . Therefore , expression of nodal is a key event that launches the gene regulatory network that specifies the D/V axis and understanding how nodal expression is regulated is essential to understand how secondary axis specification is established [3] . nodal expression is initiated around the 32/64 cell-stage and is initially rather broad , encompassing most cells of the presumptive ectoderm . Starting at the early blastula stage , nodal expression is progressively restricted to a smaller domain that corresponds to the presumptive ventral ectoderm [4] . The progressive restriction of nodal expression in the sea urchin embryo is thought to rely on the ability of Nodal to stimulate its own expression as well as that of the lefty gene , which encodes a diffusible long-range inhibitor of Nodal signaling [5–8] . Together , the auto-activation of Nodal and the long-range inhibition of Lefty constitute the basis for a reaction-diffusion mechanism , i . e . a system with self-organizing properties , able to amplify weak initial anisotropies and to generate sharp boundaries within a field of cells [9–14] . In the absence of Lefty , nodal expression is not restricted ventrally and expands towards the dorsal ectoderm [14] . Although this reaction-diffusion mechanism is clearly essential for the spatial restriction of nodal , Lefty is not the first factor acting in the process leading to the progressive restriction of nodal expression . The spatial restriction of nodal was recently shown to critically rely on the activity of a maternal TGF-ß ligand named Panda [15] . When maternal but not zygotic panda function is blocked , nodal is ectopically expressed from the very beginning of its expression until late in gastrulation despite the presence of Lefty , indicating that the reaction-diffusion mechanism between Nodal and Lefty is not sufficient to break the initial radial symmetry of the embryo . However , panda morphants nevertheless recover a dorsal-ventral axis late in gastrulation due to compensatory zygotic mechanisms . Maternal panda transcripts appear to be asymmetrically distributed in the egg and indeed , local overexpression of panda into one blastomere , but not global overexpression of panda into the egg , rescues the D/V polarity of panda morphants . The mechanism by which Panda restricts nodal expression is not elucidated but an intriguing hypothesis is that Panda may act by down-regulating the Nodal amplification loop on the dorsal side [15] . In addition to Panda and Lefty , the activity of the maternal Hypoxia induced transcription factor HIF-1 alpha has been proposed to contribute to the spatial restriction of nodal in S . purpuratus . However , HIF-1 alpha does not appear to be a crucial factor to establish the spatial expression of nodal and embryos lacking HIF-1 alpha only show a transient increase in nodal expression at early blastula stage but express nodal in a spatially restricted manner thereafter and develop with a normal dorsal-ventral axis [16] . Intriguingly , the absence of Lefty or of Panda does not cause ectopic expression of nodal in other territories that normally do not express nodal such as the animal pole region and the endomesoderm indicating that , in addition to Lefty and Panda , other factors likely contribute to the repression of nodal expression in these domains . The transcription factor FoxQ2 has been identified recently as one such factor that acts redundantly with Lefty to prevent nodal expression in the animal pole [17 , 18] . foxQ2 transcripts are expressed in the animal half of the embryo during cleavage but are progressively restricted to the animal pole region by a Wnt/β-Catenin-dependent process [18 , 19] . Interfering with foxQ2 mRNA translation does not perturb nodal expression or D/V axis formation , but disrupts specification of animal pole region and the resulting embryos lack the animal most territory called the apical tuft . In the double lefty+foxQ2 morphants , nodal expression expands to the animal pole region suggesting that Lefty and FoxQ2 act redundantly to repress nodal expression in the animal pole region . Conversely , nodal is not expressed in embryos with an expanded animal pole that retain foxQ2 expression such as in embryos animalized by inhibition of vegetal signaling . However , preventing translation of the foxQ2 transcript in these embryos is sufficient to rescue nodal expression and to restore normal patterning of the ectoderm . On the basis of these observations , it was suggested that in the unperturbed embryo , FoxQ2 links D/V axis formation to animal-vegetal patterning through its repressive action on nodal or in other words , that the vegetal signaling initiated by canonical Wnt/β-Catenin plays a permissive role in D/V axis formation by releasing the repressive action of FoxQ2 on nodal [17 , 18] . However , as mentioned above , inhibition of foxQ2 mRNA translation has no effect on the timing or the spatial regulation of nodal expression suggesting that FoxQ2 is not essential for the timing or the spatial expression of nodal in the unperturbed embryo . Consistent with this idea , FoxQ2 protein is not detectable in the early embryo and only appears in the nuclei of animal pole cells of embryos at late blastula stage , i . e . after the establishment of the spatial expression of nodal . Therefore , how the spatial expression of nodal is initiated and how nodal expression along the D/V axis is coordinated with patterning along the animal-vegetal axis remains not well understood . In this study , we report that the ETS domain transcriptional repressor Yan/Tel is an essential maternal regulator of nodal expression that links animal-vegetal signaling to the spatial expression of nodal during dorsal-ventral axis formation in Paracentrotus lividus . Furthermore , we identify the kinase GSK3 as a key component of the animal-vegetal patterning system that regulates the stability of Yan/Tel by triggering its degradation by a β-TRCP proteasome pathway . Finally , we show that Yan/Tel is epistatic to Panda and that Panda overexpression affects the phosphorylation state of Yan/Tel strongly suggesting that Panda contributes to the stabilization/regulation of Yan/Tel to regulate nodal expression . These results identify Yan/Tel as a new central regulator of nodal acting downstream of Panda and show that degradation of Yan/Tel by GSK3-mediated phosphorylation in the animal hemisphere is the key event that allows nodal expression and links patterning along the primary and secondary axes .
In the course of a cis-regulatory analysis of the nodal promoter , we noticed that , in addition to bZIP , Sox , Oct , homeobox and Smad binding sites , the 5' proximal module of the nodal promoter , called the R2 module , contains several conserved 5'-GGAA/T-3' ETS binding motifs , the function of which had not been analyzed previously ( Fig 1A ) [4] . To test the role of these sites in the transcriptional activation of nodal , we injected luciferase reporter constructs into fertilized sea urchin eggs and compared the activity of the wild-type construct with that of a mutant reporter lacking these ETS sites at blastula stage . Surprisingly , mutation of the 5 ETS sites contained in the 5' module did not reduce but instead significantly increased the activity of the nodal 5' proximal module , which reached up to 2 fold its normal value ( Fig 1B ) . This suggested that members of the ETS family with transcriptional repressor activity negatively control nodal expression by binding to the proximal module of its promoter . The sea urchin genome contains eleven genes encoding ETS transcription factors and some of them are presumed to be transcriptional repressors [20 , 21] . We focused on the gene encoding the ETS transcription factor Yan/Tel since this gene is known to encode a prototypical repressor [22–26] ( reviewed in [27] ) . The overall structure of sea urchin Yan/Tel protein is conserved with an N-terminal SAM domain , a central co-repressor binding domain and a C-terminal ETS DNA binding domain ( S1A Fig ) . The central domain is thought to be required for the recruitment of the co-repressors mSin3A , SMRT and N-CoR , which in turn can recruit histone deacetylases [28–30] . In addition , the N-terminal SAM domain of Yan and Tel oligomerizes in a head to tail manner and this polymerization has been shown to be essential for repression [31–33] . Interestingly , most of the hydrophobic residues that make up the interfaces between two monomers of Yan from Drosophila or Tel from vertebrates are conserved in the sea urchin Yan/Tel protein ( S1B Fig ) , suggesting that sea urchin Yan/Tel may also form higher order polymers as do its vertebrate and fly orthologues . Tel has been shown to be sumoylated by UBC9 on Lysine 11 and 99 and this modification has been shown to downregulate Tel and to promote its nuclear export . One of these residues , lysine 99 , is conserved in the sea urchin Yan/Tel suggesting that it may also be regulated by sumoylation ( S1 Fig ) . Yan and Tel have also been shown to be degraded primarily following ubiquitination by the ubiquitin conjugating enzyme FBL6 . However , the Ubiquitin acceptor sites of Tel have not been identified with certainty . Finally , both Yan and Tel have been shown to be phosphorylated by MAP kinases at multiple residues and this phosphorylation has been shown to downregulate Yan and Tel by promoting degradation and/or reducing DNA binding [22–26] . Intriguingly , sea urchin Yan/Tel also contains three canonical PXS/TP phosphorylation sites for MAP kinases but the positions of these sites do not appear to be conserved between the sea urchin , vertebrate and Drosophila proteins ( S1C Fig ) . Northern blot analysis and in situ hybridization revealed that yan/tel is a maternal transcript expressed abundantly and ubiquitously during cleavage ( Fig 1C and S2 Fig ) . Starting late in cleavage and during prehatching and hatching blastula stages , zygotic yan/tel transcripts were detected in a small ring of cells located in the vegetal pole region that likely corresponds to precursors of the skeletogenic mesoderm ( arrowheads in Fig 1C ) . To determine if Yan/Tel is the factor responsible for repressing the activity of the nodal promoter , we tested the effects on nodal expression of injecting antisense morpholino oligonucleotides against yan/tel . Embryos injected with morpholinos directed against the translation start site of the yan/tel transcript developed normally up to the late blastula stage . Then , while in control embryos precursors of the primary mesenchymal cells ( PMCs ) ingressed into the blastocoel , in yan/tel morphants , ingression of the PMCs was clearly delayed . At the early gastrula stage ( 24hpf ) , while in the control embryos D/V polarity was already apparent by the presence of bilateral clusters of PMCs ( arrowheads ) containing spicule rudiments and by the flattening of the presumptive ventral side , in the yan/tel morphants , no visible sign of dorsal-ventral polarity was apparent: the PMCs remained arranged in a ring around the archenteron ( arrowheads in Fig 2A ) and the embryos conserved a rounded shape . This apparent lack of dorsal-ventral polarity persisted at the late gastrula stage: when in control embryos the archenteron bent toward the presumptive ventral ectoderm , the yan/tel morphants remained radialized as evidenced by the straight position of the archenteron at the center of the blastocoel and by the radial arrangement of the PMCs . Surprisingly , despite this apparent complete failure to establish a D/V axis until late in gastrulation , the yan/tel morphants progressively recovered a D/V polarity and at the equivalent of the early prism stage ( 36 hpf ) , the first signs of dorsal-ventral polarity appeared in these embryos . A thickened ventral-like ectoderm started to differentiate on one half of the embryo and the archenteron adopted a slightly asymmetrical position closer to this thickened ectoderm . However , instead of forming bilateral clusters , the PMCs arranged themselves in an extended half-circle at the basis of the ventral ectoderm . Interestingly , the morphology of yan/tel morphants at this stage was similar to that of embryos partially ventralized by treatment with recombinant Nodal or by overexpression of low doses of nodal mRNA or to that of embryos lacking the function of the maternal factor Panda that is required to restrict nodal expression to the ventral side ( Fig 2A ) [15] . These strong effects on D/V axis formation were observed following injection of two different morpholino oligonucleotides targeting either the translation start site or the 5’ UTR and were completely suppressed by co-injection into the egg of a wild-type form of yan/tel mRNA immune against the morpholino ( S3 Fig ) . In contrast , injection of a morpholino oligonucleotide targeting a splice junction abrogated skeletogenesis ( S4 Fig ) , consistent with the zygotic expression of the gene in the PMCs , but did not affect establishment of the dorsal-ventral axis and did not prevent formation of the bilateral PMC clusters suggesting that the maternal function of Yan/Tel but not its zygotic function is required for establishment of the dorsal-ventral axis . To understand the origin of the radialized phenotype of yan/tel morphants , we examined the expression of nodal , chordin and bmp2/4 as well as of some of their downstream targets in the ectoderm such as 29D and tbx2/3 ( Fig 2B–2D ) . Strikingly , consistent with their radialized phenotype and with their resemblance to panda morphants , yan/tel morphants displayed a dramatic ectopic expression of nodal as well as of chordin , bmp2/4 and lefty , three direct downstream targets of Nodal signaling [34] . Interestingly , this ectopic expression was observed within a large ectodermal domain that encompassed the presumptive ciliary band and dorsal ectoderm , forming a large belt of cells around the embryo , but it did not extend to the animal pole region or to the endomesoderm territory ( Fig 2C ) . Also , consistent with the observed expansion of nodal expression , the expression of dorsal markers such as tbx2/3 and 29D was either eliminated or dramatically reduced in the yan/tel morphants ( Fig 2D ) . Dorsal-ventral patterning of the mesoderm was also strongly perturbed in the yan/tel morphants , which typically contained very few pigment cells . Indeed , the ventral expression of the blastocoelar cell marker gata1/2/3 was radialized while expression of the dorsally expressed pigment cell marker gcm was suppressed ( Fig 2E ) , in agreement with the radialized expression of nodal resulting from inhibition of yan/tel mRNA translation [35] . Since yan/tel morphants ectopically express nodal , we tested if local inhibition of yan/tel mRNA translation is sufficient to trigger nodal expression and to orient the dorsal-ventral axis . Since in Paracentrotus the orientation of the D/V axis is not related to the plane of first cleavage , we injected the yan/tel morpholino randomly into one blastomere of embryos at the 2 or 4 cell-stage and later scored the position of the clone of injected cells at pluteus stage . In most of the injected embryos ( n>50 ) , nodal expression was later found in a territory either congruent or overlapping with the clone of cells derived from the injected blastomere ( Fig 2F ) . Remarkably , the boundaries of the clone were always included in the nodal expressing territory and at least one of these boundaries always precisely aligned with the border of the nodal expression domain . Therefore , blocking translation of Yan/Tel mRNA caused the cell-autonomous expression of nodal , irrespective of the position of the clone . At pluteus stage , the progeny of the injected cells was always found on the ventral side of the larva . These results demonstrate that Yan/Tel is an essential regulator of nodal in vivo . They also strongly suggest a model of nodal activation by release of Yan/Tel-mediated transcriptional repression . Finally , since yan/tel morphants displayed a dramatic ectopic expression of nodal as well as of chordin , bmp2/4 and lefty , we tested if Yan/Tel exerts a global repression on several genes of the Nodal pathway . Treatment with the Nodal receptor inhibitor SB431542 blocked the ectopic expression of chordin observed in yan/tel morphants ( Fig 2G ) . Therefore , the ectopic expression of the Nodal downstream target genes like chordin following inhibition of yan/tel is most likely indirect and depends on Nodal pathway activity . nodal is likely the main gene of the Nodal pathway whose expression is directly regulated by Yan/Tel . Taken together these results show that in addition to the reaction-diffusion mechanism based on Lefty and to the activity of the maternal TGF-ß ligand Panda recently shown to be essential for the spatial restriction of nodal , the transcriptional repressor Yan/Tel plays a central role in dorsal-ventral patterning of the ectoderm and mesoderm by regulating the spatial expression of nodal . Previous studies showed that the spatial restriction of nodal expression relies on the early establishment of a reaction-diffusion mechanism between Nodal and Lefty starting at the early blastula stage [1 , 14 , 36] . To determine if the function of Yan/Tel is required early during cleavage stages for the establishment of the spatial restriction of nodal or if it is only necessary at later stages for its maintenance , we performed a time-course experiment and compared nodal expression in lefty morphants and yan/tel morphants . Embryos were injected with either the lefty morpholino or the yan/tel morpholino and the expression of nodal was examined at successive stages starting at 32-cell stage , when nodal expression is first detectable in a broad domain , up to pre-hatching blastula , when its expression is sharply restricted to the presumptive ventral ectoderm . About half of the lefty morphants embryos at 60-cell stage displayed a localized nodal expression indicating that the spatial restriction of nodal was initiated normally in these embryos ( Fig 3A ) . Nevertheless , in the absence of Lefty , the spatial restriction of nodal expression was not maintained at later stages and most embryos eventually showed a massive ectopic expression of nodal , as reported previously [14] . In contrast , in the yan/tel morpholino injected embryos , no restricted expression of nodal was visible at any stage and all the injected embryos showed a massive ectopic expression of nodal from early on ( Fig 3A ) . These results suggest that the function of Yan/Tel is required before that of Lefty for the establishment of the spatial restriction of nodal . Although its expression is not restricted along the D/V axis , Yan/Tel is required for the initial restriction of nodal expression while the function of Lefty appears to be required only later for the maintenance of the spatial restriction of this gene . Both in the unperturbed embryo and following overexpression of nodal , nodal expression is not detected in the animal pole [1 , 14 , 15 , 18 , 34 , 36–38] . The animal pole therefore appears as a territory refractory to nodal expression . It was shown previously that foxQ2 , which is expressed specifically in the animal pole , acts together with Lefty to repress nodal expression in this region [18] . Since Yan/Tel is a repressor of nodal expression and since it is expressed ubiquitously at early stages , we tested if its function is required to repress nodal expression in the animal pole domain . As shown previously [1 , 15 , 18] , embryos injected with morpholinos directed against foxQ2 or panda or lefty transcripts did not show ectopic expression of nodal in the animal pole domain ( Fig 3B ) . In contrast , in embryos co-injected with either the foxQ2 and yan/tel or with foxQ2 and panda morpholinos , nodal expression expanded dramatically to fill up the animal pole domain ( asterisks in Fig 3B ) . This suggests that in the foxQ2 morphants , nodal expression does not expand to the animal pole region because the activities of Yan/Tel and Panda prevent expansion of nodal in this region . It was shown previously that co-injection of foxQ2 and lefty morpholinos resulted in ectopic expression of nodal in the animal pole domain [18] . In contrast , embryos co-injected with yan/tel and lefty morpholinos or with panda and lefty morpholinos did not show an expansion of nodal expression in the animal pole region ( Fig 3B ) . This suggests that although Yan/Tel , Lefty and Panda cooperate with FoxQ2 to repress nodal expression in the animal pole , the contribution of these factors in the repression of nodal may not be as crucial as that exerted by FoxQ2 since only on the absence of FoxQ2 is nodal ectopically expressed in the animal pole region . Taken together these results show that Yan/Tel plays an important role as a repressor of nodal expression in the animal pole domain where it acts together with FoxQ2 , Lefty and Panda . While removing the function of either gene alone is not sufficient to cause ectopic expression of nodal in the animal pole region , removing the function of Yan/Tel and FoxQ2 triggers ectopic expression of nodal in this region . Overexpression of wild-type yan/tel mRNA had only moderate effects on development of sea urchin embryos ( Fig 4B ) . Even when injected with high doses ( 1000 μg/ml ) of this mRNA , the embryos developed into pluteus larvae and only occasionally , reduced D/V axis and ectopic skeletal elements were observed . Following overexpression of wt yan/tel mRNA , nodal expression was absent or strongly reduced during blastula stages in 30% of the embryos but it was progressively restored to normal levels during gastrula stages consistent with the apparent normal morphology of yan/tel overexpressing embryos at pluteus stages ( Fig 4B and 4C ) . This result is consistent with previous studies showing that the activity of Drosophila Yan or of vertebrate Tel is not regulated at the transcriptional level but at the post-transcriptional level by phosphorylation by MAP kinases . MAPK dependent phosphorylation of Yan and Tel is the central mechanism that regulates the activity and stability of these transcriptional repressors [22–26] ( reviewed in [27] ) . Specific phosphorylation events triggered by either ERK , JNK or p38 downregulate the transcriptional repressor function of Drosophila Yan or of vertebrate Tel , leading to their export out of the nucleus and to their degradation [22–26] ( reviewed in [27] ) . Mutations that convert the serine and threonines residues normally phosphorylated by MAPK into non-phosphorylatable residues transform these factors into constitutively active repressors while mutations that convert them into phospho-mimetic residues promote degradation of these factors . The sea urchin Yan/Tel protein contains 3 canonical consensus MAPK phosphorylation sites PXS/TP located between the Sam domain and the Ets domain and in the C-terminal region plus one additional LSTP site that has been shown to be phosphorylated efficiently by ERK [39] . To investigate the roles of these potential phosphorylation sites in the regulation of Yan/Tel transcriptional activity and or/stability in vivo , we constructed a mutant Yan/Tel , Yan/Tel4A ( Yan/Tel MAPK mutant ) , in which these putative MAPK phosphorylation sites were replaced by alanine ( Fig 4A ) . Replacing the four putative MAPK consensus phosphorylation sites by an alanine increased the ability of Yan/Tel to repress nodal at blastula stage ( absent in 58% of the embryos ) and caused a partial radialization of the embryos at 36h . However , these embryos progressively recovered a dorsal-ventral polarity as shown by their pluteus-like morphology at 72 h . We then examined carefully the sequence of Yan/Tel and noticed that in addition to MAPK sites , sea urchin Yan/Tel contains two motifs that are potentially phosphorylatable by the Proline-directed kinase GSK3 ( Fig 4A ) . The first GSK3 phosphorylation motif conforms to the well-defined β-TRCP destruction box DSGXXS . β-TRCP is a F-box subunit of the SCF ( SKP1-CUL-F-box ) complex E3 ubiquitin ligase that targets several proteins for polyubiquitinylation and proteasomal degradation including Snail , β-Catenin , Emi and IKβ [40–43] . β-TRCP recognizes the doubly phosphorylated DpSGXXpS motif of target proteins and adds ubiquitin to a proximal upstream lysine . The second motif of Yan/Tel potentially phosphorylatable by GSK3 , referred below as the Ser/Thr cluster , SSSTTPSPSPP , lies at position 669–678 , near the C terminal end of the protein , and may conform to the requirement of GSK3 for a phosphorylated Ser/Thr residue at position p+4 of the substrate . We therefore constructed a series of mutants carrying mutations in these sites alone or in combinations ( Fig 4A ) . The β-TRCP mutant , DAGHSA , bears mutations in the two residues presumed to be phosphorylated by GSK3 . The Yan/Tel “cluster” mutant , ASPTAPAPAP contains 4 mutations in the Ser/Thr -rich region of Yan/Tel . The Yan/Tel8A mutant corresponds to combinations of MAPK sites mutations and Cluster mutations while the Yan/Tel 10A mutant bears mutations in the MAPK , cluster and β-TRCP motifs . Finally , the Yan/Tel13A mutant carries mutations in the MAPK , cluster and β-TRCP motifs plus 3 additional mutations in SP sites that can potentially be phosphorylated by Proline directed kinases ( Fig 4A ) ( see materials and methods ) . These phosphorylation mutants were overexpressed at the same concentration as the MAPK mutant and the effects on the morphology of the embryos and on the expression of marker genes were analyzed . Like mutation of the putative MAPK sites , mutation of the β-TRCP or cluster motifs affected the activity of Yan/Tel and caused a partial radialization suggesting that these sites participate to the regulation of the activity/stability of Yan/Tel ( Fig 4B ) . However , none of these mutants alone was sufficient to repress nodal expression efficiently and at 72 hpf , most the injected larvae had recovered a normal pluteus morphology . Similarly , overexpression of Yan/Tel8A ( mutations in the cluster+MAPK sites ) caused a partial radialization and resulted in an incomplete loss of nodal expression ( absent in 50% of the embryos ) . Injection of mRNA encoding Yan/Tel 10A and 13A caused the strongest effects . They suppressed nodal expression in most embryos ( 91% and 85% of the embryos , respectively ) and abolished dorsal-ventral axis formation ( Fig 4B and 4C and see below ) . These observations suggest that mutations of the MAPK , cluster region and β-TRCP phosphorylation sites have synergistic effects on the activity/stability of Yan/Tel and mutations of all three types of sites are required to convert Yan/Tel into a constitutively active repressor . To analyze the effects of mutations in the MAPK , β-TRCP and cluster phosphorylation sites on the stability of Yan/Tel , we constructed tagged versions of wild-type and mutant Yan/Tel carrying three HA epitopes tag on the N-terminus . Having established that the presence of the HA tag did not alter the activity of the wild type Yan/Tel protein , we compared the effects of the various mutations on its phosphorylation state and stability . Wild type Yan/Tel typically migrated as multiple bands following SDS-PAGE with three to four slow migrating isoforms and a fourth barely detectable faster migrating isoform ( see Fig 5A ) . Mutation of the β-TRCP degradation motif did not change significantly the migration of Yan/Tel but it resulted in a marked stabilization of the protein as judged by the much stronger intensity of the band of Yan/Tel β-TRCP mutant compared to wild type ( Fig 4D ) . In contrast , replacing the four putative MAPK consensus phosphorylation sites by an alanine changed dramatically the mobility of Yan/Tel , which then migrated predominantly as a fast migrating isoform . This observation suggests that the slow migrating isoforms of Yan/Tel are isoforms phosphorylated by MAPKs . Mutation of the Ser/Thr cluster also resulted in a significant stabilization of the protein but interestingly , in this case , a fast migrating isoform was also predominant . Finally , a mutant form of Yan/Tel carrying mutations of the MAPK , Ser/Thr cluster and β-TRCP motifs showed a strong stabilization and migrated , like the cluster mutant , predominantly as a fast migrating , presumably unphosphorylated isoform ( Fig 4D ) . Taken together these results show that combining mutations in the MAPK sites , Ser/Thr cluster and β-TRCP motif produce synergistic effects on the stability of Yan/Tel and that phosphorylation on several sites of Yan/Tel likely regulates its stability and/or activity . Finally , we tested if random injection of yan/tel13A mRNA into one blastomere at the two-cell stage is sufficient to orient the dorsal-ventral axis . Indeed , in all the injected embryos ( n = 15 ) , nodal expression was found in a discrete region on the opposite side of the cells expressing yan/tel13A and the progeny of the injected blastomere later occupied a territory that precisely coincided with the dorsal ectoderm ( Fig 4E ) . Taken together these observations show that Yan/Tel acts as a major negative regulator of nodal expression . The stability of Yan/Tel is itself regulated by GSK3 , which is active in the ectoderm and which targets Yan/Tel for degradation . Therefore , Yan/Tel may act at the crossroads of both animal-vegetal and dorsal-ventral patterning . Treatments with the MEK inhibitor U0126 , or with the p38 inhibitor BIRB796 or with the JNK pathway inhibitor SP600125 had modest effects on the stability of Yan/Tel , although we noticed that the slowest migrating isoform of Yan was consistently reduced or missing following treatment with either of these inhibitors ( Fig 5A ) . Strikingly , when used in combination , U0126 , BIRB-796 and SP600125 increased the stability of Yan/Tel which then migrated predominantly as a fast migrating isoform . Western blot analysis confirmed that ERK , p38 and JNK are active in the early embryos ( Fig 5C ) . Furthermore , immunostaining experiments revealed that , in addition to being present in the skeletogenic mesoderm located at the vegetal pole , activated forms of ERK are also present in the ectoderm at early blastula stage while p38 is more broadly active and detected in most nuclei with a shallow dorsal-ventral gradient ( Fig 5D ) [38 , 44] . However , treatments from fertilization on with inhibitors of ERK , p38 or JNK reduced but did not suppress nodal expression ( Fig 5B ) , the strongest effect being observed in the case of JNK inhibition . These observations are consistent with the idea that MAPK contribute to the regulation of nodal expression by phosphorylating and destabilizing Yan/Tel but that additional kinases are likely involved in the regulation of the activity/stability of this transcription factor . Finally , treatments with MAPK inhibitors alone or in combination did not suppress the ectopic expression of nodal observed in Yan/Tel morphants consistent with the idea that MAPK act upstream of Yan/Tel ( Fig 5E ) . A plethora of studies have documented that Activin , Nodal and TGF-ß signaling secondarily induce MAPK signaling [45–48] . We therefore tested if Nodal signaling impacts on phosphorylation of Yan/Tel by specific MAPK ( Fig 5F ) . Overexpression of nodal mRNA , overexpression of a constitutively active Nodal receptor or treatment with nickel did not dramatically alter the stability of Yan/Tel but we noticed that they reproducibly promoted formation of the slowest migrating ( presumably phosphorylated ) isoforms of Yan/Tel ( Fig 5F and 5G ) . All three MAPKs were apparently required for this effect ( Fig 5F ) . In contrast treatment with the Nodal receptor inhibitor SB431542 appeared to favor formation of the faster migrating isoform of Yan/Tel , which likely represents the unphosphorylated isoform ( Fig 5G ) . Although Nodal signaling did not appear to affect the overall stability of Yan/Tel , it did change its phosphorylation state . It is therefore possible that Nodal signaling impacts on MAPK signaling which in turn modifies the activity of Yan/Tel . These results are consistent with a model in which Nodal signaling acting through MAPK signaling , indirectly promotes phosphorylation and downregulation of Yan/Tel , thereby reinforcing nodal expression . Since Yan/Tel contains a β-TRCP destruction motif ( Fig 6A and 6B ) that could potentially be phosphorylated by the Proline directed kinase GSK3 , we tested the effects of blocking GSK3 on the activity/stability of Yan/Tel . Treatment of sea urchin embryos with increasing concentrations of lithium , a potent GSK3 inhibitor , resulted in a marked and dose-dependent increase in the stability of HA-tagged Yan/Tel ( Fig 6C ) . Co-treatment with lithium and MG132 , a proteasome inhibitor , caused a synergistic accumulation of Yan/Tel . Remarkably , treatment with lithium also dramatically altered the migration pattern of sea urchin Yan/Tel protein . While in control embryos the Yan/Tel protein migrated as four distinct isoforms on an SDS-PAGE gel , following lithium treatment Yan/Tel migrated predominantly as the faster migrating isoform ( Fig 6C ) , which we presume is a non-phosphorylated form of this factor . Similarly , overexpression of β-TRCP mRNA resulted in Yan/Tel protein migrating predominantly as a fast migrating non-phosphorylated isoform consistent with the idea that the slowest migrating isoforms of Yan/Tel are isoforms phosphorylated on the β-TRCP motif that are recognized and degraded by the ubiquitin ligase β-TRCP ( Fig 6D ) . To confirm that GSK3 regulates the stability of Yan/Tel , we overexpressed sea urchin GSK3β with either wild type Yan/Tel or with the Yan/Tel10A mutant which lacks the MAPK , β-TRCP and cluster phosphorylation motifs ( Fig 6C ) . Overexpression of GSK3 significantly reduced the level of wild type Yan/Tel protein but had no detectable effect on the stability of Yan 10A reinforcing the idea that GSK3 phosphorylates Yan/Tel and targets it for degradation . In agreement with the western blot analysis , which indicated that inhibition of GSK3 stabilizes Yan/Tel , a 3h treatment with lithium performed at early blastula stage ( Fig 6F ) , when GSK3 is active ( Fig 6E ) , was sufficient to extinguish nodal expression ( Fig 6F ) . This extinction of nodal caused by lithium occurred without any concomitant expansion of foxA expression indicating that it was not caused by an indirect effect such as a change in specification of the ectoderm but more likely by a more direct effect of inhibition of GSK3 on Yan/Tel stability . Finally , treatments with lithium did not block the ectopic expression of nodal observed in Yan/Tel morphants consistent with the idea that GSK3 acts upstream of Yan/Tel ( Fig 6G ) . We conclude that phosphorylation by GSK3 and MAPK followed by proteasome-mediated proteolysis are key processes that regulate the stability and/or activity of Yan/Tel . However , it should be noted that the activity of the overexpressed Yan/Tel mutants as measured by their impact on nodal expression ( Fig 4C ) do not correlate perfectly with the effects of these mutations on the stability of Yan/Tel , with GSK3 preferentially affecting the stability and MAPKs affecting more the activity of Yan/Tel ( Fig 5A; Fig 6C ) . This highlights the complex nature of the regulation of Yan/Tel by these kinases . We have shown that sea urchin Yan/Tel contains at least two consensus sites for phosphorylation by GSK3 , that inhibition of GSK3 affects the stability of Yan/Tel and that mutation of these sites changes the activity of the protein . However , evidence that GSK3 can phosphorylate these sites was still missing . To further examine if sea urchin Yan/Tel can serve as a substrate for GSK3 , we set up an in vitro phosphorylation assay to detect phosphorylation of Yan/Tel at Ser 195 and Ser 198 . Since no phospho-specific antibody recognizing the phosphorylated motif of sea urchin Yan/Tel was available we used an alternative antibody . We reasoned first that the β-TRCP recognition motif of sea urchin Yan/Tel ( DSGHSSPDT ) shares similarity with the β-TRCP motif of human β-catenin DSGIHSGATT ( Fig 6H ) . Therefore , an antibody against the doubly phosphorylated DpSGIHpS motif of human βcatenin may recognize the DpSGHSpS motif of sea urchin Yan/Tel . We therefore produced and purified a glutathione–S-transferase ( GST ) -fused Yan/Tel peptide corresponding to the β-TRCP motif of the sea urchin protein . We used as control a GST-fused β-catenin peptide carrying the β-TRCP recognition motif of human β-catenin . These peptides were used as substrates in a kinase reaction then the reaction products were separated by SDS PAGE , blotted and immunodetection was carried out with the anti-phospho ( Ser33/Ser37/Thr41 ) -human β-catenin . Since GSK3 requires a priming phosphorylation by GSK3 to phosphorylate β-catenin , we performed kinase reactions either without or with purified Casein kinase alone , GSK3 alone or with the two kinases in combination . Western blot analysis revealed that the anti phospho-ß catenin antibody indeed recognizes the β-TRCP motif phosphorylated by GSK3 ( Fig 6H ) . However , unlike β-catenin , which requires pre-phosphorylation by Casein kinase , sea urchin Yan/Tel does not require Casein kinase to be phosphorylated by GSK3 . Intriguingly , the addition of Casein kinase actually interfered with phosphorylation of the peptide carrying the sea urchin Yan/Tel β-TRCP degradation motif . No phosphorylation was detected in control reactions in which the substrate was the non phosphorylatable Alanine mutant β-TRCP degradation box DAGHSAPDT indicating that Yan/Tel is phosphorylated most likely directly by GSK3 at positions 195 and 198 ( Fig 6H ) . Finally , we also speculated that the antibody against the triply phosphorylated ß-catenin ( Ser33/Ser37/Thr41 ) may recognize the triply phosphorylated Ser/Thr cluster of sea urchin Yan/Tel SSSTTPSPSPP . Indeed , western blot analysis detected a band that corresponded to the Ser/Thr rich cluster of Yan/Tel phosphorylated by GSK3 ( Fig 6H ) . This band was not detected when a non-phosphorylatable Alanine mutant Ser/Thr cluster peptide was used . To summarize , we have shown that the β-TRCP destruction box and the Ser/Thr rich cluster of sea urchin Yan/Tel are consensus GSK3 sites , that inhibition of GSK3 alters the stability of Yan/Tel , that mutation of these sites changes the stability and/or the activity of Yan/Tel and that these sites can be phosphorylated by purified GSK3 further reinforcing the conclusion that GSK3 is a key regulator of the stability and activity of sea urchin Yan/Tel . The similarity of the phenotypes caused by inactivation of Panda and Yan/Tel strongly suggested the possibility that Yan/Tel may act downstream of Panda . According to this model , Panda signaling may regulate nodal expression by inhibiting phosphorylation of Yan/Tel and stabilizing this factor . The fact that both Yan/Tel and Panda are required early to restrict nodal expression supports this idea ( Fig 3A see Haillot et al . 2015 ) . Furthermore , not only inactivation of either Yan/Tel or Panda causes ectopic expression of nodal but they both appear to do so in a cell-autonomous manner . To test if Panda and Yan Tel act in the same pathway and if Panda acts through stabilization of Yan/Tel , we tested the ability of Panda to orient the D/V axis in the absence of Yan/Tel and reciprocally , the ability of Yan/Tel to orient the axis in the absence of Panda . Fertilized eggs were injected either with the Yan/Tel or Panda morpholinos and at the 2-cell stage , panda mRNA or constitutively active yan/tel mRNA was injected into one blastomere at the 2-cell stage . nodal expression was then analyzed at early blastula stage and the position of the territory derived from the injected clone was later recorded at prism stage ( Fig 7A ) . In Yan morphants at prism stage , the progeny of the clones injected with Panda mRNA was found predominantly on the dorsal side indicating that Panda can orient the D/V axis of Yan morphants . However , while in control embryos local overexpression of panda mRNA efficiently oriented the D/V axis and efficiently restricted nodal expression at early blastula stage , strikingly , in yan/tel morphants injected with Panda mRNA , nodal remained expressed ubiquitously at early blastula stage ( Fig 7B ) . That Yan morphants injected with Panda mRNA into one blastomere failed to restrict nodal expression at early stages strongly suggests that Yan is epistatic to Panda , or in other words , that Yan acts downstream of Panda . Consistent with this idea , local overexpression of activated yan/tel efficiently restricted nodal expression and oriented the D/V axis in panda morphants ( Fig 7A and 7B ) . To further test if Yan and Panda act in the same pathway during D/V axis formation , we looked for synergistic effects . Injection of suboptimal doses of either the Yan or the Panda morpholino resulted in a spatially restricted expression of the nodal target gene chordin at gastrula stage . In contrast , coinjection of the Yan and Panda morpholinos at these suboptimal doses caused a dramatic expansion of chordin expression and resulted in a strong ventralization ( Fig 7D ) . These results strongly suggest that Yan/Tel and Panda likely work in the same pathway and synergize to restrict nodal expression . Finally , we tested if overexpression of Panda affects the phosphorylation of Yan/Tel by western blot . Overexpression of Panda caused Yan/Tel to migrate predominantly as a fast migrating ( presumably stabilized ) isoform , consistent with the idea that Panda acts at least in part by regulating the phosphorylation state and therefore the stability and/or the activity of the transcription factor Yan/Tel ( Fig 7C ) . In conclusion , we have identified Yan/Tel as a new central negative regulator of the early expression of nodal and possibly as a downstream effector of the maternal determinant Panda . Taken together these data suggest that the stability of Yan/Tel is regulated by the combined activities of MAP Kinases and GSK3 , and possibly of Nodal itself , and that these factors cooperate to target this repressor for degradation by a β-TRCP/proteasome degradation pathway . Finally , our data suggest a model in which Panda acts upstream of Yan/Tel to restrict nodal expression , possibly by antagonizing phosphorylation and degradation of Yan/Tel on the dorsal side . By integrating information along the animal-vegetal and dorsal ventral axis , Yan/Tel may therefore act as a factor that coordinates patterning along these two orthogonal patterning systems .
Recent studies on the promoter region of the sea urchin nodal gene have started to identify maternal transcription factors required for high level of nodal expression [4 , 49 , 50] . However , while these studies successfully identified binding sites for positive regulators of nodal expression including SoxB1 , bZIP , and Oct factors , they failed to identify any spatial regulator of nodal expression . In this study , we have identified the ETS containing transcriptional repressor Yan/Tel as a key spatial regulator of nodal expression and as a target of MAP kinases and GSK3 in Paracentrotus lividus embryos . We found that inactivation of Yan/Tel causes ectopic expression of nodal , a phenotype largely similar to that caused by the loss of the maternal determinant Panda . Indeed , we discovered that Yan/Tel is epistatic to Panda strongly suggesting that Panda acts at least in part by stabilizing Yan/Tel on the dorsal side . Furthermore , we discovered that the activity and/or stability of Yan/Tel is negatively regulated by the combined activities of MAP kinases and GSK3 and possibly of Nodal itself . Taken together , these findings shed light on the mechanisms regulating the early expression of nodal at the onset of development and suggest that Yan/Tel acts as an integrator of patterning signals along the animal-vegetal and dorsal-ventral axes . The key observation at the basis of this study was that downregulation of maternal but not zygotic Yan/Tel function by injection of morpholino oligonucleotides caused a massive but transient ectopic expression of nodal in the dorsal ectoderm , largely mimicking the effects of inactivating the maternal factor Panda and partially mimicking the effects of inactivating Lefty . The maternal TGF-β ligand Panda was recently proposed to control dorsal-ventral axis formation by restricting the early expression of nodal [15] . Intriguingly , the phenotypes resulting from inactivation of maternal Panda function are very similar to those resulting from inactivation of Yan/Tel strongly suggesting that both genes may work in the same pathway . In both Panda and Yan/Tel morphants , nodal is massively ectopically expressed throughout most of the ectoderm resulting in full radialization of the ectoderm and endomesoderm and causing ectopic expression of ventral marker genes during gastrulation . Although the phenotypes of yan/tel and panda morphants are strikingly similar , there are however important differences between the two factors . In the case of Panda , in situ hybridization revealed a D/V asymmetry of maternally deposited panda mRNA and rescue experiments confirmed that the activity of Panda is likely spatially restricted in the embryo . Therefore , Panda clearly has some properties of a maternal determinant of the D/V axis . In contrast the yan/tel maternal RNA is uniformly distributed in the egg and early embryo and rescue experiment indicate that yan/tel mRNA does not need to be spatially restricted to rescue the D/V axis of yan/tel morphants . According to one scenario , Panda and Yan/Tel may act independently , i . e . in parallel , to restrict nodal expression . Panda may first act to break the radial symmetry and to restrict nodal expression and the MAP kinase pathways and Yan/Tel may then act to maintain this asymmetry . Alternatively , Panda and Yan/Tel may act in a coordinated manner in the same pathway , as suggested by the similarity of the phenotypes caused by inactivation of either gene . According to this hypothesis , Panda may provide the initial spatial cue that restricts nodal by downregulating the activity of MAPK , thereby reinforcing the repressive effect of Yan/Tel on nodal on the presumptive dorsal side . In this study , we have provided several lines of evidence suggesting that Panda and Yan/Tel indeed likely function in the same pathway , First , we have shown that inactivation of Yan or Panda by morpholino injection caused a strikingly similar phenotype . Second , we have shown that Panda requires Yan function to restrict nodal expression while a constitutively active Yan does not require Panda function to do so . Third , we have shown that Panda and Yan strongly synergize during D/V axis formation consistent with the idea that these genes likely act in the same pathway and are linked by a functional or temporal hierarchy . Finally , we have provided biochemical evidence suggesting that overexpression of Panda causes accumulation of a predominantly non-phosphorylated , and presumably stabilized , form of Yan/Tel . Therefore , taken collectively , these findings strongly suggest a model of symmetry breaking based on the stabilization of Yan/Tel by Panda on the presumptive dorsal side . Indeed , all the evidence cited above are indirect and a definitive proof that Panda stabilizes Yan/Tel and or modulates its activity on the presumptive dorsal side will require additional studies . Nevertheless , we have identified Yan/Tel as a novel and essential regulator of nodal expression acting downstream ( temporally or functionally ) of Panda ( Fig 8 ) . It is presently unclear if the mechanisms that we have described and that implicate the maternal TGF beta Panda and the ETS domain transcription factor Yan/Tel are conserved in other species of sea urchin . We note however , that the panda and yan/tel genes are indeed present in the genomes of all sea urchin species that we have examined suggesting that this may be the case . Furthermore , the MAPK and GSK3 phosphorylation sites of Yan/Tel that we have identified in the protein from Paracentrotus are very well conserved between P . lividus and in S . purpuratus suggesting that the protein is submitted to the same regulation in both species . While Panda and Yan/Tel have not so far been implicated in D/V axis formation in S . purpuratus , early asymmetries in the distribution of mitochondria have been proposed to provide a spatial cue in the process leading to specification of the secondary axis [16 , 51 , 52] . Furthermore , in S . purpuratus , the position of the D/V axis has been shown to be correlated with the plane of first cleavage , a feature that has so far been observed only in that species [53] . Although gradients of mitochondrial activity have been described in P . lividus eggs and embryos , we have so far been unable to obtain evidence for a role of redox gradients or of hypoxia in establishment of the D/V axis in the Mediterranean sea urchin . It is therefore formally possible that different mechanisms operate in different species of echinoderms to establish the D/V axis and that S . purpuratus embryos may rely on a mechanism based redox gradients while P . lividus rely instead on Panda and Yan/Tel to set up the secondary axis . To clarify this question and to test if Panda and Yan/Tel are required in other species of sea urchin for D/V axis establishment , we have recently initiated a comparative study of the genes regulating the spatial restriction of nodal expression in different sea urchin species . Despite the massive ectopic expression of nodal during early stages , and the continued repression of dorsal marker genes until late in gastrulation , both panda morphants and yan/tel morphants progressively recover a spatially restricted expression of nodal and of dorsal marker genes between 36 and 72 h after fertilization so that on the third day of development , these embryos do not show the strong radialization that could have been anticipated from the observation of the same embryos at earlier stages . This progressive recovery of D/V polarity of yan/tel morphants raises a concern regarding the effects of the morpholino since it could be argued that injection of the morpholino might simply delay development and slow down establishment of the D/V axis . This is undoubtedly not the case . yan/tel morphants , as well as panda morphants , display a very specific and characteristic set of morphological and molecular phenotypes that are caused by the failure of a key and early step in the process of D/V axis specification: the early spatial restriction of nodal expression . Both panda and yan/tel morphants display a massive ectopic expression of nodal and Nodal target genes that starts at the 60-cell stage and lasts until late in gastrulation . Both panda and yan/tel morphants remain fully radialized up to the late gastrula stage and both adopt the typical morphology of embryos partially ventralized after treatment with recombinant Nodal protein . In conclusion , panda and yan/tel morphants are strongly ventralized , but this ventralization is not maintained during gastrulation probably because the maternal proteins involved in their function progressively disappear from the embryo allowing compensatory zygotic mechanisms to operate and the embryo to regulate . While searching for sites that may be phosphorylated by MAPK , we discovered that the sequence of sea urchin Yan/Tel contains at least two regions that may be phosphorylated by GSK3 , which targets several proteins for degradation including β-catenin and Snail . The first motif DSGHSS conforms to the degradation box recognized by the E3 Ubiquitin ligase β-TRCP that recognizes the phosphorylated residues within the motif DpSGX ( 1–4 ) pS and promotes ubiquitinylation by the Skp1-Cul1-Fbox complex . The second putative GSK3 phosphorylation motif is a short region rich in serines , threonines and prolines and is located in the C-terminal region . Although there is no strict consensus for phosphorylation by GSK3 , it has been shown that GSK3 usually requires a priming phosphorylation for processive phosphorylation [54] and many GSK3 substrates conform to the sequence S/TXXXpS . Both β-catenin and Snail contain a DSGXXS motif immediately followed by a XXXS/TXXXS/T GSK3 phosphorylation sequence that is crucial for their regulation by this kinase . Intriguingly , the sea urchin Yan/Tel phosphodegron DSGXXS is immediately followed by a PDTAED motif , which does not resemble a canonical GSK3 site . It is therefore unclear if there is a priming kinase that phosphorylates Yan/Tel before GSK3 can phosphorylate the phosphodegron and if this phosphodegron is phosphorylated as efficiently as β-catenin and Snail . Nevertheless , we have provided several lines of evidence to support the idea that GSK3 is a key regulator of Yan/Tel activity/stability and that its β-TRCP motif is indeed phosphorylated by GSK3 . First , we have shown that inhibition of GSK3 caused a dose dependent stabilization of Yan/Tel . Second , we have shown that , in vivo , overexpression of GSK3 strongly destabilized wild type Yan/Tel but did not affect the stability of a Yan13A , a phosphorylation mutant of Yan/Tel . Third , we have shown that mutation of the β-TRCP and S/T rich cluster motifs markedly stabilizes Yan/Tel . Finally , we have shown that both the β-TRCP motif and the S/T rich cluster are phosphorylated in vitro by purified GSK3 . Therefore , Yan/Tel can be added to the list of bona fide GSK3 substrates in the sea urchin . Intriguingly , neither Yan from Drosophila nor Tel from vertebrates contain a β-TRCP motif in their sequence . Indeed , in the fly or in vertebrates , phosphorylation by MAP kinases , not GSK3 , is the main mode of regulation of Yan or Tel and both Yan and Tel have been shown to be degraded primarily following ubiquitination by the ubiquitin conjugating enzyme FBL6 [55] , not by β-TRCP . The regulation of sea urchin Yan/Tel by a GSK3/β-TRCP pathway therefore appears as a novel feature of this ETS family member . The regulation of sea urchin Yan/Tel by a GSK3/β-TRCP pathway may be linked to the very early function of Yan/Tel in specification of the dorsal-ventral axis of the sea urchin embryo and to its role as a coordinator of the A/V and D/V patterning systems . The finding that Yan/Tel is targeted for degradation by a GSK3/β-TRCP pathway , like β-catenin , is particularly interesting . β-catenin turn-over has long been known to be regulated by GSK3–mediated proteolysis [56] . GSK3 is part of a multiprotein degradation complex that involves APC , Axin and Casein kinase . In the absence of Wnt , Casein kinase phosphorylates β-catenin and primes it for phosphorylation by GSK3 . Phosphorylated β-catenin is then recognized by the β-TRCP ubiquitin ligase , which targets it to degradation by the proteasome . In the presence of Wnt ligands , the function of the destruction complex is inhibited and β-catenin levels increase allowing it to translocate into the nucleus where it interacts with TCF to regulate target genes . In the sea urchin embryo , β-catenin is highly unstable in the animal half blastomeres and accumulates in the nuclei of blastomeres located at the vegetal pole of the embryo and fated to become endoderm and mesoderm [57–59] . Although the spatial distribution of maternal Yan/Tel in the early embryo remains to be determined , it is reasonable to think that Yan/Tel is subject to a similar regulation i . e . that it is targeted for degradation by GSK3 in the animal blastomeres and stabilized in the vegetal pole blastomeres where it represses transcription of target genes including nodal . One difference however , is that phosphorylation of sea urchin Yan/Tel does not appear to require a priming phosphorylation by Casein kinase , and therefore the processivity of the phosphorylation of Yan/Tel by GSK3 may not be as high as that of β-catenin . Furthermore , zygotic yan/tel mRNA , accumulates in precursors of the endoderm and mesoderm that express Wnt ligands such as Wnt1 , Wnt8 or Wnt6 and that do not express nodal [60–62] . It is therefore likely that canonical Wnt signaling in the vegetal pole region contributes to repress nodal expression in the endomesoderm by protecting Yan/Tel from GSK3-mediated proteolysis . Consistent with this idea , it has been reported that nodal is ectopically expressed in the vegetal pole region of Wnt1morphants[62] . It will be interesting in the future to test if Wnt ligands directly contribute to the stability of Yan/Tel and to further investigate if Yan/Tel is involved in the crosstalk between MAPK and Wnt signaling . Also , it will be interesting to determine if proteins such as Axin and APC participate to the GSK3-mediated regulation of Yan/Tel stability . In this study , we discovered that Yan/Tel is a key regulator of D/V patterning and that its activity/stability is regulated by GSK3 , a key regulator of patterning along the animal vegetal axis . Since GSK3 plays a central role in patterning along the animal vegetal axis by negatively regulating the stability of β-catenin in the ectoderm , our finding that Yan/Tel stability is regulated by GSK3 strongly suggests a mechanism by which Yan/Tel may integrate signals involved in patterning along the D/V and AP axis . While destabilizing β-catenin in the animal hemisphere to restrict formation of the endoderm to the vegetal pole , GSK3 may simultaneously destabilize Yan/Tel and allow the initiation of nodal expression during early cleavage . Our finding that Yan/Tel is also regulated by MAPK and the results of a recent study on the restriction of animal pole fate by Wnt/JNK signaling further suggest that , starting at 60 cell stage , a second interaction between Wnt signaling and Yan/Tel may occur [19] . Starting at 60-cell stage , early Wnt/β-catenin signaling initiates a cascade of signaling events mediated by Wnt1 and Wnt8 [19 , 61] that activates Fz5/8 /JNK signaling in the ectoderm and restricts the animal pole domain . These findings raise the possibility that early Wnt/β-catenin signaling in the vegetal pole activates JNK signaling in the overlying ectoderm and JNK in turn may promote phosphorylation and degradation of Yan/Tel , thereby allowing nodal expression ( Fig 8 ) . Our finding that nodal expression is attenuated when JNK signaling is inhibited support this model . The finding that ERK is also activated in the ectoderm during this period suggests that additional MAPK dependent signaling events may impact on Yan/Tel stability to promote nodal expression . The model presented in Fig 8 attempts to integrate our data on the regulation of Yan/Tel by GSK3 and MAPKs and by the maternal TGF-β Panda . According to this model , nodal is initially broadly expressed in the presumptive ectoderm at 32-cell stage by the activity of broadly distributed transcriptional activators such as SoxB1 and Oct1/2 as well as by the activity of GSK3 , which inactivates Yan/Tel in the presumptive ectoderm . In this model , GSK3 plays an essential role on nodal expression by releasing the repressive effect of Yan/Tel but it does not provide the initial spatial cues that restricts nodal expression to the ventral side . Instead , the spatial restriction of nodal depends on the spatially restricted activity of the maternal TGF-β ligand Panda , which breaks the radial symmetry and starts to restrict nodal expression by a mechanism that requires the activity of the transcriptional repressor Yan/Tel . Although the mechanism by which Panda antagonizes Yan/Tel is likely indirect and not well understood , we propose that Panda modulates the activity or prevents degradation of the transcription factor Yan/Tel on the presumptive dorsal side . Then , at the early blastula stage , MAPKs signaling in the animal blastomeres further cooperates with GSK3 signaling to promote nodal expression through degradation of Yan/Tel . Nodal promotes its own expression and induces lefty expression and long-range inhibition by Lefty , together with Nodal autoactivation and repression by FoxQ2 in the animal pole domain , contributes to restrict nodal expression to the presumptive ventral ectoderm . Finally , signaling from the Alk4/5/7 Nodal receptor on the ventral territory may also contribute to maintain a high level of nodal expression in this territory possibly by promoting phosphorylation and destabilization of residual maternal Yan/Tel protein . A central and novel feature of this model is the MAP-kinase/GSK3/ß-TRCP /Yan/Tel double repression mechanism of nodal regulation , which is formally similar to the double negative gate that controls the PMC gene regulatory network . In conclusion , this work identified Yan/Tel as a novel key maternal regulator of dorsal-ventral axis . Yan/Tel acts as a repressor of nodal expression whose expression activity/stability is positively regulated by GSK3 , MAPKs and ß-TRCP and likely negatively regulated by Panda . We propose that phosphorylation of Yan/Tel triggers destruction of this maternal repressor in the animal blastomeres , which in turn allows nodal expression to be initiated in the presumptive ectoderm of the zygote . Nodal signaling may further reinforce MAPK signaling and contribute to downregulate Yan/Tel in the presumptive ventral ectoderm . We have therefore uncovered an essential new function for GSK3 in patterning of the sea urchin embryo: not only GSK3 controls the differential stability of β-catenin along the A/V axis [59 , 63] , but it simultaneously controls the expression of Nodal , a central regulator of patterning along the D/V axis . Therefore , this study uncovers a key interaction between the Gene Regulatory Networks responsible for patterning of the endomesoderm and ectoderm and sheds light on the mechanisms that coordinate patterning along the primary and secondary axes of the embryo .
Adult sea urchins ( Paracentrotus lividus ) were collected in the bay of Villefranche-sur-Mer . Embryos were cultured as described in Lepage and Gache ( 1989 , 1990 ) [64 , 65] . Fertilization envelopes were removed by adding 1mM 3-amino-1 , 2 , 4 triazole ( ATA ) 1 min before insemination to prevent hardening of this envelope followed by filtration through a 75 μm nylon net . Treatments with U0126 ( Calbiochem , 10 μM ) BIRB796 ( Selleckchem , S1574 , 3 μM ) , SP600125 ( Tocris , 0 , 5 μM ) , SB431542 ( Tocris , 1614 , 10 μM ) and MG132 ( Tocris 10 μM ) were performed by adding the drug from stocks prepared in DMSO . Recombinant Nodal protein ( R&D ) was used at 1 μg/ml , NiCl2 was used at 0 , 2–0 , 3 mM and LiCl2 at 30mM-50mM . In control treatments , embryos were treated with DMSO alone . A partial yan/tel cDNA was first amplified by PCR with degenerate primers corresponding to the ETS domain . A full-length yan/tel cDNA was subsequently obtained by screening a cDNA library with conventional methods and sequencing the corresponding clones . The 5' end of the cDNA was obtained by performing 5’RACE using the Smart RACE kit ( Clontech ) . The accessions numbers of the Yan/Tel , β-TRCP and GSK3 mRNA are respectively: KF442410 , MG719522 , CAA10901 . Oligonucleotides for making the pCS2 yan/tel and pCS2 β-TRCP construct are: Yan EcoRI-ATG fw: 5'-AGGGAATTCACCATGGATCCAGCATCGGCC-3' Yan-TAG XbaI rev: I 5'-TGCTCTAGACTAGGTCTCCATCTCGGGTGA-3' β-TRCP ClaI ATG fw: 5'- GGGATCGATACCATGGAGACCAGTACTATCAATGAAG-3' β-TRCP TAG XhoI rev 5'- AGGCTCGAGCTATCGACAACTATTAGATACATATG-3' All PCR reactions were made using high fidelity DNA polymerases and the constructs were entirely verified by sequencing . To make the yan/tel rescue construct , an oligonucleotide containing 5 mismatches in the sequence recognized by the morpholino was used to amplify the coding sequence . The sequence of this oligonucleotide is: 5'-CGCGAATTCACCATGGACCCGGCGAGTGCTAGGCAGGTTCATCACTC ( mismatches underlined ) The following oligonucleotides were used to mutate the putative phosphorylation sites of Yan/Tel ( the mutated codon is in bold ) Yan MAPK1 S155A fw 5'-CAGCATGTTCCTGGAGCACCACGGGAACCGATTG-3' Yan MAPK2 S273A fw 5'-ACCCCAGTGCCGCCGGCGCCGACGAAGCGTATC-3' Yan MAPK3 S733A fw 5'- CCCAATAGTAGACCACAGGCACCCGAGATGGACACCTAG Yan β-TRCP mut fw 5'- GACGCTGGCCACAGTGCCCCCGACACGGCAGAAGAC Yan cluster mut fw 5'- GTTCACATGGCTTCGGCGCCAACAGCGCCCGCCCCAGCCCCTCCCGTAGCT Yan S12A fw 5'- GCTAGGCAGGTTCATCACGCACCGCATCCAATGATGCAG-3' Yan S238A fw 5'- CACATCCTCCACTCGTAACGCACCCCAGCA-3' Yan S594A 5'- CACATCCATCTCCACCTCTGCCCCATTGACTGTACCTTC-3' Yan S632A 5'- GGCAGAGTCTACGCTCTATCAGCTCCATCCACAACTG To introduce three copies of the HA epitope tag into Yan , the following oligos were used to amplify the whole pCS2-yan plasmid and the resulting PCR product was phosphorylated and religated . The resulting protein sequence is ( HA tags are underlined and bold sequence from Yan/Tel N-terminus ) : MYPYDVPDYAGYPYDVPDYAGYPYDVPDYAMDPASARQ 3HA-Yan fw -5’ TAC CCA TAC GAT GTT CCA GAT TAC GCT GGC TAT CCC TAT GAC GTC CCG GAC TAT GCA GGA TAT CCA TAT GAC GTT CCA GAT TAC GCT ATG GAC CCG GCG AGT GCT AGG CAG G-3’ Yan-reverse -5' CATGGTGAATTCGAATCGATGGGATCCTGCAAAAAG-3' Total RNA from staged embryos was extracted by the method of Chomczynski and Sacchi ( 1987 ) [66] . Samples of total RNA ( 20μg per lane ) were fractioned on 1% agarose gel containing 0 , 66 M formaldehyde and transferred to membrane by standard methods ( Sambrook et al . , 1989 ) [67] . A P32 labeled RNA corresponding to the Yan/Tel ORF was used as a probe . To perform the in vitro kinase assays , fragments of Yan were cloned into pGex4T and the resulting fusion proteins were purified on glutathione sepharose affinity columns . As a positive control , we used a peptide corresponding to the N-terminal region of human β-catenin which contains a β-TRCP phosphodegron followed by GSK3 and casein kinase phosphorylation sites . To insert the Yan β-TRCP and ß-catenin N-terminal sequences into the pGex4T vector , the following oligonucleotides were used: After amplification , the PCR product was phosphorylated and religated using the Q5 PCR kit from Biolabs . Yan β-TRCP wt-pGex4T1-fw ( PCSDSGHSSPDTA ) 5'-CCGTGCAGTGACTCTGGCCACAGTTCCCCCGACACGGCATTCCCGGGTCGACTCGAGCGG-3' Yan β-TRCP3SA-pGex4T1-FW ( PCSDAGHAAPDTA ) CCGTGCAGTGACGCTGGCCACGCTGCCCCCGACACGGCATTCCCGGGTCGACTCGAGCGG-3' Yan cluster wt-pGEX4T1-fw ( STTPSPSPPVA ) 5'-TCAACAACGCCCTCCCCATCCCCTCCTGTAGCCTTCCCGGGTCGACTCGAGCGG-3' Yan cluster 3A-pGEX4T1-fw ( STAPAPAPPVA ) 5'-TCAACAGCGCCCGCCCCAGCCCCTCCTGTAGCCTTCCCGGGTCGACTCGAGCGG-3' β-catenin- pGEX4T1-fw ( SYLDSGIHSGATTAPC ) 5'TCCTACCTCGATTCCGGTATCCATTCCGGAGCAACAACAGCATTCCCGGGTCGACTCGAGCGG-3' For the in vitro kinase reactions , 15 micrograms of each GST-fused Yan/Tel fragment was incubated for 3 h at 37°C with 250 units of recombinant GSK3 ( Biolabs P6040S ) and 500 units of Casein kinase ( Biolabs P6030S ) in 25 μl of 50 mM Tris pH: 7 . 5 , 10 mM MgCl2 , 0 . 1 mM EDTA , 2 mM DTT , 0 . 01% Brij 35 supplemented with 1mM ATP . The reactions were terminated by addition of SDS sample buffer and an aliquote was subjected to SDS-PAGE . Protein samples equivalent to 600 embryos per well for controls and treated embryos or to 50 embryos for embryos injected with 400ng of HA-Yan/Tel mRNA were separated by SDS-gel electrophoresis and transferred to PVDF membranes . After blocking in 5% dry milk blots were incubated overnight with the primary antibody diluted in 5% BSA in TBST . After washing and incubation with the secondary antibody , bound antibodies were revealed by ECL immunodetection using the SuperSignal West Pico Chemiluminescent substrate ( Pierce ) and imaged with a Fusion Fx7 . Antibodies used: Embryos were fixed with 4% formaldehyde for 15 min then briefly permeabilized with methanol . Anti-Phospho-p38-MAPK ( Thr180-Tyr182 ) was used at 1/50 . Anti-Phospho-p44/42 MAPK ( Erk1/2 ) ( Thr202/Tyr 204 ) was used at 1/300 . Embryos were imaged with an Axio Imager . M2 microscope . Probes derived from pBluescript vectors were synthesized with T7 RNA polymerase after linearization of the plasmids by NotI , while probes derived from pSport were synthesized with SP6 polymerase after linearization with XmaI . The nodal , lefty , chordin , bmp2/4 , tbx2/3 , gata1/2/3 , gcm and 29D probes have been described previously [1 , 2 , 34 , 35 , 68] . In situ hybridization was performed using standard methods ( Harland , 1991 ) with DIG-labeled RNA probes and developed with NBT/BCIP reagent . Detection of the lineage tracer was performed using an anti-fluorescein antibody coupled to alkaline phosphatase and using Fast Red as substrate . Control and experimental embryos were developed for the same time in the same experiments . Embryos were imaged with an Axio Imager . M2 microscope . For overexpression studies , capped mRNAs were synthesized from NotI-linearized templates using mMessage mMachine kit ( Ambion ) . After synthesis , capped RNAs were purified on Sephadex G50 columns and quantitated by spectrophotometry . RNAs were mixed with Tetramethylrhodamine Dextran ( 10000 MW ) , Texas Red Dextran ( 70000 MW ) or Fluoresceinated Dextran ( 70000 MW ) at 5 mg/ml and injected in the concentration range 100–1000μg/ml . Wild-type yan/tel and mutated yan/tel mRNAs were injected at 900–1000 μg/ml . βtrcp mRNA at 1000 μg/ml , gsk3β mRNA at 400 μg/ml , panda mRNA at 1000 μg/ml , nodal mRNA at 400 μg/ml and alk457Q265D mRNA at 400 μg/ml Morpholino oligonucleotides were dissolved in sterile water and injected at the one-cell stage together with Tetramethylrhodamine Dextran ( 10000 MW ) or Fluoresceinated Dextran ( FLDX ) ( 70000 MW ) at 5 mg/ml . For each morpholino a dose-response curve was obtained and a concentration at which the oligomer did not elicit non-specific defects was chosen . Approximately 2–4 pl of oligonucleotide solution were injected in the experiments described here . Sequences for morpholino oligonucleotides used in this study are: yan/tel-Mo5’ UTR: 5'-ATGGTGTCAGGAGTGGGATCAACAC-3' yan/tel-Mo splice: 5'-GGGACTACATACCTATACGTGAGCT-3' yan/tel-Mo 5' ATG: 5'-GATGCTGGATCCATAGGGCTTTGAA-3' lefty-Mo: 5'-GGAGCGCCATGAGATAATTCCATAT-3' foxQ2-Mo: 5'-GTTATCAATGCTGAACAGAGTCATG-3' panda Mo ATG: 5'-ATCTTTGGAATGTGCGTCGAGCCAT -3' The yan/tel morpholino was used at 0 . 75 mM , the lefty morpholino at 1 . 5 mM , the foxQ2 morpholino at 1mM and the panda morpholino at 1mM . In the experiment using suboptimal doses of the Yan and panda morpholinos , these reagents were used respectively at 0 . 5 and 0 . 6 mM . All the injections were repeated multiple times and for each experiment >50 embryos were analyzed ( see supporting S1 Table for more detail of the number of experiments and number of embryos analyzed ) . Only representative phenotypes present in at least 80% of the injected embryos are presented . The conserved predicted binding sites for ETS factors in the R2 module of the nodal promoter [4] were identified using the TransFac software and the MatInspector software from Genomatix . Mutations were introduced by PCR using the Quick Change site directed Mutagenesis kit from Stratagene . All mutations were confirmed by restriction digestion and sequencing . The following primers were used for mutagenesis of ETS sites ( bold letters indicate the mutations introduced , restrictions sites are underlined ) : Ets mut 1 , 2 Fw XhoI: 5’- CACATCTCTTCGTTTTTAAGAAAACTCGAGATGATTAATTAGTATG-3’ Ets mut 1 , 2 Rev XhoI: 5’- CATACTAATTAATCATCTCGAGTTTTCTTAAAAACGAAGAGATGTG-3’ Ets mut 3 Fw XbaI: 5’- TTAATATTCATTAAATCTAGAGTCACTCGTTTTCTTACTT-3’ Ets mut 3 Rev XbaI: 5’- AAGTAAGAAAACGAGTGACTCTAGATTTAATGAATATTAA-3’ Ets mut 4 Fw: 5’-GATTCATTGTTAATTAGTCAGACGGGTTGGGGAGATGGGTTCCTTGTG-3’ Ets mut 4 Rev: 5’-CACAAGGAACCCATCTCCCCAACCCGTCTGACTAATTAACAATGAATC-3’ Ets-mut 5 Fw MluI: 5’-GGTTGATCAATACACGCGTTGTGTAGTGGGCCGA-3’ Ets mut 5 Rev Rev MluI 5’- TCGGCCCACTACACAACGCGTGTATTGATCAACC-3’
|
Specification of the embryonic axes is an essential step during early development of metazoa . In the sea urchin embryo , specification of the dorsal-ventral axis critically relies on the spatial restriction of the expression of the TGF-ß family member Nodal in ventral cells , a process that requires the activity of the maternal determinant Panda . How the spatially restricted expression of nodal is established downstream of Panda is not well understood . We have discovered that , in the Mediterranean sea urchin Paracentrotus lividus , the spatial restriction of nodal on the ventral side of the embryo requires the inhibitory activity of a transcriptional repressor named Yan/Tel . This finding suggests a molecular mechanism for the control of nodal expression by the release of a repression . We found that this release requires the activity of two families of kinases that we identified as the MAP kinases and GSK3 , a kinase which , intriguingly , was previously known as a key regulator of patterning along the animal-vegetal axis . We discovered that phosphorylation by MAPK and GSK3 triggers degradation of Yan/Tel by a β-TRCP proteasome pathway . Finally , we find that Yan/Tel likely acts downstream of Panda in the hierarchy of genes required for nodal restriction . Our study therefore identifies Yan/Tel as a new essential regulator of nodal expression downstream of Panda and identifies a novel key interaction between the gene regulatory networks responsible for patterning along the primary and secondary axis of polarity .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"phosphorylation",
"invertebrates",
"morpholino",
"rna",
"interference",
"nucleotides",
"animals",
"blastulas",
"animal",
"models",
"developmental",
"biology",
"experimental",
"organism",
"systems",
"epigenetics",
"molecular",
"biology",
"techniques",
"embryos",
"echinoderms",
"mapk",
"signaling",
"cascades",
"research",
"and",
"analysis",
"methods",
"ectoderm",
"embryology",
"genetic",
"interference",
"proteins",
"hyperexpression",
"techniques",
"gene",
"expression",
"molecular",
"biology",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"antisense",
"oligonucleotides",
"gene",
"expression",
"and",
"vector",
"techniques",
"biochemistry",
"rna",
"signal",
"transduction",
"eukaryota",
"cell",
"biology",
"post-translational",
"modification",
"nucleic",
"acids",
"oligonucleotides",
"genetics",
"biology",
"and",
"life",
"sciences",
"sea",
"urchins",
"cell",
"signaling",
"organisms",
"signaling",
"cascades"
] |
2018
|
MAPK and GSK3/ß-TRCP-mediated degradation of the maternal Ets domain transcriptional repressor Yan/Tel controls the spatial expression of nodal in the sea urchin embryo
|
The Drosophila larva executes a stereotypical exploratory routine that appears to consist of stochastic alternation between straight peristaltic crawling and reorientation events through lateral bending . We present a model of larval mechanics for axial and transverse motion over a planar substrate , and use it to develop a simple , reflexive neuromuscular model from physical principles . The mechanical model represents the midline of the larva as a set of point masses which interact with each other via damped translational and torsional springs , and with the environment via sliding friction forces . The neuromuscular model consists of: 1 . segmentally localised reflexes that amplify axial compression in order to counteract frictive energy losses , and 2 . long-range mutual inhibition between reflexes in distant segments , enabling overall motion of the model larva relative to its substrate . In the absence of damping and driving , the mechanical model produces axial travelling waves , lateral oscillations , and unpredictable , chaotic deformations . The neuromuscular model counteracts friction to recover these motion patterns , giving rise to forward and backward peristalsis in addition to turning . Our model produces spontaneous exploration , even though the nervous system has no intrinsic pattern generating or decision making ability , and neither senses nor drives bending motions . Ultimately , our model suggests a novel view of larval exploration as a deterministic superdiffusion process which is mechanistically grounded in the chaotic mechanics of the body . We discuss how this may provide new interpretations for existing observations at the level of tissue-scale activity patterns and neural circuitry , and provide some experimental predictions that would test the extent to which the mechanisms we present translate to the real larva .
Exploratory search is a fundamental biological behaviour , observed in most phyla . It has consequently become a focus of investigation in a number of model species , such as larval Drosophila , in which neurogenetic methods can provide novel insights into the underlying mechanisms . However , appropriate consideration of biomechanics is needed to understand the control problem that the animal’s nervous system needs to solve . When placed on a planar substrate ( typically , an agar-coated petri dish ) , the Drosophila larva executes a stereotypical exploratory routine [1] which appears to consist of a series of straight runs punctuated by reorientation events [2] . Straight runs are produced by laterally symmetric peristaltic compression waves , which propagate along the larval body in the same direction as overall motion ( i . e . posterior-anterior waves carry the larva forwards relative to the substrate , anterior-posterior waves carry the larva backwards ) [3] . Reorientation is brought about by laterally asymmetric compression and expansion of the most anterior body segments of the larva , which causes the body axis of the larva to bend [2] . Peristaltic crawling and reorientation are commonly thought to constitute discrete behavioural states , driven by distinct motor programs [2] . In exploration , it is assumed , alternation between these states occurs stochastically , allowing the larva to search its environment through an unbiased random walk [1 , 4–6] . The state transitions or direction and magnitude of turns can be biased by sensory input to produce taxis behaviours [4 , 5 , 7–13] . The neural circuits involved in producing the larval exploratory routine potentially lie within the ventral nerve cord ( VNC ) , since silencing the synaptic communication within the brain and subesophageal ganglia ( SOG ) does not prevent substrate exploration [1] . Electrophysiological and optogenetic observations of fictive locomotion patterns within the isolated VNC [14 , 15] support the prevailing hypothesis that the exploratory routine is primarily a result of a centrally generated motor pattern . As such , much recent work has focused on identifying and characterising the cells and circuits within the larval VNC [16–32] . However , behaviour rarely arises entirely from central mechanisms; sensory feedback and biomechanics often play a key role [33–35] including the potential introduction of stochasticity . Indeed , thermogenetic silencing of somatosensory feedback in the larva leads to severely retarded peristalsis [36] or complete paralysis [37 , 38] . In line with the ethological distinctions drawn between runs and turns , computational modelling of the mechanisms underlying larval behaviour has so far focused on either peristaltic crawling or turning . An initial model based on neural populations described a possible circuit architecture and dynamics underlying the fictive peristaltic waves observed in the isolated ventral nerve cord [39] . A subsequent model described the production of peristaltic waves through interaction of sensory feedback with biomechanics , in the absence of any centrally generated motor output [40] , in a manner similar to earlier models of wave propagation via purely sensory mechanisms in C . elegans [41 , 42] . This model produced only forward locomotion as it incorporated strongly asymmetric substrate interaction . Recently , a model combining biomechanics , sensory feedback , and central pattern generation reproduced many features of real larval peristalsis [43] . However , this model only aimed to explain forward locomotion , and accordingly contained explicit symmetry-breaking elements in the form of posterior-anterior excitatory couplings between adjacent segments of the VNC , and posterior-anterior projections from proprioceptive sensory neurons in one segment into the next segment of the VNC . No biomechanical models of turning in the larva have yet been published , but the sensory control of reorientation behaviour has been explored in more abstract models [4 , 5 , 8 , 11–13 , 44] . No current model accounts for both peristalsis and reorientation behaviours , and no current model of peristalsis can account for both forward and backward locomotion without appealing to additional neural mechanisms . Here we present a model of unbiased substrate exploration in the Drosophila larva that captures forward and backward peristalsis as well as reorientation behaviours . We provide a deterministic mathematical description of body mechanics coupled to a simple , reflexive nervous system . In contrast to previous models , our nervous system has no intrinsic pattern-generating ability [39 , 43 , 44] , and does not explicitly encode discrete behavioural states or include any stochasticity [4 , 5 , 8 , 11–13] . Nevertheless , the model is capable of producing apparently random “sequences” of crawling and reorientation behaviours , and is able to effectively explore in a two-dimensional space . We argue that the core of this behaviour lies in the chaotic mechanical dynamics of the body , which result from an energetic coupling of axial ( “peristaltic” ) and transverse ( “turning” ) motions . Our choice not to explicitly model navigational decision-making and central pattern generation circuits is motivated by our desire to illustrate the powerful insights offered by focusing upon the mechanics of the body with which the nervous system interacts . The model neuromuscular system we have constructed is based upon simple physical arguments , yet ultimately bears a striking resemblance to known features of the larval nervous system . By starting from the mechanics of the body , and not assuming the existence of particular neural circuits , we are able to provide a new explanatory framework within which to re-interpret existing neurophysiological observations , including observations of central pattern generation within the larval VNC , and also provide unique predictions for future neurophysiological experiments . In what follows , we first outline the key components and assumptions of our model of body mechanics . We then follow simple arguments to guide the construction of a neuromuscular model capable of producing power flow into the body , and motion of the body’s centre of mass relative to the substrate . Crucially , the neuromuscular model neither senses nor drives transverse motions . In analysing the behaviour of our model , we begin by focusing on the small-amplitude , energy-conservative behaviour of the body in the absence of frictive and driving forces . In this case , the motion of the body is quasiperiodic and decomposes into a set of energetically isolated axial travelling waves and transverse standing waves . Reintroducing friction and driving forces , we demonstrate the emergence of a pair of limit cycles corresponding to forward and backward peristaltic locomotion , with no differentiation of the neural activity for the two states . We then shift focus to the behaviour of the model at large amplitudes . In this case the axial and transverse motions of the body are energetically coupled , and the conservative motion becomes chaotic . The energetic coupling allows our neuromuscular model to indirectly drive transverse motion , producing chaotic body deformations capable of driving substrate exploration . Analysis of our model supports a view of larval exploration as an ( anomalous ) diffusion process grounded in the deterministic chaotic mechanics of the body .
To explore larval crawling and turning behaviours , we choose to describe the motion of the larval body axis ( midline ) in a plane parallel to the substrate ( Fig 1 , S1 Fig ) . The larval body is capable of more diverse motions including lifting/rearing [21] , rolling [45] , digging [46] , self-righting / balancing , and denticle folding which we have recently observed to occur during peristalsis ( S1 Video ) . However , while exploring flat surfaces , the larva displays fairly little out-of-plane motion ( neither translation perpendicular to the substrate nor torsion around the body axis ) and only small radial deformations [47] . Furthermore , the majority of ethological characterisations of larval exploration treat the animal as if it were executing purely planar motion [4 , 6 , 8–13 , 48] . A planar model is thus a reasonable abstraction for the issues addressed in this paper , i . e . , the generation of peristalsis , bending , and substrate exploration . The segmented anatomy of the Drosophila larva allows us to focus our description of the midline to a set of N = 12 points in the cuticle , located at the boundaries between body segments and at the head and tail extremities . We assign each point an identical mass , and measure its position and velocity relative to a two dimensional cartesian coordinate frame fixed in the substrate ( the laboratory or lab frame ) . We therefore have NDOF = 2N = 24 mechanical degrees of freedom . We note that our assumption of a uniform mass distribution along the midline is somewhat inaccurate , since thoracic segments are smaller than abdominal segments . However , simulations with non-uniform mass distribution give results which are qualitatively close to those presented here . We assume that the larval body stores elastic energy in both axial compression/expansion and transverse bending , due to the presence of elastic proteins in the soft cuticle . We assume that energy is lost during motion due to viscous friction within the larva’s tissues and sliding friction between the body and the substrate . Sliding friction also allows shape changes ( deformations ) of the body to cause motion of the larva as a whole relative to the substrate ( centre of mass motion ) . Since the mechanical response of the larva’s tissues is yet to be experimentally determined , we assume a linear viscoelastic model . This is equivalent to placing linear ( Hookean ) translational and torsional springs in parallel with linear ( Newtonian ) dampers between the masses in the model , as shown in Fig 1 , or to taking quadratic approximations to the elastic potential energy and viscous power loss ( as in S1 Appendix ) . We note that the accuracy of the approximation may decrease for large deformations , in which nonlinear viscoelastic effects may become important . As with larval tissue mechanics , there has been little experimental investigation of the forces acting between the larva and its environment . We therefore assume a simple anisotropic Coulomb sliding friction model , in which the magnitude of friction is independent of the speed of motion , but may in principle depend upon the direction of travel . This anisotropy could be thought of as representing the biased alignment of the larva’s denticle bands , or directional differences in vertical lifting or denticle folding motions which are not captured by our planar model . A mathematical formulation of our sliding friction model is given in S1 Appendix . In addition to power losses due to friction , we also allow power flow due to muscle activation . For the sake of simplicity , we choose to allow only laterally symmetric muscle tensions . In this case , the musculature cannot directly cause bending of the midline , and can only explicitly drive axial motions . We will see later that even indirect driving of bending motion can lead to surprisingly complex behaviour , due to energetic coupling of axial and transverse degrees of freedom . The choice to neglect asymmetric muscle tensions is made in order to simplify our model and provide a clearer illustration of the potential role of body mechanics in generating complex larval behaviour . We note that there is only one way for muscle activations to be symmetric—if we were to allow asymmetry we would have to specify exactly what form that asymmetry should take , and we have little empirical or theoretical grounds on which to do so . Nevertheless , there are some interesting cases which may be considered in passing—the presence of a constant torque about the model’s segment boundaries should cause a shift in the equilibrium posture towards a resting curved shape . The presence of torques which are a linear function of the local body bending angle or local angular velocity should shift the effective transverse stiffness or viscosity of the body S5 Appendix . In this sense the model presented here could be considered to already include the effect of asymmetric muscle tensions , they have simply been incorporated into the passive stiffness and viscosity of the body . We have recently developed an extension of the model presented here which uses a similar local reflex to modulate the body’s effective transverse viscosity in proportion to a stimulus input , allowing the model to exhibit taxis behaviour [49] . Finally , we model the internal coelomic fluid of the larva . Given the extremely small speed of the fluid motion compared to any reasonable approximation to the speed of sound in larval coelomic fluid , we can safely approximate the fluid flow as incompressible [50] . This would ordinarily require that the volume contained within the larval cuticle remain constant . However , since we are modelling only the motion of the midline and neglecting radial deformations , we constrain the total length of the larva to remain constant . We note that this constraint is not entirely accurate to the larva , as the total length of the animal has been observed to vary during locomotion [47] . Nevertheless , for the sake of simplicity we will continue with this constraint in place , noting that this approximation has been used with success in previous work focused on peristalsis [40 , 43] , and that there is experimental support for kinematic coupling via the internal fluid of the larva [3] . We note that we satisfy the incompressibility condition only approximately in some sections ( Model behaviour—Conservative chaos , Dissipative chaotic deformations , and Deterministic exploration ) , by introducing an additional potential energy associated with the constraint , which produces an energetic barrier preventing large changes in the total length of the body ( see S1 Appendix for details of this approximation along with specifics of the mathematical formulation of our mechanical model ) . Note that in the absence of transverse bending , the total length constraint causes the head and tail extremities of the larva to become mechanically coupled and move in unison [40 , 43] . The axial mechanics thus has periodic boundary conditions , and the most anterior ( T1 ) and posterior ( A8 ) segments of the larva may be considered adjacent . This means , for instance , that a compression wave travelling from tail to head will cause motion of the tail on termination at the head , thus initiating a new compression wave . This view also allows us to reason about what should happen if we relax the total length constraint . In particular , if we were to replace the direct coupling of head and tail by a viscoelastic coupling , representing the capacity for storage and dissipation of energy within the internal fluid or in radial expansion of the cuticle , the axial mechanics would still have periodic boundary conditions but would now have a step change in mechanical impedance . Waves hitting such “sudden” impedance boundaries in their transmission media will generally be partially transmitted ( i . e . passing directly from head to tail in the larva ) and partially reflected ( i . e . changing direction and moving backwards from the head extremity ) , providing one possible cause of transitions between forward and backward locomotion in the animal . As will be seen , however , these transitions may occur even in the absence of an impedance discontinuity , and we will continue with the total length constraint in place in order to simplify our model . Let us now consider how we should use muscle activity to produce locomotion . There are two basic requirements . First , we must have power flow into the body from the musculature , so that the effects of friction may be overcome and the larva will not tend towards its equilibrium configuration . Second , we must be able to produce a net force on the centre of mass of the larva , so that it can accelerate as a whole relative to the lab frame . Note that in this section , we motivate the neural circuits in the model from this purely functional point of view , but will present relevant biological evidence in the discussion . To satisfy the first criterion , let us examine the flow of power into the body due to the action of the musculature P = - ∑ i = 1 N - 1 b i MF i q ˙ i ( 1 ) Here , qi describes the change in length of the i’th body segment away from its equilibrium length , q ˙ i is the rate of expansion of the i’th body segment , bi is a ( positive ) gain parameter , MFi is a ( positive ) dimensionless control variable representing muscle activation , and the product biMFi is the total axial tension across the i’th body segment . From this expression , it is clear that if we produce muscle tensions ( MFi > 0 ) only while segments are shortening ( q ˙ i < 0 ) , we will always have positive power flow into the body ( P > 0 ) . This is a mathematical statement of the requirement for the larva’s muscles to function as motors during locomotion , rather than as springs , brakes , or struts [33] . A simple way to fulfil this condition is to introduce a segmentally localised reflex circuit ( Fig 2 , [40] ) . We place a single sensory neuron in each segment which activates when that segment is compressing ( q ˙ i < 0 ) . Each sensory neuron then projects an excitatory connection onto a local motor neuron , which in turn projects to a muscle fibre within the same segment . Assuming for now that there are no other influences on the motor neurons , so that sensory activation implies local motor neuron activation , segmental shortening will produce an immediate muscle tension serving to amplify compression of the segment and thus counteract frictive energy losses . Let us now consider the second criterion for peristaltic locomotion . Assuming all segment boundaries are of equal mass , the force on the centre of mass of the larva is proportional to the sum of the forces acting on each segment boundary , i . e . F C O M ∝ ∑ i = 1 N - 1 F s e g m e n t ( 2 ) Newton’s third law tells us that any forces of interaction between segment boundaries ( i . e . viscoelastic and muscle forces ) must be of equal magnitude and opposite direction , so that they cancel in this summation and we are left only with contributions arising from substrate interaction . If the motion of the body is such that some number nf of segments move forward at a given time , against a frictional force −μf , while nb segments remain anchored or move backward , experiencing a frictional force μb , then the summation becomes F C O M ∝ n b μ b - n f μ f ( 3 ) In the limiting case of isotropic ( direction-independent ) substrate interaction we have μb = μf , and this expression tells us that the centre of mass will accelerate in the forward direction only when there are less segments moving forward than are moving backward or anchored to the substrate . Similarly , moving a small number of segments backward while the others remain anchored will result in backward acceleration of the centre of mass . Therefore , if the animal is to move relative to its substrate , it must ensure that only a limited number of its segments move in the overall direction of travel at a given time ( indeed , this matches observations of the real larva [3 , 22] ) . A more lengthy exposition of this requirement on limbless crawling behaviours can be found in [51] . We fulfil the requirement for a small number of moving segments by introducing mutually inhibitory interactions between the segmentally localised reflex circuits ( Fig 2 ) . We add a single inhibitory interneuron within each segment . When the sensory neuron within the local reflex activates , it excites this interneuron , which then strongly inhibits the motor neurons and inhibitory interneurons in non-adjacent segments , effectively turning off the local reflexes in distant neighbours . Adjacent segments do not inhibit each other in our model , allowing reflex activity to track mechanical disturbances as they propagate from one segment to the next . We comment on the plausibility of this feature of our model , given the experimental observation of nearest-neighbour inhibitory connections in the larval ventral nerve cord [28] , in the discussion . Similarly , the head and tail segments do not inhibit each other , which permits peristaltic waves to be ( mechanically ) reinitiated at one extremity as they terminate at the other ( see Discussion at the end of the previous subsection ) . This effectively introduces a ring-like topology into the neural model , matching our model of axial mechanics which couples head and tail motion through the total length constraint [40] . We now have a neuromuscular model consisting of four cell types repeated in each segment—sensory neurons , inhibitory interneurons , motor neurons , and muscle fibres . For the sake of simplicity we model all neurons as having a binary activation state governed by the algebraic relation V i = { 1 ∑ j w j V j > θ i 0 otherwise ( 4 ) where Vi is the activation of the i’th cell , θi is its activation threshold , Vj is the activation of the j’th presynaptic cell , and wj is the associated synaptic weight . Numerical values for the weights and thresholds used in our model are given in S1 Table , supplemental . Note that the muscle tension over a segment either vanishes ( when the muscle fibre is in the inactive state ) or has fixed magnitude bi ( when the muscle fibre is activated by local sensory feedback ) . For this reason we refer to bi as the reflex gain . Our choice to neglect neural dynamics is based on the large difference in timescales between the neural and mechanical dynamics . Typical motor neuron spiking occurs with a timescale on the order of 10−3 seconds . Spiking is observed to be significantly “averaged out” by the graded ( non-spiking ) muscle fibre responses , and respond on the order of ∼10−1 seconds to prolonged motor neuron spiking [52 , 53] . During locomotion , segmental compressions are driven by several longitudinal muscle fibres activating simultaneously [3 , 14 , 29] in response to largely independent motor neuron populations [54 , 55] which are unlikely to spike with identical timing . This spatial integration should further “mask” the effects of neural dynamics . Note that the lack of neural dynamics in our model immediately rules out central pattern generation . However , this does not prevent our model from producing complex , larva-like behaviour , and we consider how our model could account for observations of central pattern generation in the discussion . To summarise , the neural model we have constructed can be seen as consisting of two parts , a segmentally repeating local reflex and a mutual inhibition circuit acting between non-adjacent reflexes . The local reflex is constructed so that muscles will act as motors , amplifying segmental compressions and counteracting friction . The mutual inhibition circuit couples distant reflexes to allow only localised amplification . By limiting the number of moving segments , this should ensure that the model larva can produce a net force on its centre of mass .
One of the advantages of grounding our model of larval exploration in the body’s physics is that we are now able to apply powerful analytical results from classical mechanics in order to understand the model’s behaviour . In this section we attempt to elucidate the naturally preferred motions of the larva by focusing our attention on the conservative mechanics of the body while neglecting friction forces , which would cause all motion to stop , and driving forces , which might impose a particular pattern of motion . In this case , the general character of motion is specified by the Liouville-Arnold integrability theorem . This theorem asks us to look for a set of conserved quantities associated with a mechanical system , which remain unchanged as the system moves ( energy , momentum , and angular momentum are examples of some commonly conserved quantities ) . If we can find a number of these quantities equal to the number of mechanical degrees of freedom in our model , then the theorem tells us that the motion of the body is integrable—it can be expressed analytically , and must be either periodic or quasiperiodic . If there are not enough conserved quantities , then the system is said to be nonintegrable . In this case the motion is much more complicated and will be chaotic for some initial conditions . These chaotic motions do not permit analytical expression and must be approximated through simulation . In this section , we explicitly seek a case for which there is a “full set” of conserved quantities ( we provide only major results here , for detailed derivations see S2 Appendix ) . We begin by restricting ourselves to considering only small deformations of the larval midline , in the case where all segments are of identical axial stiffness ka , transverse stiffness kt , mass m , and length l . Under these assumptions , the total mechanical energy of the body may be written H ( x , y , p x , p y ) = 1 2 [ p x T p x + ω a 2 x T D 2 x ] + 1 2 [ p y T p y + ω t 2 y T D 4 y ] ( 5 ) where x and y are vectors giving the displacement of each segment boundary along the body axis and perpendicular to the body axis , respectively , px and py give the translational momentum associated with each direction , D2 and D4 are difference matrices arising from a Taylor series expansion of our model’s potential energy ( see S2 Appendix ) , and ω a = k a / m and ω t = k t / m l 2 are characteristic axial and transverse frequency scales . By making a linear change of coordinates {x , y , px , py} → {X , Y , pX , pY} to the eigenbasis of D2 and D4 ( see S2 Appendix ) this simplifies to H ( X , Y , p X , p Y ) = ∑ i = 1 N - 1 1 2 [ p X , i 2 + ω a 2 λ a , i X i 2 ] + ∑ i = 1 N 1 2 [ p Y , i 2 + ω t 2 λ t , i Y i 2 ] ( 6 ) where λa , i and λt , i are eigenvalues associated with the coordinate transformation . This expression is a sum of component energies , each of which is independently conserved . The Liouville-Arnold theorem immediately tells us that the motion of the body must be ( quasi ) periodic in the case of conservative small deformations . Indeed , the energy associated with each of the new coordinates Xi , Yi is in the form of a harmonic oscillator , and each coordinate executes pure sinusoidal oscillations . By transforming back to the original coordinates x , y we obtain a set of collective motions ( modes ) of the body which are energetically isolated and have a sinusoidal time dependence , corresponding to axial and transverse standing waves . We will refer to the Xi , Yi as modal coordinates since they describe the time dependence of each of the collective motions . Each transverse standing wave corresponds to a periodic lateral oscillation of the body , with a unique frequency given by ω t , i = ω t λ t , i . We determined these frequencies numerically , along with the spatial components of the lowest frequency standing waves ( Fig 3A ) . These can be seen to match the eigenmaggot shapes extracted from observations of unbiased larval behaviour [56] . The axial standing waves correspond to oscillating patterns of segmental compression and expansion . While each transverse standing wave had its own unique frequency of oscillation , the axial standing waves come in pairs with identical frequency but different spatial components—each member of the pair corresponds to a different spatial pattern of segmental compression/expansion spread across the body , but these patterns oscillate in time with the same frequency . We were able to analytically determine the frequency of oscillation of the i’th pair of axial standing waves to be ω a , i = ω a λ a , i = 2 ω a | sin ( π i N - 1 ) | , i ∈ [ 0 , N / 2 - 1 ] ( 7 ) The spatial components of the axial standing waves could also be determined analytically x k , i = 1 N - 1 cos ( 2 π i k N - 1 ) , or x k , i = 1 N - 1 sin ( 2 π i k N - 1 ) , i ∈ [ 0 , N / 2 - 1 ] ( 8 ) Where xk , i is the displacement of the k’th segment boundary for the i’th pair of standing waves . We plot the axial frequencies ωa , i and spatial components xk , i in Fig 3B . The fact that the axial oscillation frequencies come in identical pairs allows us to combine the axial standing waves with a ±90° relative phase shift to form pairs of forward and backward travelling wave solutions ( see S2 Appendix for the full derivation ) x k , i ( t ) = cos ( ω a , i t ± 2 π i k N - 1 ) , i ∈ [ 0 , N / 2 - 1 ] ( 9 ) where xk , i ( t ) gives the displacement of the k’th segment boundary as a function of time for the i’th pair of travelling waves . The choice of a plus or minus sign corresponds to the choice between forward or backward wave propagation . These solutions correspond to propagating waves of segmental compression and expansion similar to those seen during larval peristalsis . We plot the lowest frequency pair of axial travelling wave solutions in Fig 3C , and directly visualise the synthesis of travelling wave solutions from standing wave solutions in S2 Video . To summarise , in this section we have shown that for the case of conservative , small oscillations , the motion of the body may be decomposed into a combination of transverse standing waves and axial travelling waves . This is of clear relevance to understanding the behaviour of the larva , which moves across its substrate by means of axial peristaltic waves while reorienting using lateral oscillations . Indeed , the transverse modes of oscillation that we have derived here match principal components of bending computed from actual larval behaviour [56] . Our results can be interpreted as providing a physical basis for these observations—the principal components extracted from real larval data correspond to a “natural” coordinate basis that is grounded in the animal’s mechanics . Furthermore , the proportion of postural variance explained by each principal component of the experimental data decreases with increasing modal frequency in our model ( and thus increasing energy ) . We can therefore help to explain the observed ordering of principal components , as this corresponds to the larva “preferring” to occupy low-frequency , low-energy modes during most of its behaviour . We comment further on this observation in S3 Appendix in the context of axial modes . We will now focus on the small-amplitude motion of the body in the presence of energy dissipation due to friction and driving forces . Reintroducing friction will clearly lead the motions described above to eventually terminate due to energy dissipation , unless opposed by transfer of power . In a previous section ( Models—Neuromuscular system , see also S1 Appendix ) , we introduced a neuromuscular system to produce power flow into the body , but as described , it can only directly transfer power into the axial degrees of freedom . In the small deformation model we have just analysed , the axial and transverse degrees of freedom are energetically decoupled . It follows that transverse friction is unopposed and any transverse motion must eventually terminate in the case of small deformations . In this section we will therefore focus only on the axial degrees of freedom , which correspond to the peristaltic locomotion of the larva . In Fig 4 , we show the effect of coupling our neuromuscular model to the axial mechanics . For small reflex gains , the power flow into the body from the musculature is too low to effectively counteract frictive losses and the larva tends towards its passive equilibrium state , in which there is complete absence of motion . However , increasing reflex gain past a certain critical value leads to the emergence of long-term behaviours in which the larva remains in motion , away from its passive equilibrium . These motions correspond to forward and backward locomotion , driven by forward and backward propagating compression waves ( see below ) , as predicted from our earlier description of the conservative motions of the body , and as observed in the real larva [3] . The qualitative changes in behaviour that occur as reflex gain is varied are depicted in Fig 4A , where we have measured the long-term centre of mass momentum of the larva , along with the long-term relative phase of the lowest frequency modal coordinates . The exact value of reflex gain required for onset of locomotion depends on the particular mechanical parameters used in our model ( see S2 Table for parameters used in Fig 4 ) . In principle , this bifurcation point of the dynamics should be amenable to analytical investigation by studying the stability of the linearised model dynamics around the passive equilibrium state [57 , 58] . In practise , however , the presence of hard nonlinearities in the sliding friction model makes such an approximation inaccurate . For low reflex gains the centre of mass momentum tends to 0 as the body comes to a stop and enters a passive equilibrium state . The relative phase of the low frequency modal coordinates tends to either 0 or 180 degrees , which also corresponds to a loss of momentum . For larger values of reflex gain , the total momentum is either positive , zero , or negative . Positive and negative total momentum represent forward and backward locomotion , respectively , while zero momentum corresponds to two unstable cases which we discuss below . The relative phase of the lowest frequency modal coordinates tends to ±90° at high reflex gains , corresponding to the presence of forward- or backward-propagating compression waves ( see previous section ) . As in the larva [1 , 3] , forward-propagating waves drive forward locomotion while backward-propagating waves drive backward locomotion ( Fig 4B ) . We believe that these behaviours arise from the production of a pair of limit cycle attractors in the system’s phase space , which we visualise in Fig 4C by projecting the system state onto the lowest frequency pair of axial modes , and plotting the associated modal coordinates along with the centre of mass momentum . Since wave motion implies that pairs of modal coordinates should perform pure sinusoidal oscillations with equal amplitude and frequency , and a ±90° relative phase shift ( see previous section and S2 Appendix ) , the travelling wave trajectories of the system become circles in this coordinate system ( see discussion of Lissajous figures , [59] ) . Forward and backward locomotion can then be distinguished by the momentum of the centre of mass . In this model , the speed of forward and backward locomotion are equal for a fixed value of reflex gain , while in the real larva the speeds are known to differ [15] . We comment on some possible explanations for this difference in the discussion . In S2 Fig we show the neural state of the model larva during forwards and backwards locomotion . As expected given our previous exposition , we observe waves of activity in the nervous system which track the mechanical waves propagating through the body . Our sensory neurons also show a second , brief period of activation following propagation of the mechanical wave caused by a slight compression that occurs as segments return to equilibrium . This activity is “cancelled out” by the mutual inhibition circuit , so that motor neurons do not exhibit a secondary burst of activity . The larva has zero long-term total momentum in the presence of large reflex gain in only two cases , both of which are highly unstable . First , if we start the larva so that it is already in its passive equilibrium state , so that no relative motion of segment boundaries occurs , it is obvious that there will be no activation of local reflexes and the larva will not spontaneously move out of equilibrium . The second case corresponds to a pure axial standing wave . In this case the larva is divided into two regions by nodal points where the axial displacement is zero . During the first half-cycle of the standing wave , one region will experience compression while the other experiences expansion , and in the second half-cycle these roles will reverse . The neural circuit we have constructed can amplify compression during both half-cycles since they are separated by a configuration in which no compression or expansion occurs , and this allows the entire nervous system to become inactive and “reset” . Since these behaviours are extremely unstable and require very specific initial conditions to be observed , we have not visualised them here . While the mutually inhibitory connections in our model are not required for the propagation of locomotor waves , which can be maintained entirely by local reflexes [40] , these connections do greatly enhance stability . In the absence of the mutual inhibition circuit , small mechanical disturbances in any stationary body segments can be amplified , giving rise to multiple compression waves which travel through the body simultaneously . If this instability produces an equal number of forward and backward moving segments then overall motion of the larva relative to the substrate will stop , in line with the argument presented earlier . We have also observed that roughly symmetrical substrate interaction is required to produce both forward and backward locomotion in our model . If friction is too strongly anisotropic , then locomotion can only occur in one direction regardless of the direction of wave propagation . It is worth noting that the axial model presented in this section does display discrete behavioural states . However , there are no explicit decisions regarding which behavioural states to enter , since the particular neural states occupied during forwards and backwards locomotion are indistinguishable . Rather , both the apparent decision and the eventual direction of travel can only be understood by examining the dynamics and mechanical state of the body . Having successfully produced peristaltic locomotion using our model , we will now turn our attention to the larva’s turning behaviours . As before , we will start from physical principles . In a previous section ( Results—Conservative axial compression waves and transverse oscillations ) we showed that , for the case of conservative small oscillations , transverse motions of the body were energetically decoupled from axial motions , and could be decomposed into a set of periodic standing waves . We will first extend our previous analysis to the case of energy-conservative , large amplitude motions in the absence of damping and driving; and then in the following section consider the impact of energy dissipation and the neuromuscular system on transverse motion . To keep our presentation simple and allow visualisation of model trajectories , we will focus on a reduced number of the mechanical degrees of freedom . Namely , we will examine the bending angle ϕ and axial stretch q of the head segment ( Fig 5A ) . We introduce an amplitude parameter ϵ by making the substitutions q → ϵq and ϕ → ϵϕ , so that the total mechanical energy of the head may be written in nondimensional form as ( see S4 Appendix ) H = 1 2 [ p q 2 + 1 ( 1 + ϵ q ) 2 p ϕ 2 + q 2 + λ 2 ϕ 2 ] ( 10 ) where pq , pϕ are the radial and angular momentum of the head mass , and we have scaled the time axis of the model so that the natural frequency of axial oscillation is unity . The parameter λ = ωt/ωa = kt/kal2 is the ratio of transverse and axial frequencies . In the case of small oscillations , i . e . ϵ → 0 , the mechanical energy reduces to the simpler expression H = 1 2 [ p q 2 + q 2 ] + 1 2 [ p ϕ 2 + λ 2 ϕ 2 ] ( 11 ) which is clearly a sum of independent axial and transverse energies . These energies are individually conserved , so that the Liouville-Arnold theorem applies , and the motion of the head is integrable and ( quasi ) periodic . This is in clear agreement with our earlier results . For the more general case of large amplitude motion ( ϵ > 0 ) , we may have in principle only a single conserved quantity—the total energy of the system . Indeed , it should be clear from the presence of a “mixed” term in the mechanical energy ( Eq 10 ) that the axial and transverse motions are energetically coupled at large amplitudes , so that the individual energies associated with each motion are no longer independently conserved . Given that we have only one conserved quantity for a two degree of freedom system , we can no longer rely on the Liouville-Arnold theorem to prove ( quasi ) periodicity of the motion , and must accept that the system’s behaviour may be chaotic . To investigate this possibility further , we first note that conservation of energy implies that flow within the four dimensional phase space must be constrained to lie on the energy surface given implicitly by the relation H ( q , ϕ , pq , pϕ ) = E . Therefore , given a particular value E for the total energy , the system dynamics becomes three dimensional . This allows us to visualise the behaviour of the system by plotting the points at which trajectories intersect a two-dimensional Poincare section [57 , 58] . We define our Poincare section by the condition that the angular momentum vanishes pϕ = 0 ( equivalently , angular velocity vanishes dϕ/dt = 0 ) , and plot successive crossings of the section as points in the q , ϕ plane . This allows us to intuitively interpret points in the Poincare section as configurations of the head at successive turning points ( extrema ) in the transverse motion ( Fig 5B ) . In what follows , we set the total energy to be E = 1 2 so that when ϵ = 1 we can in principle obtain complete compression of the head ( q = −1 ) . We choose to set λ = e 6 ≈ 0 . 45 , giving an irrational frequency ratio . This loosely matches observations of the real larva in which the frequency of transverse oscillations is approximately half that of axial oscillations but does not satisfy an exact ( rational ) resonance relationship [44] . The results we obtain with these parameters do not differ much from results for other energies or other frequency ratios , including resonant relationships . Poincare plots for the cases ϵ → 0 and ϵ ∈ { 1 3 , 2 3 , 1 } are shown in Fig 6 . When ϵ → 0 ( Fig 6A ) , conservation of transverse energy implies that the turning points of the transverse motion must remain constant . The fact that the frequency ratio λ is irrational implies that the overall motion is quasiperiodic , and the values of q obtained at successive transverse turning points should not repeat . In accordance with these observations , the Poincare section for ϵ → 0 consists of a series of verticle lines ( Fig 6Ai ) . For ϵ = 1 3 the Poincare plot becomes distorted , but the majority of trajectories still trace out one-dimensional curves in the section ( Fig 6Bi ) , which is indicative of persistent quasiperiodic behaviour . At ϵ = 2 3 the Poincare plot appears qualitatively different . There is now a large region of what appears to be “noise” , indicating that the configuration of the head at successive transverse turning points has become unpredictable . This is a clear signature of deterministic chaos . There are , however , several regions of the section indicative of ( quasi ) periodic behaviour . These appear as one-dimensional curves or discrete points in the Poincare section ( Fig 7Ai ) . At ϵ = 1 , the region of the Poincare plot occupied by chaos has expanded , although there still appear to be some regions corresponding to ( quasi ) periodic behaviour ( Fig 7Bi ) . These results qualitatively agree with the Kolmogorov-Arnold-Moser theorem [59] , which tells us that quasiperiodic behaviour should persist under small nonintegrable ( chaotic ) perturbations of an integrable Hamiltonian , and that the region of phase space corresponding to chaotic behaviour should grow with the perturbation size ( in our case , the perturbation size corresponds to the amplitude of motion ϵ ) . However , our model as presented here does not formally meet the requirements of this theorem ( see S4 Appendix ) . Analysis by Poincare section provides an invaluable method to determine the character of overall system behaviour by direct visualisation of trajectories , but cannot be applied to higher dimensional systems . This is problematic , since we would like to assess the existence of chaos beyond our reduced model of the larva’s head . We therefore deployed a series of other methods which are possibly less reliable than the method of Poincare section but can be applied equally well to higher dimensional systems . These included estimation of the maximal Lyapunov characteristic exponent ( MLCE ) for the dynamics along with calculation of the power spectrum and autocorrelation of internal variables [57 , 58 , 60] . The MLCE can be thought of as quantifying the rate of separation of nearby trajectories , or , equivalently , the rate at which information is generated by the system being analysed [61] . A positive MLCE is generally considered a good indicator of chaotic behaviour . The power spectrum of a periodic or quasiperiodic process should consist of a “clean” set of discriminable peaks , whereas that of a chaotic process should appear “noisy” and contain power across a wide range of frequencies . Meanwhile , the autocorrelation of a periodic or quasiperiodic process should show a strong oscillatory component with an envelope that decays linearly with time , while that of a chaotic process should show a much quicker decay , similar to a coloured noise process . In Fig 6 we plot these measures at each value of ϵ , for a trajectory starting with initial conditions indicated on the corresponding Poincare plot by a filled black circle . These measures confirm increasingly chaotic behaviour as the amplitude ϵ increases , in agreement with our Poincare analysis . In Fig 8 we show a solution including all degrees of freedom in our conservative mechanical model ( i . e . not just those of the head ) . The trajectory of individual segments relative to the substrate appears qualitatively irregular , while the indirect measures we introduced above ( MLCE , power spectrum , autocorrelation ) are all indicative of deterministic chaotic behaviour . We will now reintroduce dissipative effects into our model of larval motion in the plane . We previously saw that conservative mechanics predicted axial travelling waves and transverse oscillations . These were lost in the presence of friction , but the axial travelling waves could be recovered with the addition of a neuromuscular system designed to selectively counteract frictive effects . We have now seen that conservative mechanics predicts chaotic planar motion . Although our neuromuscular model transfers power only into the axial degrees of freedom , we recall from the previous section that axial and transverse motions are energetically coupled at large amplitudes . We therefore tentatively expect that we may be able to recover the complete chaotic planar motion without requiring any additional mechanism for direct neuromuscular power transfer into tranverse motion . To investigate whether our dissipative planar model shows chaotic behaviour , we ran n = 1000 simulations starting from almost identical initial conditions ( euclidean distance between initial mechanical state vectors < 10−7 , with no initial neural activity ) and identical parameters ( see S5 Table ) . We can indeed observe that the simulated larva perform peristalsis with interspersed bending of the body ( turns ) , and that the fully deterministic system produces apparently random turning such that the simulations rapidly diverge ( S3 Video ) . Since most working definitions of chaos require strictly bounded dynamics , we here restrict our analysis to the coordinates describing deformation of the body ( segmental stretches and bending angles ) , ignoring motions of , or overall rotations about , the centre of mass ( i . e . , the trajectory through space of the body , which we will analyse in the following section ) . Qualitatively , the deformations of the large amplitude dissipative model appear irregular ( Fig 9A ) . However , there are persistent features reminiscent of the ordered small-amplitude behaviour described in previous sections . In particular , there are clear axial travelling waves and lateral oscillations . However , whereas forward- and backward-propagating axial waves previously corresponded to stable limit behaviours , the large amplitude system appears to go through occasional “transitions” between these behaviours . In addition , apparently spontaneous large bends appear occasionally , suggesting that the neuromuscular system is effectively driving transverse motion . The irregularity of the axial motion is reflected in the pattern of sensory neuron activation ( S3 Fig ) . However , the mutual inhibitory interactions in our model act to filter this input , allowing only a small window of excitability within the central nervous system . As a result , interneuron and motor neuron activity appears fairly ordered , tracking and reinforcing axial compression waves . We used four measures to assess whether our qualitative observation of irregular motion betrays the existence of deterministic chaos . First , we analysed the power spectrum of individual cooordinates ( Fig 9B ) . The power spectra of all degrees of freedom showed consistent harmonic peaks along with a strong “noisy” non-harmonic component , a pattern consistent with chaos and incommensurate with ( quasi ) periodicity ( Fig 9B shows data for head bending angle and stretch Next , we computed the autocorrelation of the same degrees of freedom . The autocorrelations of all degrees of freedom showed a periodic pattern with a peak at 0 seconds time lag followed by a rapid decay ( Fig 9C ) . This is characteristic of oscillatory chaotic behaviour , and the exponential loss of information regarding initial conditions that chaotic systems display . We then estimated the correlation dimension ( Dc ) of the limit set of our model’s dynamics . Note that we did not apply this measure to the conservative models in the previous section since the Liouville theorem rules out attracting limit sets for these systems . The distribution of correlation dimension estimates for our dissipative system across all 1000 trials is shown in Fig 9D . Estimates were clustered around ∼ 3 . 5 ( median dimension 3 . 46 ) , with 93% of estimates lying in the range [3–4] . These results are suggestive of a limit set with fractal dimension , which is a signature of “strange” chaotic attractors . Finally , we computed an estimate of the maximal Lyapunov characteristic exponent ( MLCE ) . As in the previous section , we estimated the MLCE for our system to be positive ( ∼ 13textrmbitss−1 , Fig 9D ) , a very strong indicator of chaotic behaviour . All of these results point to the presence of a chaotic dynamical regime in our dissipative large amplitude model . As the coupled biomechanical and neuromuscular system produces both forward and backward peristalsis and lateral oscillations , each simulated larva will trace out a 2D trajectory over time . As shown in Fig 10A , the chaotic deformations characterised in the previous section caused the larvae to disperse across their two-dimensional substrate , following paths reminiscent of the real animal’s exploratory behaviour . To characterise the trajectories of our model , we first investigated them at a global level , based on the centre of mass ( COM ) trajectory of each simulated larva , computing the tortuosity and fractal dimension of the paths ( Fig 10B ) [62] . We defined our tortuosity measure as T = 1 - D L ( 12 ) where D is the net displacement of the COM between initial and final times , and L is the total distance travelled by the COM along its path . Note that if the COM travels in a straight line between initial and final times we will have D = L so that T = 0 . In the limit L → ∞ , corresponding to the COM taking an extremely long path between its initial and final states , we have D L → 0 so that T → 1 . We calculated the mean tortuosity of our COM trajectories to be 〈T〉 = 0 . 43 , with a variance of 〈 ( T − 〈T〉〉 ) 2 = 0 . 05 . The lowest ( highest ) tortuosity observed was T = 0 . 05 ( T = 0 . 95 ) . We estimated the fractal dimension Db of the COM trajectories using a box-counting algorithm . The minimum expected dimension Db = 1 would correspond to curvilinear paths ( e . g . straight line or circular paths ) , while the maximum expected dimension of Db = 2 corresponds to plane-filling paths ( e . g . brownian motion ) . We calculated the mean dimension of the COM trajectories to be 〈Db〉 = 1 . 37 , with variance 〈 ( Db − 〈Db〉 ) 2〉 = 0 . 01 . The lowest ( highest ) path dimension observed was Db = 1 . 17 ( Db = 1 . 95 ) . We have plotted the tortuosity and fractal dimension of every path in Fig 10B . These results show that the trajectories of the model differed markedly from straight lines ( tortuosity T > 0 and box-counting dimension Db > 1 ) , and displayed a good ability to cover the planar substrate ( box-counting dimension 1 < Db < 2 ) . We also note that our COM trajectories display the power-law relationship between angular speed and curvature reported by [63] , with a scaling exponent ( β ≈ 0 . 8 ) falling within the range reported for freely exploring larvae ( S4 Fig ) . We next investigated the rate at which the simulated larvae explored their environment . To do this , we calculated the mean-squared displacement ( MSD ) of the COM over time ( Fig 10C ) . This is a standard measure used to characterise diffusion processes , and is defined as ⟨ d 2 ⟩ ( t ) = 1 n ∑ i = 1 n ( R i ( t ) - R i ( 0 ) ) 2 ( 13 ) where Ri ( t ) is the position of the i’th larva’s COM at time t and n = 1000 is the number of trials being averaged over . We observed an initial transient , lasting on the order of 10 seconds , during which the MSD grew as ∼ t2 ( blue line , S5 Fig ) , after which growth slowed and tended to ∼ t ( linear fit for t > 80 seconds shown by red line , Fig 10C and S5 Fig , r2 = 0 . 99 , diffusion constant D = 144segs2s−1 ) . The initial transient was not due to our particular initial conditions , since it remained even after discarding > 60s of initial data . These results therefore tell us that , although on long timescales our model appears to execute standard Fick diffusion or a Brownian random walk ( linear growth of MSD ) , on short timescales the model’s behaviour is superdiffusive ( approximately quadratic growth of MSD ) [64 , 65] . This is in good agreement with observations of the real larva [6 , 48] . The superdiffusive behaviour of the larva was previously explained in terms of a persistent random walk [6] , in which the larva’s current and previous headings are highly correlated during straight runs so that the animal follows an approximately ballistic trajectory on short timescales . We believe that persistence effects arise in our model due to the finite time required for the deterministic chaotic dynamics to destroy information regarding initial conditions . We next calculated some other standard measures found in the larva literature . To do so , we built a two-segment representation of each simulated larva by drawing vectors from the tail extremity to the A2-A3 segment boundary ( the tail vector , T ) , and from the A2-A3 boundary to the head extremity ( the head vector , H ) . We then defined the body bend , θ , to be the angle between tail and head vectors , θ = atan ( Hy/Hx ) − atan ( Ty/Tx ) . The head angular velocity ν was computed as the cross-product of the head vector and the head extremity’s translational velocity r ˙ head measured relative to the lab frame , ν = H × r ˙ head , while the tail speed v was taken to be the magnitude of the tail extremity’s translational velocity r ˙ tail measured relative to the lab frame , v = r ˙ tail · r ˙ tail . The tail speed and head angular velocity both show a strong oscillatory component , which can be seen in the time and frequency domains ( S6 Fig ) . The power spectra of these variables contains considerable “noise” over a wide spread of frequencies , in accordance with the results of the previous section . The distribution of tail speeds for our model is bimodal , similar to that of the real larva [44] . The body bend angle was observed to be symmetrically distributed ( Fig 10D ) , with roughly zero mean ( 〈θ〉 = 0 . 005 ) , small variance ( 〈 ( θ − 〈θ〉 ) 2〉 = 0 . 13 ) , slight positive skew ( SK ( θ ) = 0 . 23 ) , and high excess kurtosis KU ( θ ) = 7 . 9 . The kurtosis of our data precludes a good fit by the von Mises distribution ( maximum likelihood estimate shown by red line in Fig 10D ) . Our data appears to be better fitted by a wrapped Cauchy distribution , though this overestimates the central tendency of our data ( maximum likelihood estimate shown by blue line in Fig 10D ) . The high excess kurtosis of the body bend distribution gives a quantitative measure of the rare large bending events mentioned at the beginning of the previous section , and qualitatively matches experimentally observed distributions of real larval bends [44 , 66] . Our model also reproduces the observed overall speed of larval locomotion ( median model speed = 0 . 26 body lengths s−1 , real larval range ∼ 0 . 1 − −0 . 35 body lengths s−1 ) , the turn rate ( median model turn rate = 2 . 08 turns min−1 , real larval range ∼ 0 − −4 . 5 turn min−1 , threshold body bend for turn classification = 30deg to match relevant literature ) , and the relative distance gained during free locomotion ( median model distance gained = 0 . 14 body lengths s−1 , real larval range ∼ 0 − −0 . 2 ) , with our results being more consistent with observations of third instar than first instar larvae [66] . Finally , we computed a run-length distribution by setting a threshold body bend angle θturn = 20° ( as in [13] ) and calculating the length of time between successive crossings of this threshold . The distribution we obtained appears approximately linear on a log-linear plot ( Fig 10E , linear fit r2 = 0 . 99 with slope λ = −0 . 075 ) , and is better fit by an exponential than a power law distribution ( maximum likelihood estimates , log likelihood ratio = 5281 , p < 0 . 01 ) . Together with our observation of asymptotic linear growth of MSD , the exponential distribution of run lengths suggests that the model can be considered to be effectively memoryless on long timescales [65] . This again agrees with the observed rapid loss of information from the system due to its chaotic dynamics , as quantified by the Lyapunov exponent and autocorrelation analysis of the previous section . Ultimately , the analysis of our model supports a view of the larval exploratory routine as an ( anomalous ) diffusion process arising from the deterministic chaotic dynamics of the body . The model nervous system functions purely to recover these dynamics from the effects of frictive energy dissipation , and to ensure centre of mass motion , rather than explicitly directing exploration .
The intrinsic capabilities of an organism’s body determine the field of possibilities that neural circuits for behaviour can exploit . Here , by focusing first on the biomechanics of Drosophila larva , we find that its body already contains an inherent exploratory routine . This is demonstrated through a combined biomechanical and neuromuscular model that is the first to be able to generate both forward and backward peristalsis and turning , allowing 2D motion in the plane to be simulated . We show that , in the absence of friction , the body’s conservative mechanics alone supports both axial travelling waves and transverse standing waves . These are energetically coupled at larger amplitudes , such that no driving , sensing , or control of body bend is required for the system to start producing spontaneous coordinated bending motions . Frictional losses can be recovered , to maintain axial waves , by a neuromuscular system consisting of only simple local sensorimotor reflexes and long-range inhibitory interactions . This is sufficient to produce emergent crawling , reversal and turning that resembles larval exploratory behaviour , and which is chaotic in nature . At a population level , we observe a deterministic anomalous diffusion process in which an initial superdiffusive transient evolves towards asymptotic Fickian/Brownian diffusion , matching observations of real larvae [6 , 48] . We therefore propose that the role of biomechanical feedback in Drosophila larvae goes beyond the periphery of basic neuromuscular rhythms [40 , 43] , to provide the essential “higher order” dynamics on which exploratory behaviour is grounded . Most existing models of larval exploration abstract away from the mechanics underlying the production of runs and reorientations [4–6 , 8 , 11–13] . The larva is often described as executing a stochastic decision-making process which determines which state ( running or turning ) should be occupied , and when to initiate a change of behavioural state . In contrast , our model produces the entire exploratory routine without making any decisions ( the transverse motion is neither sensed nor driven by the nervous system ) nor introducing any stochastic process ( neural or otherwise ) . Similarly , transient “switching” is seen to occur between forward and backward peristalsis even though there is no neural encoding or control of the direction of wave propagation . In other words , the body dynamics generate the basis of a chaotic exploratory routine which only needs to be amplified by the neural circuitry , making the search for underlying stochastic or state switching circuitry superfluous for this behaviour . The work presented here also stands in contrast to previous models of larval peristalsis [39 , 43] and the prevailing hypotheses regarding this phenomenon [15 , 67] by eschewing any role for intrinsic neural dynamics . Such stereotyped and rhythmic locomotion is widely assumed to be the signature of a central pattern generator ( CPG ) , that is , a neural circuit that intrinsically generates a rhythmic output , and thus determines a particular mechanical trajectory to be followed by the body [68–70] . However it is recognised that systems vary in the degree to which coordinated behaviour is independent of biomechanical and sensory feedback [70] . Indeed , evidence from studies employing genetic manipulations to disrupt sensory neuron input suggest that proprioceptive feedback is necessary for correct larval locomotive patterns [16 , 36–38 , 71]; although in some cases coordinated waves of forward and backward peristalsis can be produced , in both intact [16 , 36 , 71] and isolated VNC preparations [14 , 15] , these are reported as abnormal with the most evident defects being time-dilation [15 , 36 , 71] and abnormal frequency in polarity changes [71] . In fact , our intent is not to adjudicate between the roles of intrinsically generated activation sequences vs . biomechanical feedback in this system , but rather to note that we should expect neural circuits of locomotion to adhere to the dynamical modes of the associated body , instead of working against them . Thus it should be unsurprising if these dynamics also exist ( potentially in a latent form ) in the neural circuitry . For example , a simple modification of the neural circuit presented here could produce instrinsic ‘peristaltic’ waves . Recall that the long-range global inhibition pattern in our model treats head and tail segments as ‘neighbouring’ nodes ( see Models—Neuromuscular system ) . If local constant input or recurrent feedback were added to each segment , the circuit would then resemble a ring attractor [72–74] and a stable activity bump would be formed . Breaking the forward/backward symmetry of the circuit , e . g . , by introducing asymmetric nearest-neighbour excitatory connections [75] , would cause the activity bump to move along the network , giving rise to intrinsic travelling waves . This would complement any mechanical compression waves travelling through the body , i . e . , remain consistent with the principles set out in this paper . Would such a network be a CPG ? The answer is unclear . On the one hand , it would show spontaneous rhythmic activity in the absence of sensory input . On the other , sensory feedback would do much more than simply correct deviations from the CPG output or provide a “mission accomplished” signal [36] . Rather , feedback would play a crucial role in orchestrating motor output to ensure power flow into the body , consistently with its dynamical modes . It is important to note that the emergence of rhythmic peristalsis and spontaneous turns in our model is not strongly dependent on the specific assumptions made in our mechanical abstraction . For example , the observation for small amplitude motions of sinusoidal axial travelling waves , along with transverse standing waves whose shapes match the experimentally observed “eigenmaggots” [56] , is a direct result of the second-order Taylor series approximation of the model Hamiltonian ( S2 Appendix ) . The small-amplitude model is thus non-unique , since many different mechanical models could have identical second-order approximations . Similarly , we expect that the deterministic chaotic behaviour derived from our conservative model for large amplitude motions will hold for other models of the larval body , given that it is conjectured that the majority of Hamiltonian systems are nonintegrable . This may also mean that our results can be applied to other animals with body morphologies and mechanics similar to the Drosophila larva . In our model we constrain the total length of the larva to be constant . This constraint is intended to represent the fact that there is minimal observable radial deformation of the larva’s body during behaviour , yet its body is filled with fluid which should conserve volume . We were further motivated by the experimental observation of “visceral pistoning” [3] in which the head and tail extremities of the larva appear to be mechanically coupled via the coelomic fluid during peristalsis . However , the total length of the real larva is known to change during behaviour [47] , and it is therefore important to consider the effect of weakening the length constraint in our model . When restricted to small-amplitude motion , the total length constraint appears as periodic boundary conditions in the axial mechanics , allowing waves of compression to propagate from head to tail and vice versa . In the complete absence of the length constraint , these waves will instead be reflected back from the head and tail extremities , leading to alternating forward and backward waves . Alternatively , replacing the constraint with a simple linear viscoelastic model to represent energy storage and dissipation within the internal fluid and in radial cuticle deformation leads to the presence of a new mechanical impedance between the head and tail . It is well known that sudden impedance changes in wave transmission media lead to simultaneous reflection and transmission of waves—in our model , this means that some amount of the axial compression wave will be transmitted between head and tail while some will be reflected . Since our neural model cannot sustain two peristaltic waves concurrently due to the presence of mutual inhibition between distant segments , this causes occasional “switching” between forward and backward peristalsis . If the extent of radial deformations is under neural control in the larva , this could provide a potential route for control or biasing of transitions between forward and backward peristalsis . As a consequence of exploiting body mechanics , our model explains a wider range of behaviour than previous models , using a simpler nervous system . The properties included in the neuromuscular circuitry were derived from basic physical considerations , i . e . , what was necessary and sufficient to produce exploration , rather than from known neuroanatomy or neurophysiology . However , it is useful to now examine what insights and predictions regarding this circuitry can be derived from our model . Firstly , we consider the connections between segments . Unlike the model from [43] , we did not require assymmetric connections to obtain forward ( or backward ) waves as these ( and spontaneous switching between them ) arise inherently in the mechanics . Rather , obtaining centre of mass motion of the entire body required the “ring attractor” layout of mutual inhibition between distant segments described above . The model thus predicts that motor output should be strongly inhibited ( by signalling from other segments ) the majority of the time , so that motor neurons only activate as the ( mechanical ) peristaltic wave passes through the corresponding body segment . This is in contrast to previous models which appealed only to local , nearest-neighbour inhibitory connections [39 , 43] . What might be the neural substrate for the proposed inhibition ? There are two currently known intersegmental inhibitory pathways in the larva . GVLI premotor inhibitory neurons synapse onto motor neurons within the same segment but extend their dendritic fields several segments further anterior along the VNC . Accordingly , the GVLIs inhibit motor neurons at a late phase during the local motor cycle [22] . Our model predicts that there should be a larger set of GVLI-like neurons within each segment , with dendritic fields tiling distant segments . Although in our model the mutual inhibition is ( for simplicity ) arranged to act on all non-adjacent segments , we would in practice expect that active compression is actually spread across more segments [3 , 22] to transfer power to the body more efficiently ( S3 Appendix ) , and this should be reflected in the inhibitory connection pattern . The second inhibitory pathway involves GDL inhibitory interneurons , which receive input from the excitatory premotor neuron A27h in the nearest posterior segment , and synapse onto A27h within the same segment while simultaneously disinhibiting premotor inhibitory neurons in distant segments [28] . Thus , GDL effectively produces both local and long-range inhibition of motor output . However , GDL receives axo-axonic connections from vdaA and vdaC mechanosensory cells within the same segment , so local inhibition is likely gated by sensory input . This would match our model , in which sensory activation within a segment should be sufficient to produce motor output when one of the neighbouring segments is active . We thus predict that simultaneous experimental suppression of GDL , GVLI , and all other long-range inhibition in the VNC should allow the propagation of several , concurrent locomotor waves in response to mechanical input . Secondly , within a segment , our model highlights the importance of the timing of neuromuscular forces relative to body motion . Specifically , during locomotion , the larva’s muscles should act primarily as motors rather than as springs , brakes , or struts ( see [33] for a discussion of these differences ) , and thus should activate in phase with the segmental stretch rate . This hypothesis could be tested by performing work-loop experiments , for which we predict the existence of a counterclockwise cycle in a plot of muscle force ( potentially measurable by calcium imaging ) over segment length during locomotion . Can our model’s requirement that neurons sensing stretch-rate provide a direct excitatory connection to motor neurons , within the same segment , be mapped to identified pathways in the larva ? One possible monosynaptic implementation of such a link are the dda mechanosensory cells which have been observed to make synapses onto aCC and RP2 motor neurons [23] . However , synapse counts show high variability both within and across individuals , so it seems unlikely to be a core component of the locomotor circuitry . A more promising candidate is the excitatory premotor interneuron A27h , which receives input from vpda and vdaC and sends bilaterally symmetric outputs to aCC and RP5 [28] . It is known that A27h activation is sufficient to activate downstream motor neurons , but it remains unknown whether proprioceptive sensory input is sufficient to activate A27h . Additionally , we hypothesise that A02 ( PMSI ) interneurons [20] , which have been recently shown to form an inhibitory sensory-motor feedback pathway between dbd mechanosensory cells and motor neurons [27] , could play a role in filtering this signal to obtain the necessary stretch-rate activation independently of stretch . General models of mechanotransduction suggest that larval mechanosensory cells may be sensitive to both rate of stretch as well as absolute stretch , depending upon the mechanical properties of the sensory dendrites and the biophysics of the relevant mechanosensitive ion channels [76] . If PMSIs have a slow-activating , integrator dynamics that encodes stretch , while A27h activate quickly in response to proprioceptive sensory input to encode stretch and stretch-rate , the combined input to motor neurons would be only stretch-rate dependent excitation , as our model requires . This could explain the observation that optogenetic disturbance of PMSIs [20] slows the timescale of peristaltic waves , as the inclusion of absolute stretch in this feedback loop would produce muscle forces that not only counteract friction but also decrease the effective stiffness of the cuticle , slowing peristalsis ( see S5 Appendix ) . It is clear the real larval nervous system exhibits many complexities not reflected in our model , and likewise that the real larva performs many more behaviours than exploration . These include appropriate ( directed ) reactions to sensory stimuli such as stopping , withdrawal and reverse in response to touch stimuli [38]; differences in the speed of forward and backward locomotion [15]; and modulation of the frequency and direction of ( large ) turns in response to sensory gradients such as odour , heat or light [4 , 8–10 , 12 , 13 , 77–82] to produce positive or negative taxis . In a previous model of taxis [44] we have shown that by a continuous coupling of the amplitude of a regular lateral oscillation to the experienced change in stimulus strength in a gradient , a larva-like response to gradients can emerge , again without requiring active switching between states . In the current model , this could be effected by incorporating direct neuromuscular driving of bending degrees of freedom , since the real larva can likely use asymmetric activation of its lateralised muscles to produce active bending torques to influence the transverse motion . Alternatively , the degree of bend could be influenced indirectly by altering the stiffness and viscosity of segments ( as explored in our upcoming paper [49] ) , or their frictional interaction with the substrate . We note that the effective viscoelasticity of body segments can be neurally controlled by local reflex arcs ( see S4 Appendix and [40] ) . Indeed , this could partially explain the experimental observation of increased bending on perturbation of a contralateral segmental reflex mediated by Eve+ interneurons [24] . The muscle activation caused by this reflex should produce bending torques which are proportional to current bend or bending rate , thus effectively modulating transverse stiffness or viscosity , respectively . Notably , in the taxis model of [44] , it is not required that the descending signal that alters turn amplitude is lateralised , but rather that it has the right temporal coordination , which itself is naturally created by the interaction of body and environment . Backward locomotion is observed to be slower than forward locomotion in the real larva [15] , yet in our model both behaviours are of equal speed for a fixed value of reflex gain . We believe that this is due to the preservation of mechanical symmetry between forward and backward motion in our model . The real larva likely experiences asymmetric substrate interaction forces . For instance , this could be due to the exact coordination of denticle folding/lifting during forward and backward locomotion ( S1 Video ) or due to the geometry of the larva’s denticle bands , which display a degree of anisotropy [83] . Alternatively , there may be asymmetries within the larva’s neural circuitry responsible for this difference . Indeed , there do appear to be neurons in the larval VNC which are only active specifically during forwards or backwards locomotion , and these may be functionally asymmetric [28] . The model presented in this paper does occasionally produce stops ( cessation of peristalsis ) during exploration , but this only occurs in concert with a large body bend ( this stored transverse energy can subsequently and spontaneously restart the peristalsis ) ; whereas in larva slowing , stopping and resumption of peristalsis ( or transition from a stop to a large bend ) can occur while the body is relatively straight [2 , 10] . As for ‘directed’ turning , this suggests that additional neural control might be needed to terminate or initiate movement in response to sensory stimuli . It is worth noting that our model predicts that peristalsis can be restarted by almost any small disturbance of the physical equilibrium provided the local feedback gain is high enough; similarly , lowering the gain means that energy losses due to friction are not compensated and the animal will stop . In general , we have found that altering assumptions about the sliding friction forces by which the model interacts with the substrate can often have unexpected and subtle effects on the motion produced , thus it would be interesting to further explore the functions provided by segmental lifting [3 , 84] , folding of the denticle bands ( S1 Video ) , and extrusion of the mouth-hooks [3 , 85] during locomotion . Indeed , detailed experimental characterisation of the substrate interaction forces experienced by the larva would be a major advance in understanding how the animal behaves . Inspiration for approaches to this problem could perhaps be taken from the literature on C . elegans substrate interaction ( see for instance [42 , 86–90] , though this list is not exhaustive ) . In the more extreme case , larva are capable of burrowing through a soft substrate , and it is clear that a complex interaction of forces , mechanics , sensing and neural control must be involved that go well beyond the scope considered here .
|
We investigate the relationship between brain , body and environment in the exploratory behaviour of fruitfly larva . A larva crawls forward by propagating a wave of compression through its segmented body , and changes its crawling direction by bending to one side or the other . We show first that a purely mechanical model of the larva’s body can produce travelling compression waves , sideways bending , and unpredictable , chaotic motions . For this body to locomote through its environment , it is necessary to add a neuromuscular system to counteract the loss of energy due to friction , and to limit the simultaneous compression of segments . These simple additions allow our model larva to generate life-like forward and backward crawling as well as spontaneous turns , which occur without any direct sensing or control of reorientation . The unpredictability inherent in the larva’s physics causes the model to explore its environment , despite the lack of any neural mechanism for rhythm generation or for deciding when to switch from crawling to turning . Our model thus demonstrates how understanding body mechanics can generate and simplify neurobiological hypotheses as to how behaviour arises .
|
[
"Abstract",
"Introduction",
"Models",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"classical",
"mechanics",
"neuroscience",
"biological",
"locomotion",
"motor",
"neurons",
"developmental",
"biology",
"reflexes",
"damage",
"mechanics",
"waves",
"traveling",
"waves",
"bending",
"animal",
"cells",
"deformation",
"life",
"cycles",
"physics",
"cellular",
"neuroscience",
"cell",
"biology",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"larvae",
"motion"
] |
2019
|
Modelling the mechanics of exploration in larval Drosophila
|
Animal African trypanosomosis , a disease mainly caused by the protozoan parasite Trypanosoma congolense , is a major constraint to livestock productivity and has a significant impact in the developing countries of Africa . RNA interference ( RNAi ) has been used to study gene function and identify drug and vaccine targets in a variety of organisms including trypanosomes . However , trypanosome RNAi studies have mainly been conducted in T . brucei , as a model for human infection , largely ignoring livestock parasites of economical importance such as T . congolense , which displays different pathogenesis profiles . The whole T . congolense life cycle can be completed in vitro , but this attractive model displayed important limitations: ( i ) genetic tools were currently limited to insect forms and production of modified infectious BSF through differentiation was never achieved , ( ii ) in vitro differentiation techniques lasted several months , ( iii ) absence of long-term bloodstream forms ( BSF ) in vitro culture prevented genomic analyses . We optimized culture conditions for each developmental stage and secured the differentiation steps . Specifically , we devised a medium adapted for the strenuous development of stable long-term BSF culture . Using Amaxa nucleofection technology , we greatly improved the transfection rate of the insect form and designed an inducible transgene expression system using the IL3000 reference strain . We tested it by expression of reporter genes and through RNAi . Subsequently , we achieved the complete in vitro life cycle with dramatically shortened time requirements for various wild type and transgenic strains . Finally , we established the use of modified strains for experimental infections and underlined a host adaptation phase requirement . We devised an improved T . congolense model , which offers the opportunity to perform functional genomics analyses throughout the whole life cycle . It represents a very useful tool to understand pathogenesis mechanisms and to study potential therapeutic targets either in vitro or in vivo using a mouse model .
Animal African Trypanosomosis ( AAT ) is a severe disease caused by several species of African trypanosomes , flagellate protozoan parasites transmitted by an insect fly vector . It affects domestic livestock in sub-Saharan Africa and consequently has a severe economic impact [1]–[3] . AAT is widely spread affecting 40 countries situated in regions that could potentially be the most productive . The main pathological symptoms of animal trypanosomosis are weight loss , anaemia , and immunosuppression , but the mechanisms involved are poorly understood . It is estimated that 50 million cattle and 70 million small ruminants are at risk , costing the continent between 1 . 5 and 5 billion US dollars per annum . In terms of severity and consequences for productivity , T . congolense is the main causative agent of AAT . This parasite has a complex life cycle: bloodstream forms ( BSF ) proliferate in the blood of the infected mammalian host and are ingested by tsetse during the blood meal , procyclic forms ( PCF ) differentiate in the insect midgut , migrate to the proboscis ( mouth parts ) where they attach as epimastigote forms ( EMF ) and finally differentiate into infective metacyclic forms ( MCF ) that are transmitted to a new mammalian host during the next blood meal . This parasite strongly hinders the agricultural development of the sub-Saharan region and an understanding of the virulence mechanisms is essential in order to block the parasite and/or the associated pathogenesis . Such studies require functional characterization of virulence genes based on genetic engineering . Genetic tools are well established in the T . brucei parasite with wide use of inducible transgene expression [4] , [5] . More specifically , gene expression inactivation using RNA interference ( RNAi ) system greatly contributed to the functional analysis of T . brucei genes to such an extent that T . brucei has become the preferred model for trypanosomiasis studies [6] , [7] . However , T . congolense is a strictly intravascular parasite whereas T . brucei can leave blood vessels and invade tissues implying differences in virulence mechanisms , host/pathogens relationships and pathogenic effects between the two species . Furthermore , while for T . brucei only the PCF and BSF stages can be cultivated in vitro and differentiation is limited to one step ( from BSF to PCF ) , the whole life cycle of T . congolense can be reproduced in vitro ( cultivation of all the developmental stages accomplishing all the differentiation steps ) [8] , [9] . Especially , the in vitro differentiation of the tsetse vector stages ( from PCF to MCF ) was described much earlier , this process is termed metacyclogenesis [10] . Therefore , it will be very informative and useful to use T . congolense not only as a model for AAT studies but also for deciphering the mechanisms underlying the different differentiation steps . The major drawback we aim to overcome is that all cultivation techniques currently available are limited even for the metacyclogenesis process ( which could take several months [10] ) resulting in the quasi-absence of genetic tools . A single strain has been developed for the inducible RNAi system in PCF ( TRUM183 strain ) [11] , and only one recent study described genetic modification of EMF ( IL3000 strain ) by transfection with a GFP vector and subsequent differentiation into MCF [12] . However , production of infectious BSF through differentiation of genetically modified parasites has never been achieved . In addition , there is an important heterogeneity between the T . congolense isolates used despite the inclination to refer to the IL3000 clone since the beginning of its genome sequencing project ( Sanger Center , UK ) . And , there are disparities regarding the in vitro differentiation especially during metacyclogenesis and transfection efficiency [13] ( our data ) . Finally , the question of the infectivity/virulence of wild type or genetically modified MCF differentiated in vitro has not been addressed until now . Another severe restriction in this research field is the lack of stable long-term BSF culture . Some authors described BSF culture systems [14] , [15] that turned out to be usable only for short-term applications eliminating the possibility of performing drug trials ( requiring continuously dividing cultures ) and above all , transfection assays , consequently resulting in the absence of functional genomics for T . congolense BSF . Because the T . congolense model displayed significant limitations , we decided to undertake a systematic analysis of all the developmental stages regarding culture methods , differentiation efficiency , and transfection success rate by using 5 different wild type isolates and 2 genetically modified cell-lines . In this work we describe the success in developing an improved in vitro model and standardized culture methods .
Primary cultures of bovine aortic endothelial cells ( BAE ) were routinely maintained in EGM2V medium ( Lonza ) in 24 well plates in a 5% CO2 humidified atmosphere at 37°C as described by the manufacturer . Cells were used between passages 9 and 20 . Co-culture with parasites was conducted in the parasite medium . Six Trypanosoma congolense strains of the Savannah type were used in this study . IL1180 originated from the Serengeti in Tanzania [16] . ILC-49 was originally isolated from a cow from the Transmara , Kenya , and was passaged in rodents [17] . This strain and its derivative , clone IL3000 , were kindly provided by the International Livestock Research Institute , Nairobi , Kenya . STIB910 is a cloned derivative of STIB249 originally isolated in 1971 from a lion in Tanzania [16] , and was kindly provided by R . Brun ( Basel , Switzerland ) . TREU1457 was derived from a stock originally isolated in Nigeria ( Zaria/67/LUMP/69 ) [18] kindly provided by C . A . Ross . TRUM183:29-13 is a genetically modified version of the TRUM183 strain , an uncloned serodeme , kindly provided by J . E . Donelson [11] . Details on culture media used are listed in Table 1 . Procyclic forms ( PCF ) were routinely maintained in TcPCF-3 medium . Cells were maintained in log-phase culture at 27°C in 25 cm2 flasks by changing the medium every 3 days . Epimastigote adherent forms ( EMF ) appeared in this medium after several weeks of culture by maintaining the culture in stationary phase ( change of half of the medium every 4 days ) . The entire culture supernatant was replaced with fresh medium every 2–3 days , after the colonies of EMF were well established . Alternatively , EMF could be cultivated in a few hours using a starvation medium named TcEMF-1 . Briefly , 5 ml of PCF culture ( ≈107 cells/ml ) was collected by centrifugation ( 1600×g , 10 min ) , placed in the same volume of TcEMF-1 medium and incubated at 27°C in 25 cm2 flasks for 2 hours . Individual adherent EMF or colonies were observed after only 1 hour . Heat-inactivated fetal calf serum ( 10% , Adgenix ) was added to the flask and this medium ( named TcEMF-2 ) was used to maintain the obtained EMF culture as described above . Metacyclic forms ( MCF ) developed in the EMF cultures after few days up to few weeks depending on the strain . MCF could be purified by DE52 anion-exchange column chromatography ( Whatman Plc . , Brentford , U . K . ) [19] . Furthermore , the number of MCF produced per ml of EMF culture were quantified using a haemocytometer . For bloodstream forms ( BSF ) cultures , endothelial cells were used as a feeder cell layer . BAE were seeded in 24-well plates . After one day , the medium was removed , cells were washed with PBS ( phosphate-buffered saline: 137 mM NaCl , 10 mM Phosphate , 2 . 7 mM KCl , pH 7 . 4 ) and incubated in 1 ml of BSF specific medium . BSF were cultured in 2 different media , the TcBSF-3 was used during the adaptation phase from the infected mouse blood or from metacyclic differentiation and the TcBSF-1 was used for well-established cultures . BSF were always maintained at 34°C in a humidified atmosphere containing 5% ( vol/vol ) CO2 . Parasites collected from infected mouse blood could be either differentiated in PCF or cultivated as BSF . Cultures were initiated with tail blood of infected mice when parasitaemia reached a minimum of 108 parasites/ml . 5 drops of tail blood were collected in 1 ml of TcPCF-1 or TCPCF-2 media for differentiation in PCF or in 1 ml of TcBSF-3 for BSF culture . A few derivatives of the TcBSF-3 medium were also used during the adaptation medium development phase by using horse , foal or lamb sera ( Invitrogen ) instead of goat serum . Centrifugation ( 200×g , 1 min ) was used to remove the majority of red blood cells . Parasites present in the supernatants were subsequently incubated in 24 well plates at 27°C for PCF , and at 34°C in a humidified atmosphere containing 5% ( vol/vol ) CO2 for BSF . For PCF , half of the medium was replaced as soon as the well-shaped parasite density in the upper part of the well reached the middle log phase ( ≈5 . 106 cells/ml ) . This process could take a few days to a few weeks depending on the strain . Following this adaptation phase , half of the medium was changed every 2 days . pLEW13 , pLEW29 with a modified integration site ( here named pLEW29c ) and p2T7Ti/αTUB were kindly provided by J . E . Donelson [11] . pLEW20 is a typical expression vector for T . brucei [20] , which allows inducible expression of proteins under the control of a strong pol I promoter regulated by two tetracycline operators . To facilitate the integration and the subsequent expression of reporter proteins in T . congolense , the integration site , the rRNA spacer region ( 1 to 121 ) , was replaced by the βTUB region ( 1 to 134 ) of the pLEW13 vector . A PCR fragment from the pLEW13 was produced using specific primers containing ScaI and NdeI restriction sites respectively , sub-cloned in pGEM-T vector ( Promega ) , and inserted in the pLEW20 ScaI-NdeI restricted plasmid . This new vector , named pLEW20c , was used to produce the enhanced green fluorescent protein ( EGFP: optimized for expression and fluorescence in mammalian cells ) or the Renilla luciferase ( RLuc ) . PCR amplified fragments corresponding to EGFP and RLuc were inserted in the HindIII-BamHI restriction sites of the pLEW20c vector to produce pLEW20c-EGFP and pLEW20c-RLuc . Stably transfected strains were obtained with the Amaxa nucleofection method . The Amaxa Nucleofector® system ( Lonza , Levallois-Perret , France ) was used as described by the manufacturer . Briefly , a pellet of 107 parasites was resuspended in 100 µl of Basic Parasite Nucleofector® solution , mixed with 15 µg of NotI linearized plasmid and subjected to nucleofection with a specific program . Amaxa solutions and programs used are listed in Table 2 . Stably transfected trypanosomes were selected using BSF or PCF media supplemented with bleomycin for pLEW20c and p2T7Ti/αTUB , G418 for pLEW29c and hygromycin for pLEW13 . Increasing concentrations of antibiotics were used for BSF selection ( from 0 . 5 µg/ml to 2 µg/ml Bleomycin; 0 . 25 µg/ml to 1 . 25 µg/ml Hygromycin and 0 . 1 µg/ml to 0 . 5 µg/ml G418 ) , while a fixed concentration was used for PCF ( 2 µg/ml Bleomycin , 2 . 5 µg/ml G418 and 6 . 25 µg/ml Hygromycin ) . During BSF selection , half of the medium was changed every two days by carefully removing it without disturbing the adherent parasites . For PCF , the medium was not changed until the parasite density reached at least 5 . 106 cells/ml . After the selection phase , single clones were generated by limiting dilution , expanded and assessed by either PCR on genomic DNA ( using primers specific of the vector amplifying the resistance gene ) or luciferase activity quantification or GFP fluorescence depending on the vector used . Transgene induction was achieved by growing the cells in the presence of 1 µg/ml of tetracycline . The Renilla Luciferase Assay System ( Promega , Charbonnières-les-bains , France ) was used to measure in vitro luciferase activity as described by the manufacturer . Briefly , non-transformed and pLEW20c-Rluc-transfected T . congolense clones were grown to a total of 5 . 106 parasites/ml and centrifuged ( 1500×g , 10 min ) . The pellet was washed with PBS , resuspended in 20 µl lysis buffer and added to 100 µl of the reaction mix . Renilla luciferase activity was monitored over 5 min every 10 sec after substrate addition by using an Optima microplate reader ( BMG Labtech , Germany ) and expressed as relative light units ( RLU ) per µg protein . PCF were collected by centrifugation ( 1600×g , 10 min ) of culture supernatants . Individual EMF were collected by brief washing of the adherent cells with PBS and gentle scraping of the colonies with a cell scraper followed by centrifugation ( 1600×g , 10 min ) . MCF were purified from culture supernatants by DE52 anion-exchange column chromatography ( Whatman Plc . , Brentford , U . K . ) [19] . BSF were collected by extensive flushing of the culture to remove adherent cells followed by centrifugation ( 1600×g , 10 min ) . To observe EMF colonies , glass coverslips were placed in 24-well plates and incubated with EMF until formation of colonies ( the process can take a few days ) . To analyze BSF/BAE interactions , BAE cells were first incubated for at least six hours on glass coverslips placed in 24-well plates to allow them to adhere before addition of the BSF and co-culturing for several hours . Cells were fixed with formaldehyde as described [21] . Slides were then incubated with mouse anti-paraflagellar rod ( PFR ) antibodies ( anti-PFR2 L8C4 monoclonal antibody diluted 1∶100 ) , kindly provided by K . Gull , ( Oxford , U . K . ) ) in PBS 0 . 1% ( vol/vol ) Triton X-100 0 . 1% ( wt/vol ) , BSA for 30 min , washed three times with PBS and incubated for 30 min with Alexa Fluor 488 goat anti-mouse IgG secondary antibody ( diluted 1∶100 ) ( Molecular probes ) . Finally , cells were incubated for 5 min with 1 mg/ml 4′ , 6′-diamidino-2-phenylindle ( DAPI ) and mounted in Vectashield ( Vector Laboratories ) . Cells were observed with a Zeiss UV microscope and images were captured using a MicroMax-1300Y/HS camera ( Princeton Instruments ) and Metaview software ( Universal Imaging Corporation ) . Direct observations of parasites in culture were also conducted: cells were observed with an Axiovert ( Zeiss ) UV microscope and images were captured by an Axiocan camera ( Zeiss ) and Axiovision 3 . 0 software ( Zeiss ) . Total protein preparations of trypanosomes were obtained by lysis of live parasites with 2% ( wt/v ) SDS ( 10 µl per 106 parasites ) and heating at 100°C in the presence of a protease inhibitor cocktail ( complete Mini , EDTA-free; Roche Diagnostics , GmbH ) . 10 µl of this lysate was loaded per well and separated by SDS-PAGE ( 12% ) before transfer onto polyvinyl difluoride ( PVDF ) membranes ( Immobilon-P , Millipore ) and processed for Western blotting as described previously [22] . Monoclonal antibodies ( mAb ) #491 and #3C6 were kindly provided by T . W . Pearson and B . Loveless ( Victoria , Canada ) . mAb #491 recognizes a carbohydrate epitope present on the protease-resistant surface molecule of T . congolense and on other surface molecules of T . congolense and T . simiae [12] , [23] . mAb #3C6 was made by B . Loveless against recombinant CESP procured by N . Inoue ( B . Loveless and T . W . Pearson , unpublished data ) . Membranes were incubated for 2 hours with dilutions of hybridome culture supernatant ( 1∶10 ) or mouse anti-T . brucei tubulin ( 1∶2 , 000 ) . Antigen-antibody interactions were revealed with the Immobilon Western chemiluminescent horseradish peroxidase substrate ( Millipore ) . Eight-week-old female Balb/c mice were purchased from Charles River Laboratories ( L'arbresle Cedex France ) , NOD/SCID ( NOD . Cg-Prkdescid Il2rgtm1 Wjl/Szj ) were bred locally in specific pathogen-free conditions and used for experiments from five to six weeks of age . All animal studies adhered to protocols approved by the University of Bordeaux 2 animal care and use committee and the commission de genie genetique ( Direction Generale de la Recherche et de l'Innovation ) . Balb/c mice either treated or not with the immunosuppressant cyclophosphamide ( 200 mg/kg , Sigma-Aldrich ) or NOG mice were inoculated by intraperitoneal injection with variable amounts of infectious trypanosomes . Inoculated parasites were: ( i ) purified MCF , ( ii ) a mix of EMF and MCF from cell culture supernatants , ( iii ) BSF cultivated in vitro or from infected mouse blood . Parasitaemia was subsequently monitored daily by microscope observation . Blood samples were collected three times a week by tail bleed in 100 µl capillary tubes coated with Na-heparin . To determine the packed red blood cell volume percentage ( PCV ) , capillary haematocrit tubes were sealed , centrifuged ( 15 , 000 rpm , 12 min ) and analyzed with the haematocrit reader supplied by the manufacturer ( Heraeus ) .
T . congolense PCF adaptation to an axenic culture system directly from infected mice blood was described previously [24] . However , with described methods and media , for most of the isolates , differentiation efficiency was low and required several months to attain stable culture . Furthermore , some strains could not undergo the adaptation process to axenic culture . Therefore , we aimed to standardize a simple method for rapid initiation from infected blood and subsequent stabilization of PCF in continuous culture . Blood samples were collected from mice experimentally infected with various strains ( IL3000 , ILC-49 , IL1180 , TREU1457 , TRUM183:13–29 , STIB910 ) . When the parasitaemia reached the mid log phase ( ∼5 . 106 cells/ml ) , a few drops of tail blood were placed in 1 ml of media ( TcPCF-1 , TcPCF-2 or TcPCF-3; Table 1 ) and inoculated in 24-wells plates at 27°C after discarding most of the red blood cells by centrifugation . TcPCF media are derived from the 109-c medium [24] with variation in serum nature and percentage , and presence of cis-aconitate and citrate . BSF were always transformed into PCF within 2 days but degenerated cells ( abnormal shape , giant or clustered cells ) , unable to divide , appeared soon after and settled at the bottom of the well . These abnormal cells constituted the major part of the culture during the first 3–10 weeks before giving way to well-shaped dividing PCF . This adaptation phase is the critical step . When the culture reached at least 5 . 106 cells/ml , the medium was changed with 1 ml of fresh medium . After a few days , the doubling time reached its maximum ( 8–15 hours depending on the strain ) and the culture was transferred to 25 cm2 flasks and standard cultivation methods in TcPCF-3 medium were applied . The presence of 3 mM of cis-aconitate and citrate shortened the adaptation phase to at least 2 weeks , probably facilitating the differentiation process as described for T . brucei [25] , [26] . Furthermore , this adaptation phase could only be conducted in TcPCF-2 ( goat serum instead of fetal calf serum ) for ILC-49 and TRUM183:13–29 . In contrast , all the other tested strains preferred the TcPCF-1 medium . For the multiplication step , all strains grew faster in TcPCF-1 . The length of time required to achieve a stabilized PCF culture in flask is reported in Table 3 for all the strains tested . We observed that even with this improved system , the time period required to achieve the process was strain dependant . Transgene expression and RNA interference ( RNAi ) in T . congolense PCF have been reported previously [11] , [27] . Specifically , the tetracycline inducible system first used in T . brucei ( constitutive expression of RNA polymerase and tetracycline repressor , [5] ) has been adapted for T . congolense by Inoue and colleagues in the TRUM183 uncloned serodeme [11] . Nevertheless , classical electroporation methods are not sufficient to ensure an acceptable success rate of transfection in several other strains ( our data ) . For example , the most used and currently sequenced ( Sanger Center ) IL3000 strain is awkward for genetic manipulation ( upon our data: less than 30% of electroporation assays lead to selection of stable transfectants ) . Because transfection efficiency could be greatly enhanced in the parasite T . brucei using the Amaxa nucleofection method [28] , we applied this technology to T . congolense . To test this method on T . congolense PCF , we tried 3 different nucleofection programs and 3 different Amaxa solutions recommended by the manufacturer with different vectors as described in material and methods: p2T7Ti/αTUB as RNAi control experiment , pLEW20c-EGFP and pLEW20c-Rluc allowing inducible transgene expression of EGFP , Renilla luciferase , pLEW13 and pLEW29c inserted successively in the IL3000 genome to devise a tetracycline inducible system . The new cell-line was named IL3000:13–29 . Stable transfection achievement has been assessed by PCR on genomic DNA and by GFP fluorescence or luciferase activity measurement . Results are reported in Table 2 . The best results were obtained with the Parasitic 2 Amaxa solution and the X-001 program: independently of the T . congolense strain used , nearly 100% of the transfection assays led to selection of stable transfectants . With the other programs , electroporation led to cell death at much higher levels . The inducible operating system was also valued as shown in Figure 1 , 2 and Table 4 . Despite the absence of tetracycline , a basal expression level of GFP fluorescence and luciferase activity was observed; this is probably the result of a leakage of the tetracycline inducible system , which has already been described previously [20] , [29] . Upon addition of tetracycline , transgene expression was enhanced as indicated by the rise in fluorescence intensity ( Figure 1 ) and the 4-fold increase of luciferase activity ( Table 4 ) . For the RNAi experiment , as expected , the α-tubulin gene expression generated an altered cell morphology called the FAT phenotype [30] ( Figure 2 ) . Without tetracycline , this phenotype is absent , after tetracycline addition , the entire population displayed the lethal FAT phenotype within 48 hours . The absence of apparent leakage in this experiment is probably due to the natural enrichment of the culture with more tightly repressed cells due to the lethality associated with the FAT phenotype . This hypothesis is reinforced by prolonged selection time for p2T7Ti/αTUB transfectants and decrease of the doubling time in the course of the selection process . Cultivation of all the life cycle stages from the tsetse vector has been described previously [8] , [9] . However , using these protocols , production of infective MCF from insect form cultures could take few weeks to several months depending on individual strains of T . congolense [13] . The differences lay in cytoadherence timing , interval between attachment , infective MCF appearance and in the number of MCF produced . In order to shorten the metacyclogenesis process . We aimed to improve the differentiation efficiency in terms of timing and standardization . First , to asses timing differences among strains , we tried to perform metacyclogenesis using the described methods with the wild type strains IL3000 , ILC-49 , IL1180 , TREU1457 , STIB 910 and the genetically modified cell lines TRUM183:13–29 and IL3000:13–29 transfected with either pLEW20c-Rluc , pLEW20c-EGFP or p2T7Ti/αTUB . PCF were inoculated with 5 . 106 cells/ml in TcPCF-3 medium containing the optimal concentrations of glutamine and proline ( 2 mM glutamine and 10 mM proline ) for metacyclogenesis as described by C . Ross [10] . These conditions were verified to be optimal ( data not shown ) . After inoculation , half of the medium was changed every 2 days . To appreciate metacyclogenesis development , we considered several criteria to characterize the EMF and the MCF . ( i ) The first observed event was the EMF typical cytoadherence: parasites attached to the bottom of the flasks , grew and formed adherent bundles . ( ii ) Cell shape , size and kinetoplast and its attached flagella localization in relation to the nucleus change during metacyclogenesis process . These main morphological features were examined microscopically . For kinetoplast positioning , cells were labeled with a flagellar marker ( anti-PFR antiserum ) and stained with DAPI for orientation . ( iii ) MCF emergence was highlighted by purification on DE52 anion exchange chromatography . MCF infectivity assay in mice is described bellow . ( iv ) Stage specific markers were analyzed by western-blot , using two specific monoclonal antibodies . According to the literature , mAb #491 recognizes carbohydrate epitope shared by different surface molecules ( such as PRS , GARP and CESP ) differentially expressed during metacyclogenesis and mAb #3C6 binds to the EMF specific protein CESP [12] , [23] . Results are presented in Figure 3 and Table S1 . We confirmed that metacyclogenesis was successful as we chiefly observed the kinetoplast repositioning through the cycle ( kinetoplast is always anterior to the nucleus in the EMF contrary to being on the posterior end in PCF and MCF ) and MCF appearance within 1–2 weeks after cytoadherence . Furthermore , the expression pattern of stage specific markers is in agreement with published data on in vitro differentiated EMF: CESP is specifically induced in EMF stage and carbohydrate epitope presence is strongly increased during the PCF to EMF differentiation [12] . Nevertheless , we observed a large discrepancy among wild type strains with , for example , a required period of 3 to 10 weeks to observe cytoadherence for wild type strains . Moreover , differentiation was more difficult for genetically modified parasites since MCF getting required several months . For all the strains , the critical step seemed to be EMF differentiation . Subsequently , we focused on the method to optimize this step by using the IL3000 strain . First , we attempted to influence the physiological state of the parasite by modifying the cell density of the culture . Unexpectedly , we observed a significant increase in the rate of cytoadherence ( few days instead of weeks ) when a density higher than 2 . 107 cells/ml was used , corresponding to the stationary phase . All the strains were subjected to this treatment and EMF differentiation followed by MCF differentiation was noted ( results are reported in Table 3 ) . With this treatment , all the strains were able to pass through metacyclogenesis and parasites fit all the stage specific features described above , but variations in EMF differentiation timing were observed . Conversely MCF always appeared within 2 weeks . One of the features of stationary phase is depletion of metabolites in the medium . Taking this into consideration , we used a medium containing only base powder , glutamine , proline and no serum , TcEMF-1 . Surprisingly , in this medium , individual cells and bundles of adherent IL3000 cells appeared after 30 minutes and reached a maximum after 2 hours . Prolonged exposure time in the absence of serum resulted in cell death . Therefore , after 1–2 hours , we added 10% FCS in this medium to produce TcEMF-2 . Half of the medium was then changed every two days and it was found that MCF always appeared after 1–2 weeks . Adherent cells induced by serum depletion ( named iEMF for induced EMF ) displayed the main EMF criteria ( cytoadherence , shape , kinetoplast positioning and ability to differentiate in MCF ) but exhibited differences in molecular markers ( Figure 3B ) . Indeed , no signal was observed either with mAb #491 or #3C6 , which implies that CESP is not expressed and the carbohydrate epitope is absent . This rapid method was tested on all the strains and worked even for transgenic strains: iEMF always appeared within 1 hour and MCF within 1–2 weeks . To confirm the potential use of the process to conduct functional analysis through the life cycle , GFP fluorescence , luciferase activity and FAT phenotype were determined in transfected strains either induced or non induced ( Figure 1 , 2 and Table 4 ) . As expected , the results are similar to those of PCF . This process provides the opportunity to test phenotypes rapidly and efficiently throughout the parasite's developmental stages of the tsetse fly . To assess infectivity and virulence of in vitro MCF , we injected 107 parasites into Balb-c mice and measured the level of parasitaemia . The different strains were not equally infective for mice ( Table 3 ) . IL3000 , TREU1457 and STIB910 produced an acute parasitaemia with a single peak , a high parasite load , a strong anaemia and death within 2 weeks ( Figure S1 ) . IL1180 resulted in a chronic infection with a fluctuating blood parasitaemia leading to a survival time greater than one month ( Figure S1 ) . As for ILC-49 and transgenic cell lines , MCF were not infective and consequently no parasites were observed in the blood . Previous parasitological studies associated T . brucei virulence differences with the capacity of sub-cellular membrane fractions to induce immunosuppression [31] . For this reason , we used immunosuppressed mice ( cyclophosphamide pretreated Balb-c ) and immunodeficient mice ( NOD/SCID ) in order to by-pass the lack of infectivity and allow these strains to develop in the murine host ( Figure 4 ) . MCF resulted in parasitaemia only in immunodeficient mice with a long prepatent period ( at least 1 month ) . On the other hand , once parasites appeared , we observed a high parasite load , development of anaemia and death within two weeks . Subsequently , 107 BSF containing blood was injected into new Balb-c mice pretreated or not with cyclophosphamide or into NOD/SCID mice . We obtained similar results as the parasitaemia only developed in NOD/SCID mice albeit with a shorter prepatent period ( after 2 weeks ) . During the second passage , BSF were observed not only in NOD/SCID mice but also in immunosuppressed Balb-c mice . Following the third passage , blood samples were infective in immunocompetent Balb-c mice . For the fourth passage we observed the same infectivity and virulence through passages resulting in the development of acute parasitaemia with parameters comparable to those obtained with blood BSF stocks never cultivated in vitro ( field isolates directly adapted to rodents ) ( Table 3 ) . To gain a better understanding of this phenomenon , we examined the possibility of a requirement of a host adaptation step for these non-infective MCF . Indeed , it has been described that MCF of the same T . congolense clone produced by tsetse fed with various host blood displayed virulence differences in mice and required an adaptation period to develop in mice [32] . Furthermore some T . vivax field isolates have been adapted to rodents using intraperitoneal inoculation of goat serum [33] . We therefore , injected 500 µl of goat serum intraperitoneally in immunosuppressed Balb-c mice one day before parasite infection and every subsequent day until BSF were observed in the blood . We observed parasites as soon as one week after MCF inoculation and mice eventually developed an acute parasitaemia . This protocol did not work with immunocompetent mice . The same treatment was required for the first passage in mice . From the second passage , goat serum injection was not essential and blood samples were infective in immunocompetent Balb-c mice . These data confirmed the existence in some strains of an adaptation process to the murine host . We aimed to develop long-term cultures of T . congolense BSF in order to conduct drug trials and genetic modifications . We followed the described procedures to obtain BSF from infected mice blood [14] . Dividing BSF were easily obtained but parasites soon began to degenerate and the culture collapsed after 3 to 10 days depending on the strain . We investigated the consequences of varying media components such as serum and base powder . We inoculated standard T . brucei BSF medium containing either IMDM or MEM base powder and derived media containing 20% of various sera ( horse , foal , lamb , goat ) as described earlier [34] . The results showed a preference for goat serum and MEM base powder ( data not shown ) . Supplementation of these media with “serum plus” increased cell density and attachment ( TcBSF-1 ) . We also examined temperature effect by cultivating parasites between 34°C and 37°C and observed higher longevity at 34°C . But despite these improvements , the culture conditions were not sufficient for long-term culture . To go one step further with the culture system , we used endothelial cells as feeder cell layer . Indeed , strong interactions between vascular endothelium have been demonstrated in vivo [35] , [36] and in vitro with BAE [37] , [38] . Addition of BAE greatly improved the BSF condition as cells promptly interacted with BAE , proliferated faster and could be cultivated for 3–4 weeks . Nevertheless , this was not sufficient for the selection process of stable transfectants since some major drawbacks remained: cultures were very sensitive and preservation through freezing was not possible . Eventually , we managed to settle BSF adaptation by adding RBC lysate to the culture ( TcBSF-2 ) , and by using fresh goat serum provided directly after sampling thus avoiding alteration through preservation or treatment ( chemical or irradiation ) of the commercial serum ( TcBSF-3 ) . Therefore , the adaptation process had to be carried out in TcBSF-3 , this phase lasted few weeks before stabilization of the culture . All the strains tested could be cultivated for at least a month but stabilization has been achieved only for IL3000 , IL3000:13–29 and STIB 910 . When the culture is stable , commercial goat serum can be used and RBC lysate and serum plus were no longer essential to sustain growth . To test our BSF culture system , we compared growth of T . brucei and T . congolense in different media ( Figure 5 , Table S2 ) . RBC lysate and haemoglobin slightly improved growth rates for both species and the absence of reducing agents affected T . brucei to a greater extent than T . congolense . The absence of BAE in T . congolense culture had a great impact on growth rate , and furthermore , the culture was more sensitive as it was not possible to inoculate with less than 2 . 105 cells/ml . It was also observed that in the presence of BAE T . congolense reached the stationary phase later and supported a higher cell density than T . brucei ( 2 . 107 cells/ml vs . 4 . 106 cells/ml ) . In addition , we subjected T . congolense MCF to differentiation with this adaptation system in order to directly obtain BSF in vitro avoiding the murine step . We obtained similar results: all the strains differentiated in BSF and were cultivated for one month and we completely stabilized IL3000 , IL3000:13–29 and STIB910 for long-term culture . Infectivity and virulence of in vitro cultured BSF were tested by inoculating mice with parasites cultured for several months . Identical results were obtained as for MCF ( Table 3 ) . To complete the in vitro cycle , the cultivated BSF were subjected to PCF differentiation as described earlier . All the strains tested passed through this step and gave rise to a stable PCF culture . Finally , we cultivated transfectant strains and managed to observe the corresponding phenotype: GFP fluorescence , luciferase activity and FAT phenotype ( Figure1 , 2 and Table 4 ) . This experiment proved unambiguously the ability to fulfill phenotype analysis of genetically modified cell lines during the complete life cycle of T . congolense . In the same way as PCF , we used the Amaxa nucleofection technology to transfect BSF . It turned out that BSF transfection was very difficult . Results are summarized in Table 2 . We tried to construct the inducible system based on the tetracycline repressor directly in BSF and managed to achieve the first step , which was the integration of the pLEW13 vector into the genome . However , we failed to insert the second vector , pLEW29c , despite numerous endeavors . We then attempted to perform a constitutive RNAi experiment or reporter gene expression . We obtained successful results with the pLEW20-Rluc vector and highlighted luciferase activity ( Table 4 ) . We also achieved inducible expression of luciferase by directly transfecting the IL3000:13–29 cell line produced from the in vitro cycle . Direct transfection of T . congolense BSF is therefore possible but remains arduous as it took several weeks and the success rate was low .
The development of a continuous culture system represents a major breakthrough in T . congolense research . Indeed , parasite biology studies imply understanding of its complex life cycle including differentiation , environment adaptation , signaling , virulence and pathogenesis , which become accessible once powerful genetic tools such as RNAi are exploitable throughout the life cycle . For the first time , we have conducted a systematic study , which allowed us to define the best in vitro culture conditions for each stage in terms of differentiation efficiency followed by long-term in vitro culture design . In this way , we succeeded in reproducing the complete in vitro life cycle allowing the study of genetically modified cell-line phenotypes throughout the cycle ( Figure 6 ) . Firstly , we designed a new enriched adaptation medium which enabled the isolation of BSF directly from the blood of infected mice and their long-term in vitro culture , which makes them amenable to in vitro drug trials , preservation through freezing and transfection assays . However , transfection of BSF remains a difficult task as we were unable to perform direct RNAi experiments but did succeed in overexpressing the luciferase reporter gene . Secondly , we improved the efficiency of PCF transfections and set up a tetracycline inducible system for the T . congolense reference strain ( IL3000 , the strain used by the Sanger Center for the genome sequencing project ) . We validated this system by the inducible expression of reporter genes ( GFP and Luciferase ) and an inducible RNAi experiment targeting the essential tubulin gene as a control . Thirdly , we standardized metacyclogenesis techniques , not only on wild type isolates , but also on genetically modified parasites ( either for transgene expression or RNAi ) and the time needed to achieve the differentiation has been dramatically shortened to ∼10 days instead of months . Finally , BSF differentiation could either be achieved by infecting mice or directly in vitro . Then , PCF transfection followed by metacyclogenesis became an efficient , rapid and reliable way to study gene function in all the developmental stages as well as during the differentiation steps . For the first time , powerful tools such as reverse genetics ( gene expression inactivation and characterization of resulting phenotypes ) , becomes now available for this devastating parasite . Adaptation of trypanosomatids to rodents for experimental infections was essential to analyze the host-pathogen interaction [39] . Specifically the use of inbred mouse models have proven to be a valuable tool in pathology research [40] . Distinctness in mouse susceptibility to trypanosomiasis among the mouse strains is always observed . For example , for T . congolense and other trypanosome species , BALB/c mice ( compared to C57Bl/6 ) are the most susceptible [41]–[43] . Discrepancy in virulence among T . congolense isolates has also been reported in ruminant livestock as well as in laboratory animals [44] , [45] . To go one step further in the understanding of virulence mechanisms , it is essential to analyze the phenotype of genetically modified parasites during experimental infection . This is the reason why we assessed the impact of long-term in vitro culture and of minor genome modification on parasite infectivity and virulence . Results showed that BALB/c mice are less susceptible to transfected cell-lines as well as to 1 of 5 tested wild type savannah strains after a complete in vitro life cycle . This deficiency in infectivity can be overcome by using immunodeficient animals or goat serum injected immunosupresssed mice . Furthermore , infectivity levels return to normal after few passages in the murine host . The survival time of infected mice is always correlated with the parasitaemia peak arising and since the parasites developed , we concluded that disease parameters like anemia are not influenced by in vitro passage . These data indicate that while in some cases the ability of the parasite to multiply and be maintained in the host is lowered , parasite virulence is not affected . We propose that infectivity decrease relies on a host adaptation step . Development of parasitaemia in the mammalian host implies for the parasite to fit the new environment in terms of energy supplying and trypanocides conditions . Impairment of the immune system could allow a time period sufficient for the parasite to adapt to its new physiological state . Goat serum injections might provide the parasite with essential factors like nutriments and create a microenvironment promoting the emergence of host adapted parasites that are able to develop in mice blood . Such phenomenon has been suggested for T . vivax adaptation to laboratory rodents using multiple passages in irradiated rats and intraperitoneal injection of goat serum [33] . This idea is in harmony with the requirement of goat serum supplemented medium for T . congolense BSF in vitro culture . Composition of this serum must suit the specific needs for parasite growth . The concentrations of vitamins and related metabolites vary from one sera to another . Goat serum is one of the most commonly used to enhance cell multiplication [46] , [47] . The greatest efficiency of fresh goat serum ( not treated ) reinforced the hypothesis that the presence of labile components is required for the parasite well-being . One can also notice that the 3 strains impaired for infectivity ( ILC-49 , TRUM183:13–29 and IL3000:13–29 ) are those which need goat serum to promote their in vitro differentiation in PCF . These findings illustrated that even for closely related strains ( all are of the Savannah genetic group ) , subtle differences in physiological requirements might exist . Another important factor to promote BSF in vitro growth is the presence of the erythrocyte lysate , which became dispensable after some time . Haemoglobin is a component of red blood cells and we showed that its presence in culture medium enhances the growth rate of the parasite cultures . The first culture media described for Trypanosoma spp . were always supplemented with blood lysate , then replaced by hemin [48] . However , neither haemoglobin nor hemin , are sufficient to ensure BSF adaptation from infected mice blood ( data not shown ) . Hemin is also essential for trypanosomatids PCF differentiation and growth [24] , [49] . The parasite's inability to synthesize heme explains this requirement , since heme containing proteins like cytochrome c [50] , are essential for parasite viability . However , erythrocyte lysate might provide the parasite with other unknown components , which are essential during the adaptation phase . Generally , T . congolense culture has to mimic the host environment . Thus , a high amino acid concentration in insect forms media reflects the high proline content in tsetse hemolymph subsequently used for energy metabolism of the parasite . An even more striking similarity is the use of BAE layer to cultivate BSF . Indeed , in infected animals , T . congolense BSF are found adhering to erythrocytes and to endothelial cells of the microvasculature by their flagellum [35] , [36] , [51] . This requirement is no longer essential after an adaptation phase . Finally , production of infective MCF requires strong adherence of EMF mimicking the bundles formed in the fly proboscis . However , parasite journey in fly midgut then in proboscis is complex and can't be fully reproduced in vitro . Especially coat antigens expression is refined and follows a defined kinetic in the different insect stages . For example , one of the major coat protein , GARP , is weakly expressed in early PCF , present in the midgut then downregulated before being strongly expressed in EMF isolated from proboscis [23] . In culture , this marker is expressed in both stages PCF and EMF , reflecting an artefactual absence of regulation during in vitro EMF differentiation [23] . Furthermore , Butikofer et al . [23] did not find reactivity with mAb #491 on EMF from tsetse , although Sakurai et al . [12] did find increased reactivity with EMF compared to PCF within in vitro culture . Similarly , the iEMF phenotype ( absence of CESP expression and of the carbohydrate marker ) may represent another illustration of the importance of the environmental factors on differentiation . Indeed , the difference observed in surface markers may result from the absence of serum at the time of differentiation or from the stress induced by the sudden environment change . Nevertheless , despite the absence of those markers iEMF are able to differentiate into infective MCF suggesting that those surface markers expression is not essential for in vitro EMF differentiation but must be involved in the in vivo process where parasites have to attach themselves to the wall of the labrum in the food canal to transform into EMF . The iEMF stage could also represent a transitory stage in the insect . Furthermore the absence of CESP expression does not hinder in vitro iEMF adhesion implying that this protein is not essential for cytoadherence and that other proteins might be involved . Successful culture and differentiation of iEMF and EMF combined with targeted gene deletion through RNAi represent powerful tools to understand the elaborated mechanism of the metacyclogenesis process encountered in the insect . Parasites display a great ability to adapt but the required time period for this adaptation is the limiting step in the initiation of stable culture of the different developmental stages as well as for experimental infections . We succeeded to go through this limiting step to secure cultures of all the T . congolense developmental stages . Standardization of media and methods using various strains , more or less demanding , allowed the design of the complete in vitro lifecycle of all the tested strains . Definition of these optimal conditions also greatly improved the success rate of PCF nucleofection , that should be very useful for routine genetics based analysis . To conclude , the ability to dominate the in vitro metacyclogenesis in combination with the transgenic PCF technique provides an essential tool to investigate the functional role of T . congolense genes throughout the cycle as well as in vivo during experimental infections in mice .
|
Trypanosoma congolense is a parasite responsible for severe disease of African livestock . Its life cycle is complex and divided into two phases , one in the tsetse fly vector and one in the bloodstream of the mammalian host . Molecular tools for gene function analyses in parasitic organisms are essential . Previous studies described the possibility of completing the entire T . congolense life cycle in vitro . However , the model showed major flaws including the absence of stable long-term culture of the infectious bloodstream forms , a laborious time-consuming period to perform the cycle and a lack of genetic tools . We therefore aimed to develop a standardized model convenient for genetic engineering . We succeeded in producing long-term cultures of all the developmental stages on long-term , to define all the differentiation steps and to finally complete the whole cycle in vitro . This improved model offers the opportunity to conduct phenotype analyses of genetically modified strains throughout the in vitro cycle and also during experimental infections .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"cell",
"biology/microbial",
"growth",
"and",
"development",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"microbiology/parasitology",
"infectious",
"diseases/protozoal",
"infections"
] |
2010
|
Complete In Vitro Life Cycle of Trypanosoma congolense: Development of Genetic Tools
|
PARP inhibition can induce anti-neoplastic effects when used as monotherapy or in combination with chemo- or radiotherapy in various tumor settings; however , the basis for the anti-metastasic activities resulting from PARP inhibition remains unknown . PARP inhibitors may also act as modulators of tumor angiogenesis . Proteomic analysis of endothelial cells revealed that vimentin , an intermediary filament involved in angiogenesis and a specific hallmark of EndoMT ( endothelial to mesenchymal transition ) transformation , was down-regulated following loss of PARP-1 function in endothelial cells . VE-cadherin , an endothelial marker of vascular normalization , was up-regulated in HUVEC treated with PARP inhibitors or following PARP-1 silencing; vimentin over-expression was sufficient to drive to an EndoMT phenotype . In melanoma cells , PARP inhibition reduced pro-metastatic markers , including vasculogenic mimicry . We also demonstrated that vimentin expression was sufficient to induce increased mesenchymal/pro-metastasic phenotypic changes in melanoma cells , including ILK/GSK3-β-dependent E-cadherin down-regulation , Snail1 activation and increased cell motility and migration . In a murine model of metastatic melanoma , PARP inhibition counteracted the ability of melanoma cells to metastasize to the lung . These results suggest that inhibition of PARP interferes with key metastasis-promoting processes , leading to suppression of invasion and colonization of distal organs by aggressive metastatic cells .
Metastatic melanoma is a fatal malignancy that is remarkably resistant to treatment; however , the mechanisms regulating the transition from the primary local tumor growth to distant metastasis remain poorly understood . Metastasis , defined as the spread of malignant tumor cells from the primary tumor mass to distant sites , involves a complex series of interconnected events . Understanding the biochemical , molecular , and cellular processes that regulate tumor metastasis is of vital importance . The metastatic cascade is thought to be initiated by a series of genetic alterations , leading to changes in cell-cell interactions that allow the dissociation of cells from the primary tumor mass . These events are followed by local invasion and migration through proteolitically modified extracellular matrix ( ECM ) . To establish secondary metastatic deposits , the malignant cells evade host immune surveillance , arrest in the microvasculature , and extravasate from the circulation . Finally , tumor cells can invade the local ECM , proliferate , recruit new blood vessels by induction of angiogenesis , and then expand to form secondary metastatic foci [1] . Several key steps in metastatic progression involve tumor-associated endothelial cells ( EC ) [2] . Both angioinvasion and angiogenesis require disruption of endothelial integrity for tumor cell transmigration across the endothelium , EC migration and EC access for mitogenic stimulation . An essential step in angioinvasion and angiogenesis is the disruption of the adherent junctions between EC . Vascular endothelial cadherin ( VE-cadherin; also known as cadherin 5 ) is the most important adhesive component of endothelial adherent junctions [3]; while ectopic expression of VE-cadherin in malignant melanoma cells confers this tumor the capability to form vessel-like structures that contributes to the lack of efficient therapeutic strategies and increases the risk of metastatic disease [4] . Epithelial-mesenchymal transition ( EMT ) is a trans-differentiation characterized by decreased epithelial markers such as E-cadherin[5] . EMT is a dynamic process resulting in the acquisition of cell motility with decreased adhesive ability for body organization that includes embryonic development and wound healing . Currently , EMT is thought to be a key step in the process of cancer metastasis [6] . Molecular markers of EMT include E-cadherin down-regulation , responsible for the loss of cell-cell adhesion , up-regulation of matrix-degrading proteases and mesenchymal-related proteins such as vimentin and N-cadherin , actin cytoskeleton reorganization , and up-regulation and/or nuclear translocation of transcription factors underlying the specific gene program of EMT , such as β-catenin and members of the Snail1 family [6] . The nuclear protein PARP-1 , known to function as a DNA damage sensor and to play a role in various DNA repair pathways , has recently been implicated in a broad variety of cellular functions , including transcriptional regulation [7] . PARP inhibitors exhibit antitumor activity in part due to their ability to induce synthetic cell lethality in cells deficient for homologous recombination repair [8] , [9] , [10] , [11] . PARP inhibitors also possess anti-angiogenic properties [12] , [13] , [14] , [15] , and recently , our group reported that PARP inhibition results in the down-regulation of Snail1 by accelerating the degradation of this protein [16] . In the present study we aimed to address the potential of PARP inhibition as modulators of metastasis [16] . The results presented here indicates that PARP inhibition , through down-regulation of the intermediary filament vimentin in both endothelial and melanoma cells , led to a reversion of mesenchymal phenotype in both cell types and prevented malignant melanoma cells from developing vasculogenic mimicry . As monotherapy , PARP inhibition displayed an anti-metastatic effect in a model of murine melanoma . Moreover , we identified vimentin as an upstream modulator of EMT: forced expression of vimentin was sufficient to induce tumor cell transformation through the ILK/GSK-3β signaling axis . The ability of PARP inhibition to modulate vimentin levels ( and hence EMT ) , the interference with vasculogenic mimicry , and the modulation of endothelial plasticity allowed PARP inhibitors to exert a multifaceted antimetastatic effect to counteract the progression of malignant melanoma .
A number of reports from various laboratories , including ours , have identified a novel and unexpected effect of PARP inhibitors on angiogenesis , raising the possibility that PARP inhibitors may be useful as anti-angiogenic agents [13] , [17] . In our present study , we disrupted PARP activation in HUVECs in an attempt to elucidate the mechanism by which PARP-1 influences endothelial cell dynamics . We have previously shown that PARP inhibitors reduced angiogenesis both in vitro and in vivo ( [13] and Figure S1 ) . To further characterize this effect of PARP inhibition on endothelial cell plasticity , we performed a proteomic analysis using primary HUVEC in the presence or absence of the PARP inhibitor DPQ ( Figure 1A , Figure 2 and Figure S2 ) . The expression levels of a number of proteins were altered following PARP inhibition , as detected by 2D DIGE electrophoresis ( Figure S2 ) and mass spectrometry analysis ( Figure 1A , Figure 2 ) . A statistically significant down-regulation of vimentin ( a class III intermediary filament ) , tropomyosin alpha-4 chain ( involved in stabilizing actin filaments ) , endoplasmin ( a molecular chaperone involved in processing and transport of secreted proteins ) , mitochondrial ATP synthase ATPB5 , protein disulfide isomerase PDIA6 , heat-shock 70 kD protein-5 ( glucose-regulated protein , 78 kD ) , heat shock protein 90 kDa alpha ( cytosolic ) , class B member 1 , and HSP90AB1 occurred following PARP inhibition . An increase in the expression of the mitochondrial heat shock protein HSPD1 was also observed after PARP inhibition . Due to its important role in the biology of endothelial cells , we focused our study on vimentin , the main structural protein of intermediary filaments . It has been reported that vimentin can be targeted for tumor inhibition due to its specific up-regulation in tumor vasculatures [18] , [19] . To confirm the results of our proteomic analysis , we performed western blot in HUVEC either treated with DPQ ( right ) or left untreated . In Figures 1B and 1C , western blot and indirect immunofluorescence analysis indicated that vimentin expression was down-regulated in HUVEC cells treated with DPQ . Figure 1B and 1D show that PARP inhibition affected not only vimentin levels but also Snail1 and VE-cadherin protein and mRNA levels . Endothelial to mesenchymal transition ( EndoMT ) is a process by which endothelial cells disaggregate , change shape , and migrate into the surrounding tissue . The process of endoMT is characterized by the loss of endothelial cell markers , such as vascular endothelial VE-cadherin , and the expression of mesenchymal cell markers , such as vimentin and Snail1 [20] . Endothelial cell migration was strongly inhibited by PARP inhibition ( Figure 1E ) . These results suggest that PARP inhibition prevented the acquisition of a mesenchymal phenotype by endothelial cells . Vimentin is a well-known marker of EMT , which is a hallmark of primary tumor progression to a metastatic phenotype . We tested the impact of vimentin down-regulation ( induced by PARP inhibition or vimentin silencing ) on EMT differentiation in various melanoma cell lines and in endothelial cells . One major event induced by PARP inhibition , in the process of EMT is the up-regulation of E-cadherin expression through the inactivation of the transcription factor Snail1 . Snail1 and vimentin levels were both down-regulated following PARP inhibition , indicating a disruption EMT in the absence of PARP activation ( Figure 3A in G361 cells and Figure S3B in B16-F10 cells ) . Down-regulation of PARP activity was confirmed in G361 following H2O2 treatment as a positive control of PARP-1 activation and poly ( ADP-ribose ) ( PAR ) synthesis ( Figure S4 ) . Vimentin and Snail1 mRNA levels were decreased after PARP inhibition ( Figure 3C and Figure S3C ) . In Figures 2B and Figure S3A , indirect immunofluorescence showed that vimentin expression was down-regulated in melanoma cells treated with DPQ or KU0058948 ( G361 cells , Figure 3B ) or PJ-34 ( B16-F10 cells , Figure S3A ) . Using two different luciferase reporter plasmids under the control of a Snail1 responsive sequence and the E-cadherin promoter , we found that PARP inhibition affected negatively the activation of Snail1 and activated the expression of the E-cadherin promoter ( Figure 3D and Figure S3D ) . Wound healing experiments also revealed decreased wound closing following treatment with a PARP inhibitor , PJ-34 ( Figure 3E ) . We have also evaluated the effect of both PARP-1 and vimentin silencing on the expression of Axl , a key determinant of cell migration and EMT promotion [21] . Following PARP-1 silencing in HUVEC or G361 cells , the EMT marker Snail1 decreased while E or VE-cadherin were upregulated ( Figure 4A and 4B respectively ) . Interestingly , Axl expression was also down-regulated in parallel with decreased levels of vimentin . Vimentin knockdown also caused a global alteration in the expression of EMT markers . Under these conditions , Axl levels were decreased ( Figure 4A and 4B ) , suggesting that vimentin down-regulation was sufficient to drive tumor cells toward an epithelial state . We next sought to determine if alterations in vimentin levels were sufficient to alter or reverse EMT progression . Vimentin is known to positively influence tumor cell migration . To test the impact of vimentin expression on cell migration and invasion we performed either silencing or over-expression in endothelial and melanoma cells . Following vimentin knockdown , wound healing closure in HUVEC cells was significantly diminished ( Figure 4C ) while its over-expression increased wound healing efficiency ( Figure 4D ) . The same approach was used in B16F10 melanoma cells where over-expression of vimentin increased significantly cell migration ( Figure 4E ) . Nonetheless , inhibition of PARP had a less impact on cell migration after vimentin over-expression , suggesting that the levels of vimentin were implicated in the effect of PARP inhibition on cell motility ( Figure 4E ) , although a multifactorial mechanism for downstream effect of PARP inhibition could not be excluded . To further confirm the role of vimentin in PARP-inhibitor-induced impaired cell migration we decided to analyze the effect of vimentin over-expression and PARP inhibition in a well-established model of epithelial cells , MDCK , that undergo EMT after hepatocyte growth factor ( HGF ) treatment , including fast movement and circularity ( scattering ) [22] . The trajectories of cell migration were determined under video-microscopy and analyzed using MetaMorph image analysis software . Global trajectories after expression of GFP-vimentin in the presence or absence of PARP inhibitor and HGF were determined . Treatment with the PARP inhibitor PJ-34 or olaparib resulted in decreased cell motility in cells transfected with empty GFP vector ( Figure 4F ) . Vimentin expression increased cell motility ( Figure 4F , right ) , and PARP inhibition was unable to prevent this increase , suggesting that vimentin down-regulation is needed for the effect of PARP inhibition in reversing the EMT phenotype . To characterize more in-depth the implications of vimentin expression in the context of EMT , we expressed GFP-vimentin in both a human melanoma cell line ( Figure 5A ) and a human breast tumor cell line with an epithelial phenotype ( MCF7 ) ( Figure 5B and 5C ) ; MCF7 cells were chosen due to the lack endogenous vimentin expression compared with melanoma G361 cells ( Figure 5A , G361 cells and Figure 5B and 5C , MCF7 cells ) . GFP-vimentin over-expression alone induced a mesenchymal phenotype characterized by Snail1 up-regulation , loss of E-cadherin , increased pGSK-3β ( inactive form ) and β-catenin expression ( Figure 5A , 5B and 5C ) . The most remarkable effect of PARP or vimentin silencing observed in our model was the down-regulation of ILK and GSK-3β ( Figure 4A and 4B ) . In order to get mechanistic information on the interaction between vimentin over-expression and the activation of EMT signaling pathway , we focused in the axis ILK/GSK-3β , which plays a central role in EMT commitment , upstream of Snail1 . Inhibition of GSK-3β was achieved by LiCl treatment while its activation was driven by silencing the kinase , ILK , which is the upstream inhibitory kinase for GSK-3β ( Figure 5 , central panel ) . Specifically , inhibition of GSk-3β ( which was confirmed by an increase in the level of inhibitory phosphorylation of GSK-3β at Ser9 ) with LiCl , activated EMT and resulted in E-cadherin down-regulation , Snail1 accumulation and increased levels of β-catenin ( Figure 5B ) ; concomitantly , E-cadherin was down-regulated following GSk-3β inhibition by LiCl ( Figure 5B ) or exogenous expression of vimentin ( Figure 5 ) . GSk-3β activation is achieved through the silencing of its upstream inhibitor integrin-linked kinase ( ILK ) . ILK knockdown resulted in Snail1 down-regulation and increased E-cadherin expression ( Figure 5C ) . Interestingly , exogenous vimentin expression completely prevented siILK-induced E-cadherin up-regulation and partially prevented the reduction of Snail1 expression . These results suggested that vimentin , when over-expressed , is sufficient to drive the phenotypic changes associated with a mesenchymal cell status , depending on the activation of GSk-3β , whose inhibition accentuated vimentin-induced changes , while its activation ( following ILK-silencing ) , abolished vimentin-induced E-cadherin decrease and Snail1 accumulation ( Figure 5C ) . The formation of patterned networks of matrix-rich tubular structures in three-dimensional culture is a defining characteristic of highly aggressive melanoma cells . It has been demonstrated that aggressive melanoma cells in which VE-cadherin was repressed , could not form vasculogenic-like networks [23] , suggesting that tumor-associated misexpression of VE-cadherin ( observed in melanoma cells ) is instrumental in allowing endothelial cells to form vasculogenic networks . We measured VE-cadherin protein levels in B16-F10 cells after treatment with the PARP inhibitor PJ-34 or KU0058948 . VE-cadherin expression was strongly down-regulated following PARP inhibition . We tried to confirm this result by indirect immunofluoresce of VE-cadherin , however the protein was barely detected , as was the case for the protein in western blot ( Figure 6A ) . Phosphorylation of VE-cadherin has been shown to correlate with loss of function of VE-cadherin and increased vascular permeability [24] , as is the case for pseudo vessels during VM . PARP inhibition was able to impact negatively on the levels of both total and phosphorylated VE-cadherin , which , indeed , had a membrane and cytoplasmic distribution ( Figure 6A ) . The consequences for the down-regulation of both total and phosphorylated VE-cadherin by PARP inhibitors during VM are now being investigated in our laboratory . VM was measured in vitro using B16F10 cells cultured in matrigel coated plates ( Figure 6B ) . All markers of VM structure formation ( covered area , tube length , branching points and loops ) were significantly decreased after inhibition of PARP with PJ-34 ( Figure 6C ) . We next aimed to examine the effect of PARP inhibition on melanoma tumor growth of cells subcutaneously implanted in C57BL/6 mice . Mice were treated every two days with 15 mg/kg ( i . p . ) of the PARP inhibitor DPQ or vehicle . A significant difference in tumor growth was found after 14 days of tumor implantation in the DPQ-treated group compared to the control ( Figure 7 and Figure S5A ) . To evaluate the direct effects of the PARP inhibitor DPQ on tumor metastasis , we used a well-characterized model of experimental lung metastasis [25] . Experimental metastasis model provide several advantages for investigation . The time course for model maturity is generally rapid , the biology of metastasis is reproducible and consistent , and we control de number and type of cells that are introduced to the circulation [26] . B16-F10 cells were tail vein injected into mice , and the mice were then treated with 15 mg/kg of the PARP inhibitor DPQ or vehicle three times per week over a three-week period . Tail vein injection results primarily in pulmonary metastases . Photon emission was acquired every two days . Seven days after B16-F10 cell injection , a photon signal was already detected in the lungs ( Figure 7B ) , and DPQ treatment significantly suppressed lung metastasis compared to the control throughout the duration of the experiment ( 21 days ) . Similar results were obtained using the clinically relevant PARP inhibitor olaparib ( Figure S6 ) . Metastatic foci were also detected in other organs upon mice autopsy . These organs included the liver , kidney , spleen , gut , stomach and heart ( Figure S5B ) . In all cases , the incidence of metastatic foci was reduced compared to lung metastasis . DPQ-treated mice exhibited a decreased incidence of extra-pulmonary metastasis compared to the control . Pathologic analysis of the lungs showed a decrease in size and number of metastatic foci ( more than 80% ) after DPQ treatment ( Figure 7C ) that was accompanied by a reduced number of tumor vessels in both primary subcutaneous tumors and lung metastasis ( Figure 7D ) , suggesting that the anti-angiogenic effect of PARP inhibition may be involved in the observed reduction in metastatic progression . Apoptotic and mitotic rate were not significantly different in tumors derived from DPQ-treated or untreated mice ( Figure S7 ) . To investigate in vivo the effect of PARP inhibition on the expression of Snail1 and E-cadherin , we performed immunohistochemistry for these EMT markers in metastatic lung tumors ( Figure 7E ) . We observed that Snail1 was highly expressed in the vessels of tumors derived from the untreated group . This expression exhibited both nuclear and cytoplasmic distribution as previously reported [27] . Metastatic lung tumors derived from DPQ-treated mice displayed reduced expression of Snail1 as well as an increase in E-cadherin expression , similar to the results obtained in cultured melanoma cells . These data indicate that the in vivo expression of EMT markers within tumors is also reduced following treatment with PARP inhibitor . We also performed a Kaplan Meyer curve to compare the mortality of both groups of mice , and we observed a statistically significant difference in the survival rate from <4 weeks in the untreated group to >8 weeks in the DPQ-treated mice ( Figure 7F ) . Survival of mice injected with B16-F10 cells stably expressing shRNA targeting PARP-1 ( Figure 7G ) , was also significantly increase . To determine the correlation between PARP-1 expression and disease progression in human melanoma , we used IHC to analyze the levels of vimentin , PARP-1 , Snail1 , E-cadherin and MITF in nodular and metastatic melanoma frozen biopsies . Vimentin was expressed in all biopsies derived from both nodular and metastatic melanoma; however , the level of expression was elevated in nodular melanoma , which is the initial stage of the disease . PARP-1 expression was positively correlated with vimentin expression , suggesting an association between the in vivo expression of both proteins ( Figure S8 , Table S1 ) . Expression of the Snail1 and microphthalmia-associated transcription factor ( MITF ) , which is a melanocyte marker , is also increased in metastatic melanoma . Interestingly , nodular melanoma did not express Snail1 while 40% of metastatic melanoma samples displayed Snail1 expression . Loss or reduction of E-cadherin and increased expression of EMT markers is frequently associated with the development of an invasive phenotype in cancer . Expression of E-cadherin in normal melanocytes is significantly reduced during the initial steps of melanoma progression [28]; however , elevated levels of E-cadherin are found at advanced stages of the disease [29] . E-cadherin expression was similar in both nodular and metastatic melanoma ( Table S1 ) , which is in agreement with previous publications . These findings suggest that in human melanoma , there is a complex interconnection between the expression levels of various disease markers and the expression of PARP-1 , although we have detected a strong correlation between vimentin and PARP-1 expression ( Figure S8 ) .
PARP inhibitors are a novel and important class of anticancer drugs , and there are now more than 40 clinical trials that are ongoing or in development to study the effectiveness of PARP inhibitors in the treatment of various cancers . Given the enormous interest in this target , it is important to understand the underlying mechanisms by which PARP-1 and other PARPs function in tumor cell biology . Until recently , the development of PARP-1 inhibitors has focused almost exclusively on the function of this enzyme in DNA repair . Emerging literature , however , indicates other activities of PARP-1 that may explain the in vivo potency of some PARP-1 inhibitors that cannot be entirely attributed to their apparent in vitro activity and that could provide additional targets for anti-cancer therapies . In addition to its direct role in DNA-damage recognition and repair , PARP-1 can regulate the function of several transcription factors , including p53 and NF-κB . In the context of certain cancers , PARP-1 interacts with the transcription factors HIF1 [13] and Snail1 [16] . The mechanisms underlying the effects of PARP inhibition on vascular plasticity and metastasis remain relatively unknown . Our current study identifies PARP-1 as a pivotal modulator of the molecular and functional changes characteristic of EndoMT ( involved in the loss of function of tumor-associated vessels ) and of the phenotypic switch that facilitates the acquisition of pro-metastatic capacities by tumor cells . Proteomic analysis of endothelial cells that have been treated with a PARP inhibitor identifies the intermediary filament protein vimentin as a target of PARP inhibition . Intermediary filaments such as vimentin and keratins are known to play non-mechanical roles in protein trafficking and signaling ( reviewed in [30] ) , which in turn influence cellular processes such as cell adhesion and polarization . Vimentin is abundantly expressed by mesenchymal cells and plays a critical role in wound healing , angiogenesis and cancer growth . Vimentin has also been described as a tumor-specific angiogenesis marker , and targeting endothelial vimentin in a mouse tumor model significantly inhibited tumor growth and reduced microvessel density [31] . Vimentin is both an EMT and EndoMT marker and is also over-expressed in tumor samples compared to normal tissues . This protein also contributes to tumor phenotype and invasiveness [18] , [19] . Our findings indicate that PARP inhibitors reduce the metastatic potential of melanoma cells , at least in part , through their ability to down-regulate vimentin expression . Vimentin expression has been shown to be transactivated by β-catenin/TCF and thus increasing the tumor cell invasive potential [19] . It has been shown that NF-κB , a key protein regulating the immune and inflammatory process , also plays an important role in regulating EMT process and its inhibition in the mesenchymal cells reversed the EMT process , suggesting the importance of NF-κB in both activation and maintenance of EMT [32] . Since vimentin is over expressed during EMT process , and NF-κB being one of the transcription factors binding to vimentin promoter , it would be tempting to speculate that this over-expression of vimentin is a result of activated NF-κB in cancer cells . Also , TGFβ1 response element was found within the activated protein complex-1 region of the vimentin promoter and was involved in regulation of vimentin expression in myoblasts and myotubes [33] . Interestingly , ADP-ribosylation of Smad proteins by PARP-1 has been shown to be a key step in controlling the strength and duration of Smad-mediated transcription [34] . Regulation of vimentin levels by PARP inhibition may also involve other transcription factor such as Snail1 and HIF-1/2 . Our results also reveal that vimentin levels are not merely a hallmark of EMT . While silencing of vimentin in melanoma cells can reverse the EMT phenotype , in part by promoting down-regulation of the protein kinase Axl that is involved in cell motility , forced expression of vimentin in tumor cells lacking this protein is sufficient to trigger the switch from epithelial to mesenchymal phenotype . GSK-3β is an upstream regulator of key factors involved in EMT such as Snail1 and β–catenin . We hypothesized that vimentin may be involved in the modulation of this upstream regulator of EMT . Indeed , vimentin expression potentiated LiCl- ( a GSK-3β inhibitor ) induced EMT ( Figure 5B ) and counteracted the inhibitory action of ILK-silencing ( leading to GSK-3β activation ) in the context of EMT ( Figure 5C ) . Mechanical signals can inactivate GSK-3β resulting in stabilization of β-catenin . Intermediate filaments are important in allowing individual cells , tissues and organs to cope with various types of stress , and they play a significant role in the mechanical behavior of cells [35] . It is possible that the signaling pathway that integrates PARP activation with altered vimentin expression and fluctuations in GSk-3β activity could be related to the capability of PARP inhibitors to inactivate AKT signaling [36] , which would result in GSk-3β activation and the modulation of its downstream signaling , ultimately resulting in the reversal of EMT . Vasculogenic mimicry , as a de novo tumor microcirculation pattern , differs from classically described endothelium-dependent angiogenesis . This is a unique process characteristic of highly aggressive melanoma cells found to express genes previously thought to be exclusively associated with endothelial cells and is characteristic of aggressive melanoma tumor cells . HIF-1α and HIF-2α , transcription factors that are stabilized during conditions of oxygen depletion ( hypoxia ) , are the master regulators of VE-cadherin . HIF-mediated transcriptional regulation during hypoxia is critical as this process induces genes that are essential for tumor cell adaptation to the stress of oxygen depletion . As a result , the expression of HIF target genes is associated with increased malignancy . Although the expression of VE-cadherin is not hypoxia-regulated , HIF-2α , but not HIF-1α , activates the VE-cadherin promoter by binding to the HRE during normoxic conditions [37] . HIF-2α expression is associated with developing endothelium , proper vascular development and increased tumor malignancy [38] , [39] , raising the possibility that it may be an important protein that functions in the induction of tumor cell plasticity . Using a mouse model of melanoma lung metastasis , we also present in vivo evidence indicating that targeting PARP strongly reduces metastatic dissemination of melanoma cells , at least in part through inducing a reduction in tumor microvessel density along with changes in the expression pattern of EMT markers ( Snail1 , vimentin and E-cadherin ) within the tumor . Snail1 is a master regulator of EMT , and the activation of this protein can mediate tumor invasiveness through the transcriptional repression of E-cadherin expression . Regulating the activity of E-cadherin repressors represents a potentially beneficial strategy to fight cancer progression , and PARP-1 inhibitors accomplish this function by interfering with Snail1 activation . Results from human tissue arrays of melanoma suggest a complex interaction between PARP-1 expression and melanoma progression . It is difficult to verify EMT experimentally in vivo due to the reversible and dynamic nature of the process . Although melanoma cells are not epithelial in nature , the EMT for this tumor is well characterized and the relevance of the cadherin switch has been previously described using several experimental approaches , demonstrating that melanoma cell lines transfected with N-cadherin are morphologically transformed from an epithelial-like shape to a fibroblast-like shape [37] . Adenoviral re-expression of E-cadherin in melanoma cells down-regulates endogenous N-cadherin and reduces the malignant potential of these cells [37] . Globally , our study shows that PARP inhibition is perturbing metastatic transformation at least at three levels ( Figure 8 ) : i ) decreasing abnormal tumor angiogenesis through its ability to counteract Endo-MT; ii ) preventing from acquisition of EMT and iii ) limiting vasculogenic mimicry in melanoma cells . Over the past few years , PARP has emerged as a strong and effective target for first line anticancer therapy . Due to its ability to regulate a number of cellular functions ( from DNA repair to cell death and transcription ) , inhibition of PARP may affect multiple facets of tumor metabolism . These findings strongly indicate that several novel activities of PARP-1 may contribute to the effects of anti-cancer therapy targeting this protein by interfering with tumor physiology and the tumor microenvironment . Given these findings , it is of vital importance that we elucidate mechanisms regulating novel functions of PARP-1 and poly ( ADP-ribose ) in tumor biology so that PARP inhibitors can ultimately make the transition to routine clinical use .
Human umbilical vein endothelial cells ( HUVEC ) were cultured in EGM-2 Endothelial Cell Growth Medium-2 ( LONZA ) . Cells were subjected to experimental procedures within passages 3–6 . B16-F10-luc-G5 cells stably expressing plasmids pGL3 control ( SV40-luc ) ( Promega ) and pSV40/Zeo ( Invitrogen ) . Human ( G361 ) , murine ( B16-F10 ) malignant melanoma cells and breast cancer ( MCF7 ) cells were cultured in DMEM containing 10% fetal bovine serum , 0 . 5% gentamicin ( Sigma , St . Louis , MO ) , and 4 . 5% glucose . All cells were cultured at 37°C ( 5% CO2 ) . The tumor cell lines have been developed as described in detail previously [40] . Melnikova et al . [41] found that unlike human melanomas , the murine melanomas cell lines did not have activating mutations in the Braf oncogene at exon 11 or 15 . All of the cell lines also expressed PTEN protein , indicating that loss of PTEN is not involved in the development of murine melanomas . This B16-F10 cell has previously been shown to be sensitive to stable depletion of PARP-1 in vivo melanoma growth [17] . Previous publication from our lab in G361 cells show similar results [16] . Cells were treated with the PARP inhibitors 3 , 4-dihydro-5-[4- ( 1-piperidinyl ) butoxyl]-1 ( 2H ) -isoquinolinone ( DPQ ) , [N- ( 6-Oxo-5 , 6-dihydro-phenanthridin-2-yl ) -N , N-dimethylacetamide] ( PJ-34 ) ( Alexis Biochemicals , San Diego , CA ) ( as described [42] , KU0058948 ( as we shown in previous publications [16] or Olaparib ( KU0059436 , Selleckchem ) for 22 hours . For capillary-like formation assays , 25 µL of Matrigel ( BD Biosciences ) were spread onto eight-chamber BD Falcon glass culture slides ( BD Biosciences ) or onto 96-well plates . Cells were seeded at 2 . 5×104 cells per well ( high density ) in eight-chamber slides and at 5×103 cells per well ( low density ) in 96-well plates and maintained in RPMI supplemented with 1% FBS [13] . These assays were performed according to previously published methods [13] . Primary antibodies used in these studies consisted of vimentin and VE-cadherin ( mouse monoclonal ) , E-cadherin ( rabbit polyclonal ) ( Santa Cruz Biotechnology ) , Snail1 and pVE-cadherin ( rabbit polyclonal ) ( Abcam ) , ILK ( rabbit monoclonal ) ( Millipore ) , Axl ( rabbit polyclonal ) , total-GSK-3β ( mouse monoclonal ) and pGSK-3β ( rabbit monoclonal ) ( Cell Signaling ) , β-catenin ( mouse monoclonal ) ( BD Transduction Laboratories ) , PARP-1 ( monoclonal ) ( Alexis ) as well as β-actin ( Sigma Aldrich ) . Quantitation of western blots was performed using Quantity One software analysis and all densitometries were normalized for loading control ( Table S2 ) . Luciferase activity was determined after transfecting the constructions into the B16-F10 cells . Firefly Luciferase was standardized to the value of Renilla Luciferase . Cells were co-transfected with 0 . 5 µg renilla as control of transfection together with 0 . 5 µg of the Snail or E-cadherin plasmid using jetPEI cationic polymer transfection reagent according to the manufacturer's instructions . The expression of Firefly and Renilla luciferases was analysed 48 h after transfection , according of the manufacturer's instructions . Cloning of the human Snail1 promoter ( −869/+59 ) in pGL3 basic ( Promega ) , was described previously [43] . E-cadherin promoter were cloned into pGL3-basic ( Promega ) to generate pGL3-E-cadherin ( −178/+92 ) . HUVEC or G361 cells were transiently transfected with an irrelevant siRNA [44] , PARP-1 siRNA or vimentin siRNA ( Thermo Scientific ) for 24 h using JetPrime ( Polyplus transfection ) according to the manufacturer's recommendations . At 48 h post-transfection , the expression of PARP-1 , vimentin , Axl , E-cadherin , Snail1 , ILK , β-catenin , pGSK-3β and total-GSK-3β was measured . Cells were washed twice in phosphate-buffered saline ( PBS ) and scraped in Laemmli buffer ( 1 M Tris , 20% SDS and 10% glycerol ) and sonicated . The protein concentration was determined using the Lowry assay . Levels of β-actin were monitored as a loading control . We used the GFP-vimentin expression vector supplied by Dr . Goldman ( Department of Cell and Molecular Biology , Chicago , Illinois ) . For transfection , JetPrime was used according to the manufacturer's protocol . 24 h post-tranfection , 5 µM of LiCl ( Sigma Aldrich ) was added in MCF7 cells and 48 hours later , the expression of vimentin , ILK , pGSK-3β , total-GSk-3β , E-cadherin , Snail1 and β-catenin was measured . In other experiment , co-transfection of GFP-vimentin and ILK siRNA ( Sigma Aldrich ) was used the according of the manufacturer's protocol . GFP and an irrelevant siRNA [44] were used as a control . HUVEC and B16-F10 cells were cultured on coverslips in six-well cell culture dishes . Monolayer cultures were stained with CellTracker Green CMFDA in HUVEC cell ( 5-chloromethylfluorescein diacetate ) ( Invitrogen ) according to manufacturer recommendations or with 4′ , 6′-diamidino-2-phenylindole dihydrochloride ( DAPI ) ( post-fixation ) . A wound was induced in the confluent monolayer cultures , and the cultures were then treated with the indicated inhibitor . The cells were fixed with 3 . 7% buffered formaldehyde and then prepared for immunofluorescence . Images were captured using a confocal microscope ( LEICA TCS SP5 Argon Laser 488 nm , HeNe Laser 543 nm ) when the cells were stained with CellTracker Green CMFDA Abs [522 nm] and Em [529 nm] and Zeiss Axio Imager A1 microscopy for cells stained with DAPI . The method used to Wound Healing using a service provided by Wimasis with permits users to upload their images online at any time and form anywhere and allows their images to be analyzed and the results uploaded back to the researcher's serve . Madin-Darby canine kidney ( MDCK ) cells ( 1 , 5×104 ) were seeded in 12-well tissue culture dish . After 24 h , cells were transfected with GFP or GFP-vimentin and 1 day after , cells were incubated with HGF ( hepatocyte growth factor , Sigma Aldrich ) or PBS . HGF is a mitogenic growth factor that is well known to induce the dissociation of islands of cells into individual cells , termed “cell scattering” or EMT . When inhibitors were used , cells were preincubated with PARP-1 inhibitor , PJ-34 or Olaparib for 2 h before addition of HGF . After 48 h , representative photographs were taken at 10× magnification using a Leica Spectral confocal laser microscope . The results were analyzed using the MetaMorph image analysis software . The effect of PARP inhibitors on the formation of tube-like structures in Matrigel ( BD Biosciences ) was determined according to manufacturer instructions . Briefly , 24-well plates were coated with 100 µl of BD Matrigel™ Basement Membrane Matrix and allowed to solidify at 37°C in 5% CO2 . Cells were treated with DPQ ( 40 µM ) or PJ-34 ( 10 µM ) . After 22 h of incubation at 37°C in 5% CO2 , the cells were fixed with 3 . 7% formaldehyde , and images were acquired using an Olympus CKX41 microscope . The formation of tube-like structures was then quantified . Each treatment was performed in triplicate , and the experiment was independently repeated at least three times . C57BL/6 mice background ( 8 weeks old ) were subcutaneously ( s . c . ) flank-injected with 600 µl of matrigel ( BD Biosciences ) supplemented with VEGF ( 100 ng/ml ) ( Peprotech ) and heparin ( Sigma , 19 U ) . The negative controls contained heparin alone . Each group consisted of four animals . After seven days , mice were sacrificed and matrigel plugs were extracted . The angiogenic response was evaluated by macroscopic analysis of the plug at autopsy and by measurement of the hemoglobin ( Hb ) content within the pellet of matrigel . Hb was mechanically extracted from pellets reconstituted in water and measured using the Drabkin ( Sigma-Aldrich ) method by spectrophotometric analysis at 540 nm . The values were expressed as optical density ( OD ) /100 mg of matrigel . This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the Bioethical Committee of CSIC . The protocol was approved by the Committee on the Ethics of Animal Experiments of the CSIC . All surgery was performed under isoflurano anesthesia , and every effort was made to minimize suffering . Eight-week-old male C57BL/6 albino mice ( The Jackson Laboratories , Bar Harbor , MN , USA ) were injected subcutaneously with B16-F10-luc-G5 cells ( 1×105 ) and intravenously with B16-F10-luc-G5 cells ( 1×105 or 5×105 ) . Three times per week mice were injected intraperitoneally with DPQ dissolved in phosphate-buffered saline/10% DMSO at a dose of 15 mg/kg body weight or olaparib at 50 mg/kg . Mice were injected intraperitoneally with D-luciferin solution dissolved in phosphate-buffered saline at a dose of 150 mg/kg body weight . After 5 to 8 minutes , the animals were anesthetized in the dark chamber using 3% isoflurane in air at 1 . 5 L/min and O2 at 0 . 2 L/min/mouse , and animals were imaged in a chamber connected to a camera ( IVIS , Xenogen , Alameda , CA ) . Exposure time was 3 min in large binning , and the quantification of light emission was performed in photons/second using Living Image software ( Xenogen ) . Tumor growth was monitored at 0 , 2 , 7 , 14 and 21 days by in vivo imaging and bioluminiscence measurement . After 21 days , mice were sacrificed , and their organs were removed and stored in buffered formalin ( 3 . 7% ) until histological staining . Immunostaining for vimentin , VE-cadherin , pVE-cadherin , Snail1 and E-Cadherin was performed on cells plated onto coverslips and grown for 22 h prior to experimental treatments . The culture medium was removed , and the cells were fixed ( Paraformaldehyde 3% , Sucrose 2% in PBS ) for 10 minutes at room temperature . Permeabilization was performed using 0 . 2% Triton X-100 in PBS . The coverslips were rinsed three times in PBS prior to incubation with primary antibody for 1 h at RT and then rinsed three times in PBS before incubation with the secondary antibody . Secondary antibodies were FITC-conjugated anti-mouse IgG or anti-rabbit ( Sigma , St . Louis , MO ) . Antibodies were diluted in PBS containing 2% bovine serum albumin . Nuclear counterstaining with 4′ , 6′-diamidino-2-phenylindole dihydrochloride ( DAPI ) was performed after removal of excess secondary antibody . Slides were prepared using Vectashield mounting medium ( Vector Lab . , Inc . , Burlingame , CA 94010 ) , cover slipped and stored in the dark at 4°C . Immunofluorescence images were obtained in the linear range of detection to avoid signal saturation using a fluorescent microscope ( Zeiss Axio Imager A1 ) or confocal microscopy ( Leica SP5 ) . For conventional morphology , three buffered 4% formaldehyde-fixed , paraffin-embedded skin longitudinal tissue sections were stained with periodic acid schiff ( PAS ) at the end of treatment . The study was done in blinded fashion on 4-µm sections with light microscopy . The mitosis and apoptosis cells were assessed by examining their number in ten high power field ( hpf ) at 600× magnifications . The results were expressed as number of cells per mm2 . For evaluation of blood vessels density , tissue sections of different groups were dewaxed , hydrated , and heat-treated in 0 . 01 M citrate buffer for antigenic unmasking . The rest of the procedure was carried out using an automatic immunostainer ( Autostainer480 , Labvision , Fremont CA , USA ) . The incubation time with lectin Ulex europaeus biotin conjugated was 60 min , the dilution was 1∶200 , and the streptavidin-biotin-peroxidase method ( Master Diagnóstica , Granada , Spain ) with diaminobezidine was used as visualization system . A millimeter scale in the eyepiece of a microscope BH2 ( Olympus ) with 40× objective was used to count the vessel per mm2 of tissue section . The morphological and immunohistochemistry study was done in a double-blinded fashion by two pathologists . For data shown in Figure 7 and FigureS7 we have fitted the values of the average number of tumors per mouse during carcinogenesis treatment using the Mann-Whitney u-test . Statistical analysis of other experiments used unpaired Student's t-test .
|
Metastasis is the spread of malignant tumor cells from their original site to other parts of the body and is responsible for the vast majority of solid cancer-related mortality . PARP inhibitors are emerging as promising anticancer therapeutics and are currently undergoing clinical trials . It is therefore important to elucidate the mechanisms underlying the anti-tumor actions of these drugs . In our current study , we elucidated novel anti-neoplastic properties of PARP inhibitors that are responsible for the anti-metastatic effect of these drugs in the context of malignant melanoma . These effects appear to be the result of PARP-1's ability to regulate the expression of key factors , such as vimentin and VE-cadherin , involved in vascular cell dynamics and to limit pro-malignant processes such as vasculogenic mimicry and EMT .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"oncology",
"medicine",
"cancer",
"genetics",
"rna",
"interference",
"antiangiogenesis",
"therapy",
"gene",
"expression",
"genetics",
"cancer",
"treatment",
"biology",
"dna",
"modification",
"dna",
"transcription"
] |
2013
|
PARP-1 Regulates Metastatic Melanoma through Modulation of Vimentin-induced Malignant Transformation
|
Brucella melitensis is a facultative intracellular bacterium that causes brucellosis , the most prevalent zoonosis worldwide . The Brucella intracellular replicative niche in macrophages and dendritic cells thwarts immune surveillance and complicates both therapy and vaccine development . Currently , host-pathogen interactions supporting Brucella replication are poorly understood . Brucella fuses with the endoplasmic reticulum ( ER ) to replicate , resulting in dramatic restructuring of the ER . This ER disruption raises the possibility that Brucella provokes an ER stress response called the Unfolded Protein Response ( UPR ) . In this study , B . melitensis infection up regulated expression of the UPR target genes BiP , CHOP , and ERdj4 , and induced XBP1 mRNA splicing in murine macrophages . These data implicate activation of all 3 major signaling pathways of the UPR . Consistent with previous reports , XBP1 mRNA splicing was largely MyD88-dependent . However , up regulation of CHOP , and ERdj4 was completely MyD88 independent . Heat killed Brucella stimulated significantly less BiP , CHOP , and ERdj4 expression , but induced XBP1 splicing . Although a Brucella VirB mutant showed relatively intact UPR induction , a TcpB mutant had significantly compromised BiP , CHOP and ERdj4 expression . Purified TcpB , a protein recently identified to modulate microtubules in a manner similar to paclitaxel , also induced UPR target gene expression and resulted in dramatic restructuring of the ER . In contrast , infection with the TcpB mutant resulted in much less ER structural disruption . Finally , tauroursodeoxycholic acid , a pharmacologic chaperone that ameliorates the UPR , significantly impaired Brucella replication in macrophages . Together , these results suggest Brucella induces a UPR , via TcpB and potentially other factors , that enables its intracellular replication . Thus , the UPR may provide a novel therapeutic target for the treatment of brucellosis . These results also have implications for other intracellular bacteria that rely on host physiologic stress responses for replication .
Brucellosis is a chronic debilitating disease with protean manifestations and insidious onset most frequently caused by the facultative intracellular bacteria Brucella melitensis [1] . With over 500 , 000 new infections per year , brucellosis is the most prevalent zoonosis worldwide [2] . Brucellosis is most often acquired by consumption of contaminated dairy products . Following ingestion , Brucella infects macrophages and dendritic cells that constitute the replicative reservoir [3] . The intracellular replicative niche thwarts immune surveillance , complicates vaccine development , and renders the organism refractory to antibiotics [1] . A greater understanding of host-pathogen interactions is critical for elucidating disease pathogenesis and thus improving therapeutic strategies . Brucella establishes its stealthy intracellular lifestyle through virulence factors . Brucella expresses a weakly endotoxic smooth LPS that directs bacterial uptake via class A scavenger receptor in lipid rafts [4]–[6] . Inside macrophages , ∼90% of bacteria are killed within the first 4 h . However , some Brucella-containing vesicles ( BCV ) avoid end-stage lysosomes and ultimately fuse with the endoplasmic reticulum ( ER ) [7] . Fusion appears to involve an early ER to Golgi vesicular compartment , as GAPDH and the small GTPases Rab2 and Sar1 are essential for replication [8] , [9] . Replicative BCV contain ER markers including calnexin , calreticulin and sec61β [7] . Correct trafficking and ultimately replication depend upon de novo bacterial protein expression following cellular infection . In particular , BCV acidification activates the type IV secretion system encoded by the VirB operon [10] . VirB mutant BCV fail to fuse with the ER and VirB mutants are greatly attenuated in vivo [7] . Within 48 h of infection , Brucella induces a marked reorganization of the ER with ER membrane accretion around replicating bacteria [7] . The mechanism by which Brucella disrupts ER structure is currently unknown . Somehow , the host cell adapts to this perturbation , as Brucella infection inhibits apoptosis . Although the bacterial factors leading to successful infection are beginning to be clarified , the host pathways supporting replication remain poorly understood . The requirement for ER fusion and dramatic restructuring of the ER suggest Brucella most likely disrupts ER homeostasis . To cope with physiologic and stressful perturbations of ER function , cells mobilize a conserved stress response called the Unfolded Protein Response ( UPR ) [11] . The UPR is initiated when unfolded proteins within the ER excessively bind the chaperone BiP/glucose regulated protein ( Grp ) 78 , titrating it away from three primary ER membrane resident stress sensors , inositol requiring kinase 1 ( IRE1 ) , activating transcription factor ( ATF6 ) , and PKR-like endoplasmic reticulum kinase ( PERK ) . IRE1 is both a kinase that phosphorylates targets such as Jun kinase ( JNK ) , and an endonuclease that cleaves 26 nucleotides from the X-box binding protein 1 ( XBP1 ) mRNA , thus removing a premature stop codon [12] . Spliced XBP1 mRNA encodes the full-length transcription factor . Upon release of BiP , ATF6 traffics from ER to Golgi , where site-specific proteases cleave it to an active transcription factor . PERK phosphorylates eukaryotic initiation factor 2α , resulting in global translational attenuation apart from select open reading frames ( e . g . ATF4 mRNA ) . The three primary stress sensor-dependent biochemical pathways regulate the following: 1 ) UPR target gene transcription , including chaperones and co-chaperones ( e . g . BiP and ERdj4 ) that increase folding capacity , 2 ) molecules involved in ER associated degradation and 3 ) pro-apoptotic factors such as C/EBP homologous protein ( CHOP ) . The UPR exerts a profound effect on multiple cellular processes including autophagy , apoptosis , ER and Golgi biogenesis , and lipid and protein synthesis . If ER stress remains unresolved despite these adaptive measures , the UPR initiates apoptosis [11] . One study suggests the UPR may play a role in Brucella replication . Brucella replicate less efficiently in IRE1 knockdown insect cells and IRE1 deficient murine embryonic fibroblasts [13] . The IRE1 axis of the UPR regulates autophagy , which appears to support replication in non-phagocytic cells [14] , [15] . The role of autophagy in supporting Brucella survival and replication efficiency in macrophages remains somewhat controversial , though compelling work implicates early autophagy pathway proteins in completion of the Brucella intracellular life cycle [16] , [17] . Serum starvation enhances bacterial replication in HeLa cells , which may reflect a component of ER stress [18] . However , the relevance of the UPR to Brucella replication in physiologic host cells ( e . g . macrophages ) remains unknown . Although viral manipulation of the UPR has been extensively studied , very little is known about the effect of bacterial infection on the host UPR . Evidence for UPR activation has been detected in histologic sections from patients infected with M . tuberculosis [19] . However , the relationship between infection and host response was not clear . In one report , intracellular bacteria Francisella , Listeria , and Mycobacteria induced XBP1 mRNA splicing via toll like receptor ( TLR ) signaling [20] . Deficiency of the TLR-adaptor protein myeloid differentiation primary response gene 88 ( MyD88 ) ablated TLR2 and decreased TLR4-dependent XBP1 splicing . TLR-dependent XBP1 splicing was not accompanied by downstream UPR target gene induction , although there was evidence supporting a role for XBP1 in synergistic cytokine induction . XBP1 was essential for optimal cytokine production and immune control of Francisella in vivo . The exact mechanism underlying this selective XBP1 pathway activation is unknown . Extracellular Listeria monocytogenes has also been shown to induce a more complete UPR , involving all three signaling axes , via production of listeriolysin [21] . However , the intracellular life cycle of this bacterium differs greatly from Brucella . In this study , we evaluated induction of the host UPR by Brucella infection in macrophages . We detected activation of all three axes of the UPR , stemming from activation of IRE1 , PERK and ATF6 , as evident by increased UPR target gene expression and XBP1 mRNA splicing . Although XBP1 splicing appeared to be largely MyD88-dependent , UPR gene expression was independent of the TLR-signaling adaptor molecule . Optimal UPR target gene induction required both live bacteria and expression of the microtubule-modulating Brucella protein TcpB . Finally , tauroursodeoxycholic acid ( TUDCA ) , a pharmacologic chaperone that inhibits the UPR , substantially decreased replication . Together these data suggest Brucella actively induces a UPR that enables its intracellular replication within the ER in macrophages .
A previous study documented dramatic reorganization of the ER within 48 h of Brucella infection [7] . We observed ER fragmentation and condensation even within 24 h of infection ( Figure 1 ) . The replicative requirement for ER-BCV fusion and the ER structural reorganization following infection raised the possibility that Brucella triggers the host cell UPR . The UPR directs an adaptive program through the induction of target gene transcription . Although the three primary biochemical signaling pathways have overlapping functions , several of the UPR gene targets appear to be relatively pathway specific; thus activation of PERK , IRE1 , and ATF6 can be detected by downstream induction of mRNA for CHOP , ER localized DnaJ homologue 4 ( ERdj4 ) , and BiP , respectively [22]–[24] . XBP1 spliced and unspliced mRNA species can be resolved by high-percentage agarose gel electrophoresis , and this method is often used to detect IRE1 endonuclease activity [25] . To test for UPR activation , RAW264 . 7 macrophages were infected with B . melitensis for 24 h ( Figure 2 ) . Of the genes examined , CHOP showed the most robust induction and mRNA expression correlated with a marked increase in CHOP protein . The induction of BiP , CHOP and ERdj4 mRNA and evidence for XBP1 mRNA splicing supports the hypothesis that Brucella induces a UPR involving all three primary signaling axes in macrophages in vitro . To determine if Brucella infection induces a detectable UPR in vivo , splenic CD11b+ cells ( containing macrophages ) were isolated 24 h following infection and UPR gene expression assessed by qPCR ( Figure 3 ) . TNF-α expression served as a positive control that is expected to increase early with infection . Induction of BiP , CHOP , and ERdj4 expression was evident following in vivo infection , consistent with activation of the UPR in vivo . A recent study described XBP1 splicing in response to TLR2 and TLR4 agonists ( Pam3cysK4 and LPS ) as well as various intracellular bacteria [20] . Interestingly , although XBP1 was required for optimal TLR-stimulated cytokine production , TLR ligation decreased BiP and CHOP induction by pharmacologic UPR inducers . Thus the TLR-MyD88-XBP1 pathway appears antagonistic towards the rest of the ER stress response . Another report also documents selective suppression of ER stress signaling by LPS [26] . Brucella stimulates both TLR2 and TLR4 and the TLR adaptor MyD88 appears to be essential for controlling infection in vivo [27] . To further elucidate the role of TLR-MyD88 signaling in UPR induction by Brucella , XBP1 splicing and UPR target gene expression was examined in primary bone marrow derived macrophages from MyD88 deficient mice ( Figure 4 ) . The phenotype of these mice was confirmed by diminished IL-6 expression following infection of macrophages in vitro . Although Brucella induced XBP1 mRNA splicing was impaired in MyD88−/− macrophages , induction of other UPR target genes , e . g . CHOP and ERdj4 , was intact . In fact , induction of CHOP was slightly greater in the MyD88−/− macrophages ( p = 0 . 034 ) . This result is consistent with the described suppression of UPR target genes by TLR agonists . BiP expression was not upregulated at 24 h following infection in these experiments . XBP1 mRNA splicing in response to a pharmacologic UPR inducer , tunicamycin , was equivalent in the two mouse strains ( 82±1% in MyD88−/− vs 85±4% in wild type , data not shown ) . Together , these data suggest Brucella induces XBP1 splicing through TLR-MyD88 signaling; however induction of the other UPR target genes examined ( CHOP and ERdj4 ) proceeds through a MyD88 independent pathway . These results further demonstrate Brucella-dependent UPR induction in primary macrophages , thus validating the RAW 264 . 7 macrophage cell line data . One prediction of these data is that the bacterial surface of Brucella ( containing LPS ) will be sufficient to stimulate XBP1 splicing . However , the induction of other UPR-dependent events ( e . g . increased CHOP expression ) may involve other surface or intracellular components . To begin testing this premise , RAW macrophages were infected with heat killed Brucella ( Figure 5A ) . Heat killed Brucella induced XBP1 splicing to a similar extent as living Brucella , as predicted , but induced significantly less UPR target gene expression . These data suggest other heat-labile or newly produced factors besides bacterial LPS are responsible for activating the host UPR . The need for living bacteria for optimal UPR target gene induction may reflect the involvement of de novo bacterial protein/factor production following infection . De novo expression of virulence factors directs the distinctive trafficking and replicative events that result in chronic infection . In particular , products encoded in the VirB operon appear to be essential for fusion of BCV with ER membranes and subsequent replication [7] . To test the requirement for VirB , UPR induction was assessed in a VirB4 mutant ( Figure 5A ) [28] . This mutant displays attenuated virulence , with defects in in vivo persistence . Consistent with above results , XBP1 splicing was intact . Downstream UPR target gene induction following 24 h infection with the VirB deletion mutant ( ΔVirB ) was variable , and not statistically different compared to the wild type control . These data suggested another Brucella factor , besides those encoded by VirB , or utilizing the VirB-dependent type IV secretion system , must be involved in UPR induction . The B . abortus protein Btp1 ( Brucella-TIR-Protein 1 ) was originally characterized by its ability to inhibit dendritic cell maturation and to antagonize TLR2 signaling [29] . TcpB ( Toll/Interleukin 1 like receptor domain containing protein ) , the correlating protein in B . melitensis also antagonizes TLR signaling and NF-κB activation [30] . We have recently shown that TcpB co-localizes with plasma membrane and microtubules and exerts a microtubule stabilizing effect similar to paclitaxel ( Taxol ) [31] . Besides co-localizing with cytoskeletal elements , exogenously expressed TcpB also co-localizes by immunofluorescence with the ER protein calreticulin ( Figure S1 ) . ER structure is microtubule dependent [32] . In the context of cancer research , microtubule-stabilizing agents such as paclitaxel have been shown to induce ER stress [33] , [34] . Microtubules also regulate intracellular vesicular trafficking [35] . Brefeldin A , which blocks egress from the ER is commonly utilized to induce the UPR [36] . Thus we reasoned that TcpB might contribute to UPR induction through microtubule-related modification of ER structure . As shown in Figure 5A , infection with the TcpB deletion mutant ( ΔTcpB ) resulted in ≈60% decreased expression of BiP , CHOP and ERdj4 as compared to wild type Brucella . Note , some CHOP up-regulation by the TcpB mutant was still detectable ( p≤0 . 005 vs . NI ) . Complementation of the TcpB mutant with exogenous TcpB recovered UPR gene expression ( Figure 5B ) . These results were consistent with a role for TcpB protein in UPR induction . We hypothesized that UPR induction and ER restructuring are related events . In this case , the diminished UPR induction by the TcpB mutant should correlate with decreased effect on ER structure . Indeed , infection with the TcpB mutant did not induce the same degree of condensation and vacuolization as observed upon infection with wild type Brucella ( Figure 6 , Figure S2 ) . The ER remains lacy , reticular and more evenly distributed compared to wild type infection . Trafficking of the Brucella within the cell however appears relatively intact , as the Brucella still migrate centrally to form a ring around the nucleus . Thus trafficking may not depend upon dramatic ER restructuring . To directly test the role of TcpB in UPR induction and ER restructuring , RAW 264 . 7 macrophages were treated with purified TcpB protein using a concentration previously shown to affect microtubules and NF-κB signaling ( Figure 7A ) [30] , [31] . TcpB protein was sufficient to upregulate BiP , CHOP , ERdj4 and spliced XBP1 . The relative magnitude of effect appeared much greater for BiP and CHOP than for ERdj4 and spliced XBP1 . Triggering of UPR events correlated well with effects of TcpB on ER structure as detected by immunofluorescence microscopy ( Figure 7B ) . Compare the diffuse lacy reticular pattern extending throughout the cell in untreated or the MBP treated cells ( Figure S3 and 7B ) to the circumscribed circular area with large holes and more defined compact structures in TcpB treated cells . The majority of cells examined appeared similarly affected . Overall ER area appears enlarged , particularly at the lower dose of TcpB ( Figure S4 ) . ER condensation and fragmentation increases with dose of purified TcpB . Similar effects were observed by 12 h of treatment ( not shown ) . This effect on ER structure was qualitatively similar to that induced by infection of macrophages with wild type Brucella ( Figures 1 and 6 ) . Together these results implicate TcpB in both ER fragmentation and UPR induction . It was unclear how the ER disruption related to the UPR . Were the ER structural changes a result of ER stress or is the UPR downstream of the ER disruption ? To begin addressing this question , macrophages ( or in some experiments D17 osteosarcoma cells ) were treated with the ER stress inducer tunicamycin , a potent N-linked glycosylation inhibitor ( Figure 8 ) [37] , [38] . Although tunicamycin caused ER vacuolization , most likely related to proteins being retained in the ER , the disposition of ER calreticulin in the cell was different compared to TcpB treatment ( or infection , see above ) : in the tunicamycin treated cells , the ER did not condense in a sphere but remained distributed into the macrophage processes . Thus TcpB induced disruption does not simply reiterate an ER stressor . If TcpB-induced ER restructuring were upstream of the UPR and not dependent on UPR , then blockade of the UPR should have no effect on ER disruption . To address this hypothesis , we inhibited the UPR with tauroursodeoxycholic acid ( TUDCA ) , a chemical chaperone widely utilized in vitro and in vivo to modulate the UPR . The ability of TUDCA to impede BiP and CHOP induction by tunicamycin was confirmed ( Figure 8 ) . Inhibition of XBP1 splicing was more variable . TUDCA also inhibited tunicamycin-dependent cytokine induction ( Figure S5 ) as expected . TUDCA treatment mitigated the effect of tunicamcyin ( less vacuolization and size increase ) but had no apparent effect on TcpB-related ER restructuring . These results suggest that ER restructuring is not UPR dependent . If ER structure and UPR are interdependent , ER disruption must occur upstream of UPR induction . The above data suggests Brucella induces the UPR at least in part via TcpB . However , it was not clear if the host mounts a UPR in response to infection , or if the UPR benefits the bacteria ( or both ) . Viral infections manipulate the UPR in a variety of ways , including capitalizing on host protein production and folding machinery to enhance replication . One report utilizing insect cells and mouse embryonic fibroblasts suggests the IRE1 branch of the UPR supports Brucella replication , but the relevance to macrophages was unclear [13] . Brucella may not behave exactly the same in macrophages and non-phagocytic cells [15] , [39] . TcpB mutant Brucella are defective at spreading systemically early during infection in vivo [27] . However , the other effects of TcpB , in particular inhibition of TLR signaling in the setting of an in vivo immune response , complicate the interpretation . To determine if TcpB plays a role in intracellular replication in macrophages in vitro , RAW 264 . 7 cells were infected with wild type B . melitensis or the TcpB mutant . Select cultures were also treated with very low dose tunicamycin to enhance the UPR ( Figure 9 ) . Initial uptake of the TcpB mutant was greater than wild type ( p = 0 . 008 ) , but the replication growth curve plateaus below the level observed in wild type . This slowed growth resulted in decreased CFU later during the culture period ( p≤0 . 001 after 24 h ) . Tunicamycin treatment enhanced recoverable TcpB mutant CFU at all time points ( p = 0 . 04 at 4 h and p≤0 . 006 thereafter ) . This effect of tunicamycin is consistent with previous reports documenting enhanced Brucella replication by serum starvation ( nutrient deprivation ) [18] . The pleiotropic effects of TcpB within an individual cell may have multiple effects on initial uptake , early bacterial destruction , trafficking , replication and ultimate recoverable CFU . For instance , the cytoskeletal disruption and potential alteration of vesicular trafficking induced by TcpB could initially impede Brucella infection . This interpretation is consistent with the increased initial CFU observed in the TcpB mutant cultures . An initial negative effect of TcpB on invasion could obscure a later positive effect on replication ( and thus detected CFU ) . As a separate issue , although TcpB may play a role in inducing the UPR , other molecules may compensate in the absence of TcpB , as evident by the residual CHOP induction ( Figure 5 ) and another recent report implicating VceC [40] . To more directly assess the role of the UPR in Brucella replication in macrophages , the cells were treated with TUDCA to inhibit the UPR . We confirmed that TUDCA pre-treatment decreases Brucella induced BiP and CHOP expression at 24 h ( Figure 10 ) . The effect of TUDCA on XBP1 splicing was variable ( similar to the effect on tunicamycin-induced UPR ) , and ERdj4 expression increased . During a 24 h period , TUDCA had minimal impact on RAW cell viability ( 93±10% untreated ) and none on Brucella at 500 µg/mL ( Figure S6 ) . At earlier time points ( 12–16 h ) , the effect of TUDCA on replication was modest but reproducible ( 5 . 4±1 . 8 fold mean decrease for 4 experiments ) . However , TUDCA pre-treatment significantly decreased recoverable Brucella CFU , typically by a log or more ( p≤0 . 02 in 4 independent experiments , range 4-fold to 3 logs ) by 24–36 h . TUDCA exerted a similar effect on Brucella CFU in the osteosarcoma D17 cell line ( Figure S7A ) [37] , [38] . TUDCA did not appear to inhibit Brucella trafficking to a peri-nuclear location in those cells containing visible bacteria . Together , these data are consistent with a critical role for UPR pathways in enabling Brucella intracellular replication inside macrophages .
B . melitensis infection mobilizes all three UPR signaling axes in macrophages , stemming from the activation of IRE1 , PERK and ATF6 . Oxidative stress also strongly activates the PERK pathway , thus the UPR is often referred to as an “integrated stress response” [41] . However robust induction of target genes from the three distinct biochemical signaling pathways is most consistent with the UPR [42] . One report demonstrated IRE1 phosphorylation and PERK pathway activation in M . tuberculosis infected macrophages in vivo . However the direct link between infection and induction of host UPR was not established [19] . Another study implicated the IRE1 pathway in supporting Brucella replication , consistent with the results obtained in this study , however the relevance to macrophages was unclear [13] . Although previous data reported XBP1 mRNA splicing by intracellular bacteria such as Francisella [20] , this is one of the first reports of more widespread UPR induction resulting directly from intracellular bacterial infection rather than toxin production . Brucella induced XBP1 mRNA splicing appears to proceed predominantly through the previously described MyD88 ( TLR ) dependent pathway [20] . The unusual smooth Brucella LPS contains reduced negative charges and unusually long aliphatic hydrocarbon chains in the Lipid A core ( C28 as compared to C12-16 in enterobacteria ) [4] . Related to these properties , Brucella smooth LPS displays reduced TLR4 agonist activity [43] . Thus smooth LPS may be a relatively weak inducer of XBP1 splicing . Consistent with this prediction , the genomic island 2 deletion Brucella mutant that expresses rough LPS triggers much more robust XBP1 splicing ( data not shown ) [44] . In comparison with XBP1 splicing , downstream CHOP and ERdj4 target gene induction was entirely MyD88-independent . BiP induction was not detected in these particular experiments , potentially related to timing ( BiP upregulation is an early transient event ) , mouse strain , or differences in macrophage type [45] . Thus , as noted by others , all signaling pathways encompassed by the UPR are not always coordinately regulated [20] . “UPR” signaling events such as XBP1 splicing may be triggered by non-UPR agonists and UPR signaling pathways are not always activated in their entirety . Other examples of XBP1 splicing-downstream target disconnection come from the viral literature , and Hepatitis C in particular [46] . The mechanism underlying the dissociation remains unknown . In the present study , it was curious that the canonical XBP1 target gene ERdj4 was up regulated in the absence of significantly detectable XBP1 splicing in the MyD88−/− bone marrow macrophages . There are several possible explanations . First , sufficient TLR4-TRIF dependent XBP1 activity remains to induce ERdj4 . Second , only part of XBP1 splicing is MyD88 dependent and the assay is not sufficiently sensitive to detect minor differences . The low level XBP1 splicing induced by purified TcpB is consistent with this idea , as we would assume this is ER stress rather than MyD88-related . Indeed , TcpB would be expected to antagonize TLR-MyD88-dependent signaling [29] , [30] . Third , ERdj4 may be induced in an XBP1 independent manner [20] , [47] . Given the evidence that even weak TLR signaling by Brucella still induces XBP1 splicing , it will be interesting to determine the role of XBP1 in Brucella-induced cytokine production . Our results reveal a new role for the TcpB protein in regulating host stress responses and ER structure . Indeed , the ability of TcpB to fragment and condense the ER may be the underlying mechanism for the dramatic ER restructuring first reported almost a decade ago [7] . Based on 1 ) our data correlating ER disruption and UPR induction in response to purified TcpB , 2 ) the diminished UPR and ER structural impact in the absence of TcpB , and 3 ) the capacity of an analogous microtubule disrupting drug paclitaxel to induce ER stress , it is highly likely that TcpB induced ER restructuring and UPR are causally linked [33] , [34] . The comparison with tunicamycin treatment and the lack of TUDCA effect on TcpB induced ER restructuring also suggest that TcpB-induced UPR occurs following , or downstream of ER structural disruption . However , it remains possible that ER stress is not directly related to TcpB-induced ER structural changes . Purified TcpB was more effective at upregulating CHOP and BiP than IRE1 dependent events such as XBP1 splicing and ERdj4 . Indeed XBP1 splicing was intact in the TcpB mutant infected cells and minimally induced in cells by TcpB protein . The vast majority of XBP1 splicing appears to proceed through the TLR-MyD88 pathway . Thus the readout of “UPR” reflects contributions from multiple bacterial factors . The effect of TUDCA on replication also suggest that the delayed virulence of the TcpB mutant in a susceptible IRF1−/− deficient mouse model may reflect both altered replication and enhanced cytokine production . It is a testament to bacterial efficiency that one protein product may both antagonize host immune signaling and induce host stress responses that support bacterial replication . The ultimate role of TcpB in replication , given the pleiotropic effects of this molecule , remains unclear . Initially , uptake of the mutant is much greater in macrophages , but the growth curve plateaus below the level of the wild type . Growth of the ΔBtp1 B . abortus was not impaired in dendritic cells [29] . This may reflect timing , a difference in macrophages vs . dendritic cells , cell line vs . primary cells on another strain background , or differences in B . abortus vs . B . melitensis . The TcpB mutant is clearly attenuated in vivo , though immune modulation complicates the interpretation [30] . Although TcpB appears to play a pivotal role in regulating UPR target genes , other virulence factors ( e . g . VirB ) may contribute . Indeed , CHOP expression was not reduced to the non-infected level in the TcpB mutant infection , consistent with the existence of other UPR inducing molecules [40] . Also , TUDCA inhibited growth of the TcpB mutant ( Figure S7B ) . The experimental variability obtained with the VirB mutants may reflect a timing issue ( important earlier or later than our experimental window ) or sensitivity . The proportion of cells infected and number of bacteria/cell will affect UPR detection . Since the VirB mutant traffics abnormally and fails to survive inside macrophages , fewer bacteria will be available to produce UPR-inducing factors [7] . Interestingly , in the IRF1−/− mouse model , patterns of in vivo virulence differed between the VirB and TcpB mutants consistent with roles in different parts of the bacterial life cycle . TcpB appears to regulate early spread of infection whereas VirB contributes more to bacterial persistence [28] , [30] . The requirement for living bacteria to optimize UPR target gene induction suggests the UPR is an active process supported by new protein ( s ) or other factor ( s ) produced following infection; it is not just a host response to components present in dead bacteria . The residual UPR induction by heat-killed bacteria may reflect TcpB , or an unidentified factor produced by the bacteria during growth in broth . The UPR may support the intracellular life cycle of Brucella in a number of ways . First , the UPR mobilizes amino acid transport and supports lipid biogenesis . Second , the UPR also initiates autophagy , thus providing more nutrients . As described by Starr et al , the UPR regulated autophagy may participate in completing the Brucella intracellular life cycle , allowing spread to neighboring cells [17] . Third , the UPR enhances protein-folding capacity through induction of chaperones and other folding machinery . Fourth , the UPR allows cells to cope with oxidative stresses . Finally , as a means of physiological adaptation , the UPR may enable host cells to survive the disruption of ER structure and function . The UPR encompasses anti-apoptotic mechanisms and only promotes apoptosis when stress is severe or prolonged . In the viral literature , Dengue activates all three UPR pathways , yet suppresses downstream apoptosis [48] . It will be interesting to determine if some of the same apoptosis modulating mechanisms apply to Brucella . Another possibility is that Brucella LPS may sufficiently temper CHOP induction to avert apoptosis [20] . In this study , TUDCA pre-treatment exerted a dramatic effect , decreasing recoverable CFU in culture . The simplest interpretation is that the host UPR plays an absolutely critical role in supporting Brucella replication . This hypothesis is consistent with the work from Qin et al . showing decreased Brucella CFU following IRE-1 knockdown . We also have preliminary data suggesting this UPR axis supports replication in macrophages , most likely through the IRE1-kinase-JNK signaling pathway rather than through XBP1 ( data not shown ) . The contrasting effect of TUDCA on Brucella replication and XBP1 splicing/ERdj4 expression is consistent with our preliminary XBP1 RNAi data showing no effect on replication . The XBP1 variability in response to TUDCA may reflect multiple mechanisms of XBP1 splicing induction . However the apparent effect of TUDCA on Brucella-induced BiP and CHOP expression may also result from greatly diminished numbers of bacteria . Although TUDCA is widely utilized to assess the role of the UPR in vivo , and is approved for use in humans , the drug may affect other cellular processes besides the UPR [49] . The non-specificity of TUDCA is one limitation of this study . However , these results supply strong rationale to further investigate which specific UPR-related molecules might be involved in supporting Brucella replication in macrophages . Also , despite non-specificity , TUDCA may be useful therapeutically , particularly in view of safety and cost . It may prove important to inhibit multiple arms of the UPR , as inhibition of one specific signaling axis may not be sufficient . Successful inhibition of Brucella virulence in vivo by TUDCA or other more selective UPR modulation would open a new avenue of drug development . TUDCA has an excellent safety profile and is being studied in humans to counteract UPR-related metabolic syndromes [49] , [50] . It will be essential to determine whether TUDCA mediated inhibition of replication outweighs the effect of UPR blockade on inflammatory cytokine production in vivo . We , along with other researchers , have described dramatic augmentation of interferon and inflammatory cytokine production by the UPR [20] , [51] . Indeed , the UPR has been implicated in numerous inflammatory and autoimmune diseases [52] , [53] . Currently , little is known about the role of the UPR in immune responses to Brucella , and the formation of immune memory [54] . The concept that subverting the host UPR enables bacterial replication in macrophages , thus promoting infectious success represents a paradigm shift for the field that merits further investigation . The results from this study have broad implications for other bacteria that establish an intracellular replicative niche , particularly those that interact with the ER [55] .
The RAW264 . 7 murine macrophage and D17 canine osteogenic sarcoma cell lines ( both ATCC ) were maintained in RPMI 1640/high glucose with 4 mM L-glutamine , sodium pyruvate ( Hyclone Laboratories ) and supplemented with 10% FBS ( Hyclone ) , 100 U/mL penicillin , and 100 µg/mL streptomycin . B . melitensis 16M , the engineered bioluminescent strain GR019 ( VirB mutant ) , or the TcpB deletion mutant were grown in Brucella broth ( BB , Difco ) supplemented with 50 µg/mL kanamycin [28] , [30] . To heat kill bacteria , B . melitensis in BB was incubated at 65°C for 60 min . The purification of TcpB protein has been described [30] . MBP-TcpB was used at a concentration of 50 µg/mL , with maltose binding protein ( MBP ) as a control [30] . Mice were kept in facilities at the University of Wisconsin-Madison that are accredited by the American Association of Laboratory Animal Care . Mouse experiments were performed with oversight and approval of the University of Wisconsin-Madison School of Medicine and Public Health and School of Veterinary Medicine Animal Care and Use Committee ( NIH assurance number: A3368-01 ) , in accordance with recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . MyD88−/− femurs were a gift from Laura Knoll , University of Wisconsin-Madison . Bone marrow cells from C57BL/6 wild type or MyD88−/− femurs were isolated on Histopaque 1083 ( Sigma-Aldrich , St . Louis , MO ) and differentiated for 7 days in RPMI 1640 with 10% FBS and 50 ng/mL recombinant murine M-CSF ( Peprotech ) . For in vivo infections , 4–7 mice/group of 6–8 week old BALB/c mice were injected i . p . with PBS or 107 GR023 ( bioluminescent B . melitensis ) . After 24 h , spleens were pooled within groups , homogenized , and splenic macrophages were isolated using CD11b+ magnetic cell separation ( Miltenyi ) according to manufacturer's protocol . Cells were immediately resuspended in Trizol for further processing . RAW 264 . 7 or bone marrow derived macrophages ( BMDM ) were cultured in 6-well dishes ( unless otherwise indicated ) overnight prior to infection . Macrophages were infected with either 10∶1 or 100∶1 multiplicity of infection ( MOI ) with late log or stationary phase Brucella for times indicated and then harvested for RNA analysis or CFU evaluation . Cultures were incubated at 37°C with 5% CO2 . Select cultures were treated with 0 . 01 µg/mL tunicamycin ( Sigma ) 30 minutes prior to infection . Note: although gentamycin is routinely used during Brucella infections , it decreases detection of UPR induction ( particularly CHOP ) . The coding sequence of Bmel 1674 encoding TcpB1 was inserted into the Brucella plasmid pNstrcD [56] . The plasmid was electroporated into B . melitensis ΔTcpB1 by standard methods . RAW 264 . 7 or J774A . 1 ( both from ATCC ) mouse macrophage cell lines were seeded in 6 well tissue culture plates at 3×105 per well 1 day prior to infection . Cultures of B . melitensis , B . melitensis ΔTcpB1 , and B . melitensis ΔTcpB1+pNstrcD/BmeI1674 ( 3 ml each in BHI media with appropriate antibiotics ) were seeded 2–3 days before infection to be in late log phase at the time of infection . Macrophage cells were infected at 100 MOI and cultured for 24 h . Cells were then washed 1× in PBS , lysed and harvested in 1 ml/well of Trizol ( Invitrogen ) for RNA processing . Real time PCR: Following culture , supernatant was removed and samples were resuspended in TRIzol ( Invitrogen ) . RNA was purified according to manufacturer's instructions and treated with DNaseI ( Invitrogen ) to remove genomic DNA . RNA was reverse transcribed using random primers ( Promega ) . Relative cDNA was quantified using SYBR Green ( Bio-Rad ) and detection in MyiQ , or CFX96 real time PCR machines ( both Bio-Rad ) . Primers were designed using Beacon Design software ( Premier Biosoft ) and are as follows: 18S rRNA: forward , 5′-GGACACGGACAGGATTGACAG-3′ and reverse , 5′-ATCGCTCCACCAACTAAGAACG-3′ . Hprt1: forward , 5′-GTTAAGCAGTACAGCCCCCAAA-3′ and reverse , 5′-AGGGCATATCCAACAACAAACTT . BiP: forward , 5′-AGGATGCGGACATTGAAGAC-3′ and reverse , 5′-AGGTGAAGATTCCAATTACATTCG-3′ . CHOP: forward , 5′-CATCACCTCCTGTCTGTCTC-3′ and reverse , 5′-AGCCCTCTCCTGGTCTAC-3′ . ERdj4: forward , 5′-AGGGAAGGATGAGGAAATCG-3′ and reverse , 5′-ACTGTTGTTGCCGTTTGG-3′ . IL-6: forward , 5′-ACGATGATGCACTTGCAGA-3′ and reverse , 5′-GTAGCTATGGTACTCCAGAAGAC-3′ . XBP1 splicing was assessed through 3 assays: 1 ) Agarose gel assay: XBP-1 primers for conventional PCR: forward , 5′-ACACGCTTGGG- AATGGACAC-3′ and reverse , 5′-CCATGGGAAGATGTTCTGGG-3′ . PCR amplified cDNA was resolved on 3% gel and optical density ( OD ) quantified using Image Quant ( GE Healthcare ) . % XBP splicing is spliced cDNA OD/total ( spliced+unspliced ) OD×100 . 2 ) Quantification of separate species by Agilent . 3 ) qPCR assay: XBP1 ( t ) : forward , 5′-TCCGCAGCACTCAGACTATGT-3′ and reverse , 5′-ATGCCCAAAAGGATATCAGACTC-3′ . XBP1 ( s ) : forward , 5′-GAGTCCGCAGCAGGTG-3′ and reverse , 5′-GTGTCAGAGTCCATGGGA-3′ . % splicing = XBP1 ( s ) /XBP1 ( t ) ×100 [57] . Non-infected and B . melitensis infected RAW cells were harvested and resuspended in a buffer containing 20 mM Tris-HCI [8 . 0] and 0 . 5% SDS . Samples were boiled for 20 min . and mixed with equal amount of sample buffer . Cell lysates were resolved on a 4–20% SDS PAGE and transferred to immobilon PVDF membrane ( Millipore ) . The membrane was blocked with Tris-buffered saline containing 0 . 1% Tween 20 ( TTBS ) and 5% nonfat milk for 1 h at room temperature followed by three washes with TTBS . The membrane was incubated with anti-CHOP antibody ( Cell Signaling Technology ) in blocking buffer over night at 4°C . After washing three times with TTBS , the membrane was incubated with HRP-conjugated anti-mouse IgG ( Pierce ) in blocking buffer for 1 h at room temperature . After three washes with TTBS , protein bands were detected using SuperSignal West Pico Chemiluminescent Substrate according to manufacturer's instructions ( Pierce ) . The membrane was re-probed with anti-actin ( Santa Cruz Biotechnology ) . Chemiluminescence was detected by CL-XPosure Film ( Thermo Scientific ) . RAW264 . 7 cells , D17 cells or bone marrow derived macrophages were seeded into chamber slides ( Lab Tek and Ibidi ) and allowed to adhere 16–24 h . For infections , macrophages were infected ( 1000 MOI ) with either wild-type B . melitensis or B . melitensis containing a TcpB gene deletion for 24 h . Both strains express YFP under control of the trcD promoter . TUDCA ( 500 µg/mL ) pre-treatments were 30–60 min . For purified protein treatments , the medium was then replaced with fresh medium ( 1 ml ) containing purified maltose binding protein ( MBP ) , MBP-TcpB protein ( 10 or 50 µg/ml ) , or tunicamycin ( 1 or 10 µg/mL ) and the plates were incubated over night ( 12–24 h ) . The cells were washed 3× with PBS , fixed with 4% paraformaldehyde for 10 min , followed by permeabilization with 0 . 1% Triton X100 for 10 min . Cells were treated with blocking buffer containing 5% normal serum and 50 mM NH4CI in 1X PBS for 30 min , then washed and incubated with 1∶100 dilution of anti-calreticulin antibody ( Thermo Scientific ) in PBS containing 0 . 1% normal serum for 1 h . Cells were washed 3X with PBS and incubated with 1∶1000 dilution of Alexa Fluor 488 goat anti-rabbit IgG ( Invitrogen ) or secondary conjugated to DyLight 550 ( Thermo Scientific ) for 1 h , or anti-rabbit 550 ( Cell Signaling ) overnight , washed 3X with PBS , and mounted in ProLong Gold antifade reagent ( Invitrogen ) . Select samples were mounted in ProLong Gold antifade reagent with DAPI ( Cell Signaling ) . Images were collected using either a Radiance 2100 MP Rainbow confocal/multiphoton microscope ( Bio-Rad ) or Nikon A1R confocal laser microscope . To determine the effect of chemical chaperones on Brucella viability , Brucella were plated in 96 well dishes at 5×106 cells/well in RPMI containing serial dilutions of tauroursodeoxycholic acid ( TUDCA , Sigma ) . BacTiter-Glo assay ( Promega ) was performed to determine ATP content ( viability ) as assessed by luminescence . To determine the effect on RAW cell viability , cells were plated in 96 well plates at 104 cells/well one day prior to challenge . The medium was then replaced with fresh medium containing serial dilutions of chemical chaperones . CellTiter-Glo assay ( Promega ) was performed to determine ATP content ( viability ) as detected by luminescence . Effect of TUDCA on Brucella viability in broth was also determined utilizing this assay . To confirm TUDCA inhibition of UPR , RAW cells were pre-treated 30 min . with TUDCA , then stimulated with 10 µg/mL tunicamycin ( Sigma ) for 6 h or B . melitensis for 24 h prior to processing in TRIzol . Inhibition of replication: RAW 264 . 7 macrophages were plated in 24 well dishes at 5×105/well the day prior to infection . Cells were pre-treated with 500 µg/ml TUDCA ( 4 experiments ) or 4 mg/mL TUDCA ( 1 experiment , no significant RAW cell viability effect ) for 30 min . prior to infection with either 10 or 100 MOI of stationary phase B . melitensis . After 30 min . , cells were washed 4X with warm PBS and fresh media with 50 µg/ml gentamycin added with or without TUDCA . To evaluate colony-forming units ( CFU ) , cells were washed 3X with PBS and then lysed in 1% Triton-X 100 in water . CFU were determined by serial dilution plating on agar after 3–4 days . In parallel , samples were lysed in TRIzol to determine effect of TUDCA on UPR target gene induction at 24 h . Differences between data were evaluated using Students T-test with p<0 . 05 considered significant .
|
Brucella melitensis is an intracellular bacterium that invades and replicates within macrophages and dendritic cells . With over 500 , 000 new infections per year , brucellosis is the most prevalent zoonosis worldwide and incurs significant human morbidity and economic loss . The intracellular location of Brucella renders the organism resistant to antibiotics . A safe and effective human vaccine does not exist . Thus , better understanding of the host-pathogen interactions supporting establishment of the intracellular replicative niche is critical . In this study , we found that infection of macrophages with Brucella induces a host stress response called the Unfolded Protein Response ( UPR ) , a conserved stress response originating in the endoplasmic reticulum ( ER ) . Full induction of the UPR requires live bacteria and expression of a microtubule modulating protein , TcpB . Inhibition of the UPR with the drug tauroursodeoxycholic acid significantly diminished Brucella replication . Together these results suggest Brucella induces the UPR to enable its own replication within host macrophages . Thus the UPR may represent a novel therapeutic target for the treatment of brucellosis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
Brucella Induces an Unfolded Protein Response via TcpB That Supports Intracellular Replication in Macrophages
|
Nearly 45% of people living at risk for lymphatic filariasis ( LF ) worldwide live in India . India has faced challenges obtaining the needed levels of compliance with its mass drug administration ( MDA ) program to interrupt LF transmission , which utilizes diethylcarbamazine ( DEC ) or DEC plus albendazole . Previously identified predictors of and barriers to compliance with the MDA program were used to refine a pre-MDA educational campaign . The objectives of this study were to assess the impact of these refinements and of a lymphedema morbidity management program on MDA compliance . A randomized , 30-cluster survey was performed in each of 3 areas: the community-based pre-MDA education plus community-based lymphedema management education ( Com-MDA+LM ) area , the community-based pre-MDA education ( Com-MDA ) area , and the Indian standard pre-MDA education ( MDA-only ) area . Compliance with the MDA program was 90 . 2% in Com-MDA+LM , 75 . 0% in Com-MDA , and 52 . 9% in the MDA-only areas ( p<0 . 0001 ) . Identified barriers to adherence included: 1 ) fear of side effects and 2 ) lack of recognition of one's personal benefit from adherence . Multivariable predictors of adherence amenable to educational intervention were: 1 ) knowing about the MDA in advance of its occurrence , 2 ) knowing everyone is at risk for LF , 3 ) knowing that the MDA was for LF , and 4 ) knowing at least one component of the lymphedema management techniques taught in the lymphedema management program . This study confirmed previously identified predictors of and barriers to compliance with India's MDA program for LF . More importantly , it showed that targeting these predictors and barriers in a timely and clear pre-MDA educational campaign can increase compliance with MDA programs , and it demonstrated , for the first time , that lymphedema management programs may also increase compliance with MDA programs .
There are 1 . 3 billion people living at risk of infection with the parasites that cause lymphatic filariasis ( LF ) and an estimated 40 million suffering from the long-term complications of the disease [1] , [2] . In 2000 , the Global Programme for Elimination of LF ( GPELF ) began its campaigns to interrupt transmission of the parasite using a strategy of annual mass drug administration ( MDA ) to those at risk and to control or prevent LF-related disability through morbidity management programs [3] . India's National Vector Borne Disease Control Programme has scaled up MDA to interrupt LF transmission over the past several years and recently began adding albendazole to diethylcarbamazine ( DEC ) therapy where available with the monumental goal of providing mass drug treatment to all 590 million Indians living at risk for infection [4] . Although the program has distributed sufficient quantities of DEC tablets , problems have remained with achieving sufficient levels of adherence to DEC regimens in many regions in India [5]–[10] , including Orissa State [11] , [12] . Published estimates reporting drug coverage are often more accurately characterized as estimates of drug distribution which overestimate the actual drug consumption or compliance with MDA of the population [6] , [13] . Mathematical models suggest interrupting transmission is dependent on the baseline population prevalence of LF infection and on overall population compliance with MDA programs [14] , [15] . The lower the compliance with the MDA and the higher the baseline prevalence of LF , the more rounds of MDA required to interrupt transmission . Ensuring maximal compliance is critical to programmatic success . In some areas of India , the MDA program is restricted to tablet distribution , and issues such as drug adherence , drug side effects , and LF education of the populace are not comprehensively addressed [13] . For this reason , in 2007 the Church's Auxiliary for Social Action ( CASA ) partnered with the Indian Ministry of Health in Orissa State to enhance adherence to the DEC regimen . CASA developed a community-based educational campaign for the populace in three sub-districts in Khurda District of Orissa State . The campaign sought to increase awareness about the occurrence of the MDA , about the transmission and prevention of LF , about who should take DEC and the potential side effects , and about mosquito control . The message was distributed over a four-week period prior to the December 2007 MDA via radio and newspaper advertisements , street plays , leaflet distributions , broadcasting local songs incorporating health messages , posters , wall paintings , and village educational sessions . CASA partnered with the Centers for Disease Control and Prevention ( CDC ) to evaluate the effectiveness of their program . This evaluation found that adherence to the DEC regimen was less than 60% and failed to detect a significant impact of the community-based education campaign [16] . However , it identified barriers to and predictors of adherence . The major barriers were fear of side effects and a lack of recognition of the benefit of adherence . A model of predictors of adherence to the DEC regimen found that people who knew about the MDA in advance of its occurrence , people who knew that the MDA was for LF prevention , and people who knew that mosquitoes transmit LF were significantly more likely to adhere to the medication . The model also suggested that those who knew everyone was at risk for LF were also more likely to adhere , though this was not statistically significant . Other studies have found similar barriers and predictors using a variety of methodologies [5] , [9] , [11] , [12] , [17]–[21] , though relatively few provided a quantitative measure of association [22] , [23] . Based on these data , the CASA community-based educational message was refined to focus on these predictors and barriers . For the December 2008 MDA , CASA expanded its community-based educational campaign adding three new sub-districts to the original three in Khurda District . These new sub-districts received the same educational campaign described for the 2007 pre-MDA campaign , but the educational messages incorporated these refinements . Additionally , in early 2008 , CASA initiated a lymphedema management program which included both a community-based education component for the entire populace and patient self-care component focused on foot and leg hygiene for affected individuals and their families . This program was initiated in the three original sub-districts only . This evaluation was designed with the following objectives: 1 ) to assess the effectiveness of community-based LF education and community-based lymphedema management education in increasing compliance with the MDA program and 2 ) to validate the importance of predictors of and barriers to adherence to the DEC regimen identified in the previous evaluation .
The program was approved by the National Center for Zoonotic , Vector-borne , and Enteric Diseases ( NCZVED ) Human Subjects Committee at CDC , Atlanta , Georgia , USA , prior to the implementation of the survey . Permission for the survey was obtained from the Orissa State Department of Health and Family Welfare . Participants were asked to give their written consent prior to participation . For those unable to write , consent was documented by recording the person's fingerprint or marking the signature line with an ‘X’ and by countersignature of survey personnel . Consent procedures were approved by the Human Subjects Committee . The 2008 MDA for Orissa occurred from December 28th to 30th . The coverage survey occurred from February 19th to 28th , 2009 . A random 90-cluster sample design was utilized , with 30 villages selected in each of three areas: the community-based pre-MDA education plus community-based lymphedema management education ( Com-MDA+LM ) area , the community-based pre-MDA education ( Com-MDA ) area , and the Indian standard pre-MDA education ( MDA-only ) area . The Com-MDA+LM area included the three original sub-districts―Khurda , Balianta , and Balipatna―that received both the community-based pre-MDA educational campaign and the community-based lymphedema management program . The Com-MDA areas included the three new sub-districts―Bologarh , Begunia , and Jatni―that received only the pre-MDA educational campaign . The MDA-only area was composed of one sub-district―Banapur―which did not border any of the other six sub-districts and received only the standard Indian Ministry of Health MDA campaign . Villages were selected in each area using probability proportionate to size methodology [24] , [25] . In villages with hamlets , which are areas of a village separate from the main village , probability proportionate to size methodology was used to determine whether the hamlet or the main village was sampled . Fifteen households were randomly selected using the Expanded Programme on Immunization ( EPI ) random walk methodology [24] . Two quantitative surveys were performed: a household ( HH ) survey , in which every member of the household was included , and a knowledge , attitudes , and practices ( KAP ) survey , in which one person in each household over the age of 17 was randomly chosen to participate . Replacement of non-responders was not permitted . The survey was designed to detect a difference in drug adherence of 15% between the areas . Calculations were adjusted to allow for two 2-way comparisons and to account for a design effect of 12 , based on the design effect found in the 2008 study [16] . Thus the study had 80% power to detect 15% difference with an alpha of 0 . 025 if 3 , 181 persons were enrolled in each study area . Data were entered into EpiInfo v3 . 5 . 1 ( Stone Mountain , GA ) and analysis was performed in SAS v9 . 2 ( Cary , NC ) . All results presented from the HH and KAP surveys were adjusted for stratification and clustering , except for tests for medians . All KAP survey results were also weighted by the size of the eligible population in the household . For differences between the three areas , dichotomous variables were evaluated using chi-square tests or Rao-Scott Likelihood Ratio Tests and continuous variables were evaluated using tests for medians . Multivariable logistic regression analysis of KAP data was performed to assess predictors of adherence to DEC . All predictors that were statistically significant ( p≤0 . 05 ) in univariable analysis were included in the final model . Any demographic variable not found to be a univariable predictor of adherence but which differed across the three groups was included in the initial model . Interaction terms were created and removed by examining the Wald chi-squares for the individual components of the interaction terms . After evaluation of the interaction terms , demographic variables that were not found to be predictors of adherence were removed if removal did not change the adjusted odds ratios ( OR ) for the major predictors by at least 10% and if removal improved the precision of the estimates for the major predictors . The model was adjusted for weighting , clustering , and stratification .
In Com-MDA+LM areas 449 ( 99 . 7% ) households participated in the HH survey , in Com-MDA areas 427 ( 94 . 9% ) participated , and in MDA-only areas 409 ( 90 . 9% ) participated . There were 2949 , 2863 , and 2481 persons included in the survey , respectively . The groups were similar in age distribution ( median 29 years , range 0 . 1 years–105 years ) , and sex distribution ( 52 . 0% male ) . Tablets were received by 2784 ( 94 . 4% ) , 2671 ( 93 . 3% ) and 2105 ( 84 . 9% ) persons , respectively ( p<0 . 0001 ) . There were three ( 0 . 1% ) , 67 ( 2 . 3% ) and 49 ( 2 . 0% ) persons eliminated from further analysis because they did not live in the respective area at the time of the MDA . Adherence to the DEC regimen differed significantly between the three areas ( Table 1 ) , with 90 . 2% adherence in Com-MDA+LM areas , 75 . 0% adherence in Com-MDA areas , and 52 . 9% adherence in MDA-only areas . Among those who took DEC , 217 ( 3 . 6% ) reported side effects , the most common of which was headache ( 125 , 2 . 1% ) . All three groups reported similar rates of side effects ( p = 0 . 2 ) . No one required hospitalization for side effects . Persons who did not take DEC were asked why . All reasons provided by more than 5% of the population are shown in Table 1 . The most common reason given in all areas was fear of side effects , though persons in the Com-MDA+LM area were the least likely to give this reason . Com-MDA+LM persons were more likely to state that they were sick at the time of the MDA , which is a legitimate contraindication in the Indian program . In Com-MDA+LM areas 445 ( 98 . 9% ) persons participated in the KAP survey , in Com-MDA areas 423 ( 94 . 0% ) participated , and in MDA-only areas 401 ( 89 . 1% ) participated . One Com-MDA+LM person was eliminated from the analysis because her answers could not be weighted . The demographic breakdown of KAP participants is shown in Table 2 . There were statistically significant differences in the sex , age , caste , educational level , and literacy level distributions between the three groups . Although households reporting at least one household member with a swollen leg ranged from 9 . 7% to 18 . 6% to 23 . 5% , this difference was not statistically significant ( p = 0 . 28 for overall comparison across the three groups ) . Participants who complied with the MDA program were asked why they took DEC . The most common reasons given were as follows: 1 ) to prevent LF ( 463 , 48 . 0% ) , 2 ) because the MDA distributor told me to take DEC ( 344 , 32 . 9% ) , and 3 ) because a family member told me to take DEC ( 211 , 22 . 6% ) . Participants who did not take DEC were asked to specify why and what they would need to be told to change their minds . The top five reasons given were as follows: 1 ) fear of side effects ( 80 , 30 . 3% ) , 2 ) lack of trust of DEC ( 45 , 16 . 9% ) , 3 ) sick at the time of the MDA ( 29 , 9 . 5% ) , 4 ) not at home when DEC was distributed ( 25 , 9 . 2% ) , and 5 ) not sick and therefore DEC was not needed ( 25 , 9% ) . They reported they would comply if convinced that taking DEC would help them ( 151 , 52 . 3% ) , if convinced that taking DEC would help their family ( 53 , 17 . 3% ) , or if taught to manage side effects ( 22 , 9 . 8% ) . Participants were asked questions about their knowledge of LF , MDA , and lymphedema management . Responses are shown in Table 3 . Com-MDA+LM participants had greater knowledge than Com-MDA participants that LF is transmitted by mosquitoes , everyone is at risk for LF , and there are specific treatments for lymphedema such as leg exercises , leg washing , and leg elevation . Com-MDA participants were much more likely than MDA-only participants to know about the MDA in advance of its occurrence , mosquitoes transmit LF , and antibiotics can be used to help manage acute attacks . Com-MDA and MDA-only participants were equally likely to know everyone was at risk for LF . Demographic and knowledge variables were examined to determine if they were univariable predictors of adherence to the DEC regimen . Results are shown in Table 4 . Neither caste nor male sex was a univariable predictor . Some quartiles of age , having 11 to 12 years of education , reading well , and having a household member with lymphedema were found to be predictors . More importantly , five factors that could be addressed in educational campaigns were found to predict adherence . They are , in decreasing order of strength of association: knowing about the MDA in advance of its occurrence ( OR = 8 . 1; 95% CI: 5 . 2–12 . 6 ) , knowing the MDA was for LF ( OR = 7 . 5; 95% CI: 4 . 3–12 . 9 ) , knowing one or more components of lymphedema management ( OR = 7 . 4; 95% CI: 3 . 8–14 . 6 ) , knowing everyone was at risk for LF ( OR = 3 . 7; 95% CI: 2 . 5–5 . 3 ) , and knowing mosquitoes transmit LF ( OR = 3 . 2; 95% CI: 2 . 2–4 . 7 ) . Multivariable modeling was then performed . A statistically significant interaction between knowing about the MDA in advance of its occurrence and knowing everyone was at risk for LF was found and therefore was kept in the model . Male sex did not influence the model and was removed . Caste , literacy , and having a household member with lymphedema did not predict adherence in multivariable analysis . Some quartiles of age and multiple educational levels influenced adherence . Significant predictors , which could be addressed in an educational campaign , included knowing both about the MDA in advance and that everyone was at risk for LF ( adjusted OR = 16 . 1; 95% CI: 8 . 8–29 . 3 ) , knowing about the MDA in advance ( adjusted OR = 4 . 8; 95% CI: 3 . 8–8 . 1 ) , knowing everyone was at risk for LF ( adjusted OR = 2 . 2; 95% CI 1 . 0–4 . 8; p = 0 . 04 ) , knowing the MDA was for LF ( adjusted OR = 3 . 3; 95% CI: 1 . 7–6 . 6 ) , and knowing at least one component of lymphedema management self-care ( adjusted OR = 3 . 3; 95% CI: 1 . 6–6 . 9 ) . To further examine the impact of the lymphedema management programs on adherence , two sub-analyses were performed . In the first sub-analysis , the knowledge of univariable predictors of adherence was compared between those who knew at least one of the three components of lymphedema leg care and those who did not know any . Only persons in Com-MDA+LM and Com-MDA populations were included . As shown in Table 5 , persons who knew at least one component of leg care had greater knowledge of all four of the univariable predictors of DEC adherence . The second sub-analysis drew from this same population . Multivariable analysis which included all predictors of adherence from the main model was performed among those who had a household member with leg swelling and among those who did not . Having knowledge of at least one component of leg care predicted increased adherence both among those who had a household member with leg swelling ( adjusted OR = 11 . 1 , 95% CI: 1 . 4–86 . 1 ) , and among those without ( adjusted OR = 5 . 1 , 95% CI: 1 . 5–17 . 4 ) .
The evaluation of the December 2007 MDA in Orissa led to the description of several predictors of and barriers to compliance with the MDA program . These predictors and barriers were used to refine a pre-MDA community-based educational campaign that was then implemented in six blocks in Khurda District , three of which had received the early version of the campaign―the Com-MDA+LM area―and three of which were new to the campaign―the Com-MDA area . The results were remarkable . In the Com-MDA+LM areas MDA compliance increased from a 2007 baseline of 59 . 5% [16] to 90 . 2% , well above that target of 80 . 0% compliance among the entire population . There was also a marked increase in compliance in the Com-MDA areas to 75 . 0% , which is close to the target and much improved from the 2007 baseline of 52 . 2% [16] . It is likely that the baseline MDA compliance for rural areas in this district is around 52% , and it is clear that both intervention groups had a significant increase in adherence over that baseline . This study not only makes an important and direct contribution to the effort to interrupt the transmission of LF in India , it also serves as an example that can be used by other programs to overcome barriers to MDA compliance in affected populations . The KAP survey allowed identification of predictors of and barriers to adherence to a DEC regimen Factors identified in the previous evaluation were targeted by an educational campaign delivered one month prior to the 2008 MDA . The increased adherence during the 2008 MDA campaign provided not only the proof-of-concept that the targeted educational program worked , but it also validated the previously identified predictors and barriers . This assessment demonstrates how critical operational research is to any health program , particularly one whose success depends on changing health behaviors . Fortunately , this research can lead to simple and effective solutions . Developing messages that address key concepts for improving compliance with the MDA program is essential . In Orissa , these include: 1 ) making people aware of the occurrence of the MDA in advance of its occurrence―the CASA program is launched one month prior to the MDA , 2 ) making people aware of the purpose of the MDA medication , 3 ) making people aware that everyone is at risk for infection , 4 ) making people aware that one can be infected and still feel well , and 5 ) making people aware that side effects of DEC are infrequent and mild . Additionally , data from those who did not take DEC suggested that the medication's benefit needs to be personalized . The person who takes the medication needs to feel that they or a close family member stands to benefit directly . Lofty national goals did not speak to those who did not take DEC in this evaluation population . Individualized programs will need to be developed to address the specific needs of each location . One unique and important finding from the 2008 evaluation is that community-based lymphedema management programs positively affect MDA compliance independently of such programs' effects on the other predictors of compliance . Even after multivariable modeling controlling for all of the other LF and MDA knowledge predictors , knowing any one of the three components of the management of leg lymphedema predicted adherence to the DEC regimen . This positive impact of community-based lymphedema management education persisted even among those who had no household members with lymphedema . Additionally , the Com-MDA+LM area had the highest level of persons adhering to the DEC regimen in this study ( 90 . 2% ) . Admittedly , part of the explanation may be that the Com-MDA+LM had received a pre-MDA educational campaign two years in a row , but the campaign in the year 2007 , which did not focus on predictors of adherence , was largely ineffectual ( as evidenced by DEC adherence of 59 . 5% in the area in 2007 ) . Previous authors have suggested that morbidity control programs could improve MDA compliance [3] , [26] , but this study is the first to provide data wholly consistent with , if not unequivocally substantiating , that hypothesis . Perhaps these programs are effective because they help maintain awareness of LF and its chronic manifestations in the community and reinforce LF messages taught in the pre-MDA programs . Or perhaps they enhance trust at the community and individual level by providing programs benefiting a generally marginalized and stigmatized population , those who suffer from lymphedema and elephantiasis . Lymphedema management programs could provide an ideal platform for both LF and MDA education to improve MDA program compliance . As India approaches LF elimination , there will be a continued need to assist LF patients with clinical disease . Integrating lymphedema management with LF elimination efforts could be a more cost-effective way to ensure that MDA compliance remains high , even if political pressure to continue funding elimination efforts diminishes . There are several factors that could influence compliance that merit further comment . Persons in the Com-MDA+LM areas had the fewest number of people who reported no education and the highest number who reported reading well . In univariable analysis reading well influenced the decision to take DEC and education level had relatively little impact on the decision; in multivariable analysis the relationship reversed . Possibly the ability to weigh the risks and benefits of MDA compliance is more directly related to education level than to literacy . Additionally , the mechanisms utilized to distribute the educational message included many verbal routes ( i . e . street plays , auto-rickshaws , etc ) . However , an assessment of literacy , or health literacy , using a validated tool might allow a more thorough examination of this complex relationship . Multivariable analysis suggested that those with less education were more likely to comply . Perhaps those with less education are more likely to accept public health messages . It is important to note that although the relationship with education level is statistically significant , because of smaller numbers in each group the confidence intervals around the ORs are wide and in many cases approach one . It may be that the impact of education on compliance is much less than suggested by our analysis; this is an issue that should certainly be examined in future studies . Knowing a household member with leg edema could also influence one's perception of risk for LF . While the prevalence of leg edema in a household member did not differ statistically across the three groups , the highest prevalence was reported in the Com-MDA+LM group . Whether this represents actual increased prevalence or increased recognition of the condition because of the lymphedema management program is unclear . Even though this factor was found to be a predictor of MDA compliance in univariable analysis , it was not significant in multivariable analysis . One possible reason for this is that the CASA educational message emphasized that everyone was at risk for infection and that one might be infected even if one felt well . Finally , there was an interaction between knowing about the MDA in advance and knowing everyone was at risk for LF . Those who only knew everyone was at risk for LF had a small increase in MDA compliance . Those who only knew about the MDA in advance had a larger increase . Those who knew both had a synergistically larger increase . Why this was so is not clear . Perhaps those who understood both messages had a heightened sense of benefit or felt more empowered to achieve their own health goals because they felt at risk for infection and that they had the opportunity to avail themselves of preventive medication . Perhaps the interaction reflects the influence of another factor , such as an understanding of side effects and their management . In either case , the interaction points to the importance of addressing risk of LF and opportunity to access the beneficial MDA medication in any educational message . The limitations of this evaluation are similar to other retrospective evaluations that use the EPI random walk cluster method . Selection bias was reduced by defining a strict set of rules governing household selection and replacement of non-participants was not allowed . The evaluation was cross-sectional , so causality cannot be assumed . However , given that most of the predictors identified in this evaluation were the same as in the prior evaluation and that the knowledge of the predictors in the Com-MDA+LM area was higher in this evaluation that in the prior one , it is likely that the predictors are causal . The generalizability of the results may be limited to rural areas as urban areas were not included . Finally , although there is a definitive baseline MDA compliance for the Com-MDA+LM area , the baseline for the Com-MDA and the MDA-only areas are based on less direct empirical data . The Com-MDA baseline is derived from the Bologarh MDA compliance of 52 . 2% for the 2007 MDA . The fact that the compliance in Banapur for the 2008 MDA was 52 . 9% suggests that MDA compliance in rural areas of Khurda District is approximately 50–55% . Determining the predictors and barriers of adherence to the DEC regimen distributed in the MDA allowed for identification of key educational messages that were incorporated into a pre-MDA community-based LF educational campaign and resulted in a marked increase in regimen adherence . An added benefit was demonstrating that community-based lymphedema management programs independently enhanced adherence . Although further work is needed to determine exactly which components of lymphedema management programs influence MDA program compliance , one should not wait for those results before investing in such programs which address the twin goals of improving the lives of those suffering from filarial disease and increasing compliance with MDA programs to the level needed for the interruption of LF transmission .
|
Global elimination of lymphatic filariasis requires giving drugs at least annually to populations who live at risk of becoming infected with the parasite . At least 80% of people at risk need to take the drugs annually for 5 or more years to stop transmission of the infection . People suffering from the long-term effects of infection , such as swollen legs , benefit from programs that teach self-care of their affected limbs . In this study , we assessed the impact of an educational campaign that , after addressing previously identified predictors of compliance , significantly improved drug compliance . The specific factors improving compliance included knowing about the drug distribution in advance , knowing that everyone is at risk for acquiring the infection , knowing that the drug distribution was for lymphatic filariasis prevention , and knowing at least one component of leg care . We also found that areas with programs to assist people with swollen legs had greater increases in compliance . This research provides evidence that program evaluation can be used to improve drug compliance . In addition , our work shows for the first time that programs to benefit people with swollen legs caused by lymphatic filariasis also increase the participation of people without disease in drug treatment programs .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"public",
"health",
"and",
"epidemiology/preventive",
"medicine",
"public",
"health",
"and",
"epidemiology/global",
"health",
"public",
"health",
"and",
"epidemiology/infectious",
"diseases"
] |
2010
|
Increasing Compliance with Mass Drug Administration Programs for Lymphatic Filariasis in India through Education and Lymphedema Management Programs
|
Cryptococcus neoformans is an opportunistic fungal pathogen that causes serious human disease in immunocompromised populations . Its polysaccharide capsule is a key virulence factor which is regulated in response to growth conditions , becoming enlarged in the context of infection . We used microarray analysis of cells stimulated to form capsule over a range of growth conditions to identify a transcriptional signature associated with capsule enlargement . The signature contains 880 genes , is enriched for genes encoding known capsule regulators , and includes many uncharacterized sequences . One uncharacterized sequence encodes a novel regulator of capsule and of fungal virulence . This factor is a homolog of the yeast protein Ada2 , a member of the Spt-Ada-Gcn5 Acetyltransferase ( SAGA ) complex that regulates transcription of stress response genes via histone acetylation . Consistent with this homology , the C . neoformans null mutant exhibits reduced histone H3 lysine 9 acetylation . It is also defective in response to a variety of stress conditions , demonstrating phenotypes that overlap with , but are not identical to , those of other fungi with altered SAGA complexes . The mutant also exhibits significant defects in sexual development and virulence . To establish the role of Ada2 in the broader network of capsule regulation we performed RNA-Seq on strains lacking either Ada2 or one of two other capsule regulators: Cir1 and Nrg1 . Analysis of the results suggested that Ada2 functions downstream of both Cir1 and Nrg1 via components of the high osmolarity glycerol ( HOG ) pathway . To identify direct targets of Ada2 , we performed ChIP-Seq analysis of histone acetylation in the Ada2 null mutant . These studies supported the role of Ada2 in the direct regulation of capsule and mating responses and suggested that it may also play a direct role in regulating capsule-independent antiphagocytic virulence factors . These results validate our experimental approach to dissecting capsule regulation and provide multiple targets for future investigation .
Cryptococcus neoformans is an opportunistic fungal pathogen [1] . The disease it causes , cryptococcosis , is contracted by inhalation of infectious particles ( spores [2] or dessicated cells ) , which initiate a pulmonary infection . In the setting of immune compromise the fungus disseminates , with particular predilection for the central nervous system where it can cause a fatal meningoencephalitis . In otherwise healthy hosts , the infection may remain latent for extended periods , emerging in the event of immune compromise [3] . The impact of the disease is significant , especially in populations with limited access to health care , leading to an estimated 600 , 000 deaths per year [4] . A variety of factors have been implicated in cryptococcal virulence . These include melanin synthesis [5]; urease and phospholipase secretion [6] , [7]; titan cell formation [8] , [9]; and the ability to survive at host body temperature . Additionally , the main feature that distinguishes C . neoformans from other pathogenic fungi is an extensive polysaccharide capsule that surrounds the cell wall and is required for virulence [10] . Capsule size varies tremendously with growth conditions , becoming particularly large during mammalian infection [11] . Capsule expansion can be induced in vitro by mimicking aspects of the host environment such as low iron availability , the presence of mammalian serum , and physiological concentrations of carbon dioxide [12]–[14] . Strain virulence correlates with capsule size in vivo [15] , implicating the regulation of capsule formation as a critical factor in the pathophysiology of cryptococcal disease . Our current knowledge of capsule regulation derives primarily from studies where mutations of specific genes yield cells with abnormal capsules . A variety of readily assayed phenotypes that are related to the size or nature of the capsule ( including cell sedimentation behavior [16] , antibody reactivity [17] , India ink staining , and colony morphology ) has enabled the identification of a wide array of such mutants . Most of these have reduced virulence , emphasizing the central role of the cryptococcal capsule in pathogenesis . Capsule size is regulated by distinct and overlapping signaling pathways , including those typically associated with stress response . The best-characterized of these , the cAMP pathway , responds to amino acid starvation , low glucose , and elevated carbon dioxide [18] . Stimulation of this pathway leads to high intracellular cAMP levels , which activate the kinase Pka1 [19] . This enzyme in turn activates the C2H2 zinc finger transcription factor Nrg1 , leading to the transcriptional induction of genes that are directly involved in capsule assembly [20] . Pka1 also activates another transcription factor , Rim101 , which is necessary for capsule enlargement . Interestingly , activation of Rim101 requires elements of both the cAMP pathway and the pH-responsive Rim signaling pathway [21] . Deletion of the genes encoding Pka1 , Nrg1 , or Rim101 leads to reduced capsule size . Iron sensing mechanisms also influence capsule formation . Transcription factors Hap3 and Hap5 are involved in both iron homeostasis and capsule regulation; deletion of the corresponding genes leads to a reduction in capsule size [22] . In addition to Hap3 and Hap5 , the iron responsive transcription factor Cir1 also regulates capsule [23] , in part by transcriptionally regulating the cAMP pathway . Recently , ChIP-chip studies revealed that Cir1 is directly regulated by another transcription factor , Gat201 [24] . Strains lacking either Cir1 [23] or Gat201 are hypocapsular [25] . Capsule regulation is also influenced by the HOG pathway . Several proteins in this pathway ( including Hog1 , Pbs2 , and Ssk2 ) negatively regulate capsule size [26] . Epistasis analysis shows that the cAMP pathway is required for this HOG-dependent influence on capsule , but the mechanism of the cross-talk between these two central signaling pathways is unknown . Normal capsule formation also requires proteins in pathways related to temperature sensing [27] , sexual development [28] , and cell wall integrity [29] , [30] . More broadly , chromatin remodeling has been implicated in capsule regulation , by the observation that cells lacking the histone acetyltransferase Gcn5 are hypocapsular [31] . Gcn5 is a member of the well-conserved SAGA complex , which acts in transcriptional regulation from fungi to humans [32] . Sequence analysis suggests that other SAGA proteins are present in C . neoformans , but Gcn5 is the best conserved and the only one that has been characterized [31] . Over 60 genes have been identified as important players in capsule formation due to the effects of their deletion on capsule structure or morphology; we refer to such genes as ‘capsule-implicated’ ( Table S1 ) . However , because the majority of cryptococcal transcription factors and signaling proteins are uncharacterized , it is likely that important elements of the capsule regulatory network are missing from this group . Furthermore , some capsule-implicated genes may be required for other primary functions , such as cell wall synthesis , that have incidental effects on capsule formation . As reviewed above , components of several known signaling pathways are required for capsule formation , but there is no model that accounts for the integration of these pathways to regulate capsule growth . To begin constructing such a model , we have identified genes whose RNA levels are correlated with capsule size over a range of in vitro conditions . We term this set of genes the transcriptional signature of capsule . This signature includes previously capsule-implicated genes as well as multiple uncharacterized genes encoding putative regulatory factors . We chose to analyze one uncharacterized gene , ADA2 , which encodes a putative DNA-binding protein . We now show that Ada2 is a novel regulator of capsule and of other virulence-related features of Cryptococcus . Analysis of downstream targets of Ada2 and other capsule regulators by RNA-Seq and ChIP-Seq suggests the context of Ada2 in the capsule regulatory network and illustrates the effectiveness of this approach in unraveling complex regulatory networks .
We reasoned that the transcript abundance of many genes involved in the regulation and synthesis of capsule would correlate with capsule size across multiple growth conditions . To test this hypothesis , and potentially identify capsule regulatory genes beyond those previously reported , we cultured the C . neoformans serotype A reference strain H99 in four conditions known to stimulate capsule formation to varying degrees . For each condition , we also cultured the cells in a similar medium that stimulates capsule formation to a lesser extent . The eight conditions used were low iron medium ( LIM ) with and without the chelating agent ethylenediaminetetraacetic acid ( EDTA ) ; phosphate-buffered saline ( PBS ) with and without fetal bovine serum ( FBS ) ; Dulbecco's Modified Eagle's Medium ( DMEM ) in room air ( RA ) or in 5% CO2; and Littman's medium ( LIT ) with two concentrations of thiamine ( LO-THI / HI-THI ) . After 24 h the average capsule radius in each culture was assessed by light microscopy . The remaining cells were used to isolate total RNA for hybridization against a C . neoformans serotype A/D microarray ( http://gtac . wustl . edu/services/microarray/rna-analysis/cryptococcus-neoformans . php ) . To identify genes whose transcript abundance correlated with capsule size , we compared the transcription profiles over all eight conditions to the quantitative measurements of capsule radius ( Figure 1 ) . Our analysis revealed 880 genes whose transcript abundance correlated significantly with capsule size ( Table S2 ) , which we considered the transcriptional signature of capsule induction . Within this set , we identified 316 genes whose transcription correlated positively with capsule radius and 564 genes whose transcription correlated negatively . Among the positively correlated genes , most are involved in responses to stress , including DNA damage repair , trehalose biosynthesis and sugar transport . In contrast , many of the negatively correlated genes are involved in mitochondrial function and ribosome biogenesis . We expected that some of the genes in the transcriptional signature would specifically influence the formation of capsule ( see Discussion ) . Consistent with this hypothesis , the set of genes whose RNA levels correlated positively with capsule size was enriched for capsule-implicated genes ( p < 0 . 02; see Methods ) ; no such enrichment was observed among genes that correlated negatively . Positively correlated genes that are capsule-implicated included the genes encoding regulatory proteins Cir1 , Hap5 [22] , and Ste20 [33] and the phosphodiesterases Pde1 and Pde2 [34] ( see Discussion ) . The transcriptional signature of capsule included previously uncharacterized genes that encode putative transcription factors , signaling proteins , and sugar transporters ( Table S2 ) . It is likely that many of these genes are involved in capsule regulation and assembly . We were particularly interested in one previously uncharacterized gene , CNAG_01626 , which encodes a putative DNA binding protein . Expression of CNAG_01626 correlated positively with capsule size ( Figure 2 ) . For comparison , Figure 2 also shows the correlations obtained for two cryptococcal transcriptional regulators , CIR1 and SSN801 , whose roles in capsule regulation have previously been demonstrated . CIR1 showed significant positive correlation with capsule size , consistent with the hypocapsular phenotype of cir1Δ mutants [23] , while SSN801 exhibited a negative correlation with capsule size , consistent with the hypercapsular phenotype of the corresponding deletion mutant [25] . Given the strong correlation of CNAG_01626 transcription with capsule size , we suspected that the corresponding gene product was a regulator of capsule formation . Because this gene encodes multiple putative DNA-binding domains ( Myb-like and SWIRM ) , we further expected that it might act at a transcriptional level . This hypothesis was supported by the 33% homology we noted between the amino acid sequence predicted for CNAG_01626 and that of the Saccharomyces cerevisiae Ada2 protein . In S . cerevisiae , Ada2 is a member of the Spt-Ada-Gcn5 Acetyltransferase ( SAGA ) complex that mediates histone acetylation [35] . Within SAGA , Ada2 is required for proper catalytic activity of the acetyltransferase Gcn5 [36] . Based on the homology between the cryptococcal gene and S . cerevisiae ADA2 , we decided to refer to CNAG_01626 as ADA2 . Since transcription of the cryptococcal ADA2 gene positively correlates with capsule size , we hypothesized that deleting ADA2 would yield hypocapsular cells . To assess the role of this putative transcriptional regulator in capsule formation , we replaced the ADA2 genomic coding sequence with a nourseothricin-resistance marker ( NAT ) in the serotype A strain KN99α , derived from the serotype A reference strain H99 [37] . We then incubated ada2Δ mutant cells under capsule-inducing conditions and examined capsule size by negative staining with India ink . Consistent with the microarray analysis ( Figure 2 ) , ada2Δ mutant cells had dramatically reduced capsule compared to wild type ( Figure 3 ) . This phenotype was reversed by complementation with the ADA2 genomic coding sequence ( ada2Δ::ADA2 in Table S3; see Methods ) . To facilitate comparison of ada2Δ to strains lacking other capsule regulators , we also deleted CIR1 , NRG1 , and SSN801 in KN99α ( see Methods ) . Consistent with earlier reports , the ssn801Δ capsule was enlarged , while the cir1Δ and nrg1Δ capsules were reduced , similar to the capsule produced by ada2Δ ( Figure 3 , panel A ) . Having demonstrated that cryptococcal Ada2 influences capsule expansion , we proceeded to further investigate its role . Given the function of the SAGA complex in histone acetylation in S . cerevisiae [35] , we expected that cryptococcal Ada2 would reside in the nucleus . To test our hypothesis , we integrated a hemagglutinin ( HA ) epitope-tag sequence at the 3′ end of the ADA2 genomic coding sequence and examined the localization of the tagged protein ( Ada2-HA ) by immunofluorescence microscopy . Consistent with the nuclear role of Ada2 in S . cerevisiae , the tagged cryptococcal protein colocalizes with nuclear DNA ( Figure 4 ) . In S . cerevisiae , the SAGA complex activates transcription of stress-responsive genes by acetylating specific lysine residues at the N-terminal tails of histones H2B and H3 [35] , [38] . One of these modifications is the acetylation of lysine 9 of histone H3 ( H3K9 ) . To assess whether Ada2 is involved in similar histone acetylation in C . neoformans , we analyzed the abundance of acetylated H3K9 in the ada2Δ mutant by immunofluorescence microscopy using an antibody specific for this modification . We found that the fluorescence intensity of mutant cell nuclei was reduced by at least 50% compared to nuclei of both wild type and complemented cells ( Figure 5 ) , a result we confirmed on the population level by immunoblotting with the same antibody ( not shown ) . In contrast , H4 acetylation , which is not SAGA specific [39] , [40] , showed no difference between the ada2Δ mutant and either the wild type or complemented strains ( not shown ) . These results demonstrate the role of CNAG_01626 in histone acetylation , likely in the context of C . neoformans SAGA , and strongly support our identification of this novel capsule regulator as the cryptococcal homolog of the S . cerevisiae ADA2 . In S . cerevisiae and other fungi , the SAGA complex regulates the response to stress conditions such as elevated temperature , high salt concentration , and oxidative damage [41] , [42] . We found that the ada2Δ mutant grew normally compared to wild type on rich medium ( YPD ) at 30°C ( Figure 6 ) . However , the mutant exhibited a subtle growth impairment at 37°C , and a moderate attenuation of growth at 39°C . In all cases , the complemented strain behaved like wild type ( Figure 6 ) . To further compare the phenotype of ada2Δ to known fungal SAGA mutants , we next tested a panel of stress conditions for their effect on growth of the wild type , mutant , and complemented strains at both 30 and 37°C ( Figure 6 ) . We found that mutant cells were highly sensitive to alkaline pH , with no growth at pH 8 . 8 at 37°C , while growth at physiological or acidic ( 5 . 5 ) pH was like that of wild type ( not shown ) . Growth of ada2Δ at 37°C was also abolished when 6% ethanol was included in the medium , in notable contrast to the growth of wild type cells under this condition , and was impaired at 0 . 4 M CaCl2 . Conditions that challenge cell integrity , including media containing calcofluor white ( 0 . 2% ) , congo red ( 0 . 5% ) , low levels of SDS ( 0 . 01% ) , or high sorbitol ( 2 M ) , had no effect on mutant growth ( not shown ) . Similarly , KCl ( 1 . 2 M ) and NaCl ( 0 . 4 M ) did not alter growth ( not shown ) , although high NaCl concentrations ( 1 . 2 M ) did reduce growth at 37°C compared to wild type ( Figure 6 ) . The ada2Δ mutant also showed enhanced growth on caffeine and LiCl at 30°C , although this difference was not observed at the higher temperature tested ( see Discussion ) . The ability of C . neoformans to withstand nitrosative and oxidative stress is required for the virulence of this yeast [43] , [44] . We therefore tested the effect of Ada2 absence on cryptococcal sensitivity to compounds that induce such stress . Growth of the ada2Δ mutant was not affected by NaNO2 ( 0 . 5 mM ) at 30°C but exhibited a significant defect at 37°C . The mutant was highly sensitive to oxidative stress ( 0 . 5 mM H2O2 ) , with growth attenuated at 30°C and absent at 37°C . ( Figure 6 ) . We also examined the ability of this mutant to produce melanin , a feature of C . neoformans that is associated with virulence [5] . We observed no difference in melanin production on medium containing L-3 , 4-dihydroxyphenylalanine ( L-DOPA; not shown ) . Finally , we tested the sensitivity of the ada2Δ strain to several pharmacological agents . These included fluconazole , amphotericin B , and flucytosine , all antifungal compounds used to treat cryptococcal infections . Growth in all cases was comparable to that of wild type , in contrast to the increased fluconazole sensitivity observed upon deletion of ADA2 in Candida albicans [40] . We also tested the sensitivity of ada2Δ to FK506 , a compound that inhibits calcineurin signaling . A C . neoformans gcn5Δ strain has been shown to be FK506 sensitive , suggesting a defect in this pathway [31] ( see Discussion ) ; ada2Δ cells were even more sensitive to this compound ( data not shown ) . While back-crossing the ada2Δ mutant , we noticed that this strain was slow to filament . To investigate the potential role of Ada2 in cryptococcal sexual development , we crossed mating type a and α cells bearing the ada2Δ mutation to KN99a and KN99α cells and to each other ( Figure 7 ) . Deletion of ADA2 in either mating type dramatically impaired the formation of dikaryotic filaments in unilateral crosses between the mutant and wild type . A bilateral cross between two ada2Δ mutants of opposite mating type showed no visible hyphal development even after 13 days , while the complemented strain behaved identically to wild type . The ada2Δ mutant displays a smaller capsule , demonstrates reduced resistance to oxidative and nitrosative stress , and grows more slowly at 37°C compared to wild type . Based on these characteristics , we hypothesized that the mutant would also be attenuated for virulence . Indeed , we found that pulmonary growth of the ada2Δ mutant was impaired by almost 100-fold compared to the wild type and complemented strains in an inhalational mouse model of cryptococcosis ( Figure 8 , panel A ) , although it did grow slightly better than a completely acapsular mutant ( cap59Δ ) . To pursue this observation , we conducted a survival study with the same four strains . By three weeks post-inoculation , all mice infected with the wild type and complemented strains had succumbed to the infection ( Figure 8 , panel B ) . In contrast , mice infected with the ada2Δ or cap59Δ mutants remained healthy throughout the study , confirming the requirement for Ada2 in the virulence of this yeast . To identify the genes and processes regulated by Ada2 , we used RNA-Seq to perform transcriptome analysis of the ada2Δ mutant and wild type cells cultured in either capsule-inducing or capsule non-inducing conditions ( see Methods ) . The majority ( 92% ) of the resulting short reads mapped to the C . neoformans serotype A reference sequence [45] , indicating the excellent quality of the data . The average 300-fold coverage of the cryptococcal transcripts we obtained in these studies allowed confident sequence identification , and will help improve annotation of the C . neoformans genome . Gene expression analysis revealed 460 genes that were differentially expressed in the ada2Δ mutant compared to wild type under the capsule inducing condition; 675 genes were differentially expressed between the two strains under the capsule non-inducing condition ( Table S4 ) . We examined the genes whose expression was significantly affected in one or both conditions . Most of these ( 73% ) were regulated in a sign consistent manner in the two conditions ( e . g . if gene expression was reduced in the ada2Δ mutant in non-inducing conditions it was also reduced in the mutant in inducing conditions ) , although the magnitude of changes did vary . Gene ontology ( GO ) analysis ( see Methods ) indicated that processes significantly enriched in the response to loss of Ada2 included ribosomal protein synthesis , sugar transport , and carbohydrate metabolism . Consistent with the filamentation defect we observed in the ada2Δ mutant ( Figure 7 ) , we noted several genes downstream of Ada2 that are involved in cryptococcal sexual development ( Table 1 ) . Two mating type-specific genes ( encoding the homeodomain regulator , Sxi1α , and the pheromone receptor , Ste3α ) showed decreased expression in the ada2Δ mutant . A variety of genes that are independent of mating type but are implicated in the pheromone response pathway were also found to respond to loss of Ada2 ( Table 1 ) . Our initial interest in Ada2 was stimulated by its importance in capsule synthesis . In the ada2Δ mutant , we observed a reduction in transcript abundance for a number of genes that , when deleted , yield small capsules ( Table 1 ) . These observations are consistent with the hypocapsular and avirulent phenotypes of the ada2Δ mutant . The ada2Δ mutant also showed reduced expression for genes involved in oxidative stress; this agrees with the hypersensitivity to oxidative stress observed in the mutant and may also contribute to the avirulent phenotype . Expression of two genes ( BLP1 and GAT204 ) , which have recently been implicated in capsule-independent mechanisms of cryptococcal virulence [24] , was also reduced in the ada2Δ mutant ( see Discussion ) . To place Ada2 in the context of the broader capsule regulation network , we performed RNA-Seq analysis on mutants that lack the transcriptional regulators Cir1 and Nrg1 . We chose these transcription factors because , like the ada2Δ mutant , both the cir1Δ and the nrg1Δ mutants are hypocapsular , demonstrate attenuated avirulence , and exhibit defects in mating . We identified 1265 genes that were differentially expressed in the nrg1Δ mutant compared to wild type under the capsule inducing condition and 1084 under the non-inducing condition ( Table S5 ) . For the cir1Δ mutant these values were 1257 and 529 , respectively ( Table S6 ) . Cryptococcal sexual development is regulated by Cir1 and Nrg1 [23] , [20] , as well as by Ada2 ( Figure 7 ) . To identify common regulatory targets shared by these three transcription factors , we examined the gene expression data from the ada2Δ , cir1Δ , and nrg1Δ mutants . Among genes previously implicated in cryptococcal sexual development , we found that only SXI1α was downstream of all three regulators , with its transcription reduced in nrg1Δ and ada2Δ but increased in cir1Δ . Transcription of the pheromone receptor STE3α was similarly reduced in ada2Δ and elevated in cir1Δ although it was not significantly changed in nrg1Δ . Genes regulated by Nrg1 included the cell type-specific p21-activated protein kinase STE20α , as well as other mating type-independent genes that are involved in sexual development , but these were not regulated by Ada2 or Cir1 . By comparing mutants generated in the same strain background and grown in the same conditions , we were able to confidently identify capsule-implicated genes that are downstream of Cir1 or Nrg1 , some of which are also regulated by Ada2 . For example , Nrg1 and Ada2 share downstream targets that include CAS4 , CAS32 , CPL1 , MAN1 , NSTA , and CHS3 . Similarly , CAP10 , CAS1 , CAS4 , and CPL1 are all downstream of both Cir1 and Ada2 . Notably , CAS4 and CPL1 are shared targets of all three regulators ( Ada2 , Nrg1 and Cir1 ) . In addition to genes that are likely to be directly involved in capsule biosynthesis , we found many genes whose expression was affected by the loss of Cir1 or Nrg1 that are involved in regulating capsule formation . For example , the nrg1Δ mutant showed altered transcription of genes in the cAMP pathway , including increased transcription of RIM101 and decreased transcription of PKA2 and PDE2 . Consistent with previous reports [22] , we also observed altered transcript levels in the cir1Δ mutant that correspond to a number of pH-specific pathway genes , including RIM9 and RIM20 . The latter gene product is involved in proteolytic activation of Rim101 [21] . Finally , we discovered that Cir1 and Nrg1 regulate the expression of two HOG pathway genes: absence of either protein led to reduced transcription of HOG1 and increased transcription of PBS2 . Additionally , both Cir1 and Nrg1 appeared to enhance the expression of TUP1 [46] , which encodes a regulator that may operate in the HOG pathway . Interestingly , data from a previous microarray study indicated that ADA2 ( at that time uncharacterized ) increased in expression upon deletion of HOG pathway members ( HOG1 or SSK1 ) [47] ( see Discussion ) . Ada2 is required for the majority of H3K9 acetylation in C . neoformans ( Figure 5 and immunoblotting data not shown ) . We reasoned that localizing Ada2-dependent occurrences of this modification would lead us to genes that are directly regulated by Ada2 . We therefore used chromatin immunoprecipitation ( ChIP ) to isolate DNA directly associated with acetylated H3K9 in ada2Δ and wild type cells that we could analyze by short read sequencing ( ChIP-Seq ) . We obtained 84 million short reads from our ChIP-Seq studies , which we aligned to the serotype A reference sequence and analyzed to identify genomic regions with statistically significant coverage ( “peaks” ) in IP samples compared to input DNA . From triplicate experiments , we identified an average of 2014 peaks in wild type cells , compared to only 364 in ada2Δ . This 82% reduction is consistent with our earlier observations on the Ada2-dependence of most H3K9 modification ( Figure 5 ) . Consistent with H3K9 acetylation in S . cerevisiae [48] , [49] , the majority of the peaks identified in wild type ( 75% ) were within 500 bp of at least one transcription start site ( TSS ) as annotated [45] ( see Table S7 for a summary of TSS neighboring H3K9 acetylation for wild type or mutant ) . Most peaks in wild type were also located in the 5′ region immediately downstream of the TSS , with a strong depletion near the TSS and a modest enrichment upstream of the TSS ( Figure 9 , black bars in panel A ) . In contrast , only 28% of peaks in ada2Δ were within 500 bp of a TSS and almost none of these were downstream of the TSS ( Figure 9 , red bars in panel A ) . Thus , not only is histone acetylation in this mutant depleted throughout the genome , the pattern of acetylation is also changed , with the most dramatic depletion occurring in the region immediately downstream of the transcription start site . The loss of histone acetylation in ada2Δ cells suggested Ada2-dependent transcriptional activation at specific loci ( see example in Figure 9 , panel B ) . We anticipated that some of these genes would also show reduced transcription by RNA-Seq in the ada2Δ mutant; this was indeed the case ( p < 0 . 003 ) . In contrast , we found no such relationship for genes with increased transcription in ada2Δ ( i . e . , genes that are directly or indirectly repressed by Ada2; p > 0 . 99 ) , consistent with the generally activating function of the SAGA complex . Overall , we found that genes differentially expressed in the ada2Δ deletion strain that also lost histone acetylation near the TSS were twice as likely to exhibit reduced transcriptional abundance as genes that did not lose histone acetylation ( Figure S1 ) . We were particularly interested in genes that were activated by Ada2 according to our RNA-Seq analysis and also showed Ada2 dependent H3K9 acetylation in our ChIP-seq analysis . This set is significantly enriched for genes that are directly regulated by Gat201 ( p < 0 . 0001 ) , including BLP1 and GAT204 [24] . The genes implicated by both RNA-seq and ChIP-seq also include a number with known capsule phenotypes , such as CPL1 , HXT1 , STE3α , and UGT1 ( see Discussion ) .
We analyzed gene expression in C . neoformans yeast cells cultured over a diverse set of growth conditions that stimulate capsule production to varying degrees and identified a transcriptional signature of capsule formation . Gene ontology ( GO ) analysis shows that this signature is enriched for genes involved in stress response , as expected from the conditions we used to induce capsule formation . The signature also contains a significant number of genes that have previously been implicated in capsule regulation; the expression of most of these correlates with capsule in a manner consistent with the null phenotype . The phosphodiesterases Pde1 and Pde2 are exceptions to this pattern: their transcript levels correlated positively with capsule size , while their disruption increases capsule size [34] . Pde1 and Pde2 hydrolyze cAMP to AMP and thereby inhibit the cAMP-dependent activation of regulators known to stimulate capsule formation . Elevated levels of cAMP occurring under capsule inducing conditions may lead to elevated transcription of PDE1 and PDE2 , which would ultimately attenuate the cAMP signal . Feedback inhibition of cAMP signaling via post-translational activation of phosphodiesterases has been documented in both S . cerevisiae and C . neoformans [50] , [34] . One sequence in the transcriptional signature that correlated significantly with capsule size ( Figure 2 ) encoded the putative transcriptional regulator , Ada2 . This protein has been characterized most extensively for its role within the SAGA complex , which broadly regulates the transcription of genes involved in stress response and development in multiple organisms [32] . This pattern holds true for C . neoformans , based on the increased sensitivity of mutants that lack either ADA2 ( this work ) or GCN5 [31] to reactive oxygen species , ethanol , alkaline pH , elevated temperature , and CaCl2 ( Figure 6 ) . All of these sensitivities are shared by S . cerevisiae SAGA mutants [42] , [51] , and the last two also are shared by SAGA mutants in other fungi including C . albicans , S . pombe and S . kluyveri [40] , [42] . Despite many conserved functions of the SAGA complex across fungal species , several phenotypes of ada2 mutants in C . neoformans differ markedly from those observed in other fungi , perhaps reflecting the specific evolutionary pressures of the cryptococcal niche . Whereas ada2 mutants in C . neoformans display increased caffeine resistance ( Figure 6 ) , for example , disruption of SAGA components in S . cerevisiae , S . pombe and S . kluyveri has the opposite effect . Also , ada2 mutants in C . neoformans show an increase in LiCl resistance but no change in KCl resistance ( Figure 6 ) , while other fungi defective in SAGA typically exhibit normal growth in LiCl but are KCl sensitive relative to wild type [42] . Interestingly , ada2 mutants in C . neoformans have wild type sensitivity to fluconazole in contrast to ada2 mutants in C . albicans , which have increased sensitivity [40] . Finally , C . neoformans ada2Δ differs from other fungi in its regulation of sexual development . In S . pombe , the ada2Δ mutant is enhanced for mating , probably through a mechanism that does not directly involve histone acetylation [52] . In the C . neoformans ada2Δ strain , we instead found dramatically decreased sexual development ( Figure 7 ) , reduced transcript abundance of the pheromone receptor STE3α , and loss of H3K9 acetylation at the STE3α promoter . These results suggest that these two fungi differ in both the direction and the mechanism of Ada2's influence on sexual development . Recently , another component of the SAGA complex , Gcn5 , was shown to play a role in capsule formation and virulence in C . neoformans [31] . H99 cells lacking Gcn5 , like our mutant lacking Ada2 , are hypocapsular and hypovirulent . To compare the roles of these proteins , we examined genes that are differentially expressed by ada2Δ and gcn5Δ upon growth in DMEM , using our RNA-Seq data for ada2Δ ( Table S4 ) and published microarray data sets for gcn5Δ [31] . We found a significant overlap in the sets of genes whose expression is affected by each mutation ( p < 1e-5 ) , supporting the idea that some genes are jointly regulated by Gcn5 and Ada2 , probably due to the coordinated role of these proteins in SAGA-mediated histone acetylation . In addition to shared characteristics , we observed important differences between the ada2Δ and gcn5Δ mutants at both the phenotypic and transcriptional levels . The ada2Δ mutant is more resistant to high temperature , showing ∼10-fold growth inhibition on rich medium at 39°C compared to wild type ( Figure 6 ) , a condition where gcn5Δ does not grow at all [31] . In contrast , ada2Δ is more sensitive than gcn5Δ to the calcineurin inhibitor FK506 . ( The minimal inhibitory concentration ( MIC ) for gcn5Δ is 10-fold below that of its H99 parent [31] , while the MIC for ada2Δ ( performed as in [31] ) is at least 67-fold below that of KN99α; data not shown . ) . We also found that expression of both STE3α and SXI1α responds to the loss of Ada2 ( Table 1 ) , whereas no sexual development genes have been reported to be downstream of Gcn5 [31] . Consistent with this difference , ada2Δ is severely defective in filamentation ( Figure 7 ) while gcn5Δ filaments normally ( T . R . O'Meara and J . A . Alspaugh , personal communication ) . Furthermore , two genes involved in the recently described ‘antiphagocytic response’ [24] , GAT204 and BLP1 , showed a loss of both H3K9 acetylation ( Figure 9 , panel B and Table S7 ) and expression ( Table 1 ) in ada2Δ but no change in expression in gcn5Δ [31] . It will be interesting to determine whether these transcriptional differences manifest phenotypically . The phenotypic differences between ada2Δ and gcn5Δ may be due to Gcn5-independent functions of Ada2 in C . neoformans . Acetylation at some loci may rely on Ada2 partnering with a histone acetyltransferase ( HAT ) other than Gcn5 , or it may be that the regulation of these loci is independent of acetylation altogether . For example , in S . cerevisiae Ada2 regulates gene silencing by preventing the spread of repressive chromatin [53] . Such mechanisms remain to be investigated in C . neoformans . Given the importance of SAGA in virulence , the roles of Ada2 , Gcn5 and other SAGA subunits in C . neoformans biology are worthy of further investigation . After identifying Ada2 as a novel regulator of capsule , we sought to identify elements downstream of it in the capsule regulatory network . To do this , we performed RNA-Seq on the ada2Δ mutant and wild type strains , considering genes differentially expressed between these two strains to be downstream of Ada2 . To identify probable direct targets of Ada2 , we performed ChIP-Seq using antibodies specific for H3K9 acetylation , comparing the ada2Δ mutant and wild type strains . We reasoned that genes that lose histone acetylation near their transcription start sites in the ada2Δ mutant are likely direct targets of Ada2 via the SAGA complex or another histone acetyltransferase ( HAT ) complex involving Ada2 . The ada2Δ mutant strain revealed a dramatically altered landscape of H3K9 acetylation compared to the wild type , with more than an 80% reduction in acetylated sites across the genome ( see Results ) and even greater reduction around transcription start sites ( Figure 9 ) . This nearly total loss of H3K9 acetylation in the ada2Δ mutant is consistent with the established global HAT activity of SAGA in S . cerevisiae [54] , [55] . In contrast to its broad histone modification activity , SAGA only influences expression of 10% of S . cerevisiae genes [39] . RNA-Seq analysis of the ada2Δ mutant strain revealed that Ada2 influences transcription of 14% of the genes in C . neoformans , indicating that the transcriptional regulatory role of SAGA in C . neoformans is also locus specific . ChIP-Seq data further suggest that Ada2 exerts the minority of its influence through direct regulation: only 3% of cryptococcal genes exhibit both altered H3K9 acetylation and expression in ada2Δ cells , while 11% exhibit altered expression only . This large indirect response could be mediated in part via the 8 putative transcription factors that Ada2 directly regulates as evidenced by our studies . Consistent with the activating role of SAGA , the set of genes with reduced expression in ada2Δ was significantly enriched for those that lost H3K9 acetylation . ( In contrast , genes with increased expression in ada2Δ showed no significant overlap with those that lost H3K9 acetylation . ) Some genes , including the capsule-implicated gene UGT1 , showed increased expression together with loss of H3K9 acetylation in the ada2Δ mutant , perhaps because H3K9 acetylation at certain loci makes repressor binding sites more accessible . Alternatively , these genes may be directly activated by Ada2 through H3K9 acetylation yet also indirectly repressed by Ada2 , which could yield net repression . We observed phenotypic changes in the ada2Δ mutant in sexual development , capsule formation , stress response , and virulence; we also found genes with known roles in these processes to be directly regulated by Ada2 as evidenced by ChIP-Seq and RNA-Seq . For example , we found that Ada2 directly regulates genes encoding proteins implicated in capsule formation , including HXT1 [56] , CPL1 [25] , and UGT1 [57] , consistent with the capsule defect of the ada2Δ mutant ( Figure 3 ) . We also identified the gene encoding pheromone receptor Ste3 as a direct target of Ada2 in the mating type α ( MATα ) cells used in these studies , consistent with the observed filamentation defect in ada2Δ ( Figure 7 ) . Ste3 has also been implicated in mating in MATa [28] . Ste3a has further been shown to regulate virulence factors including titan cell [8] and capsule formation [28] , although no such relation has been reported for Ste3α . If Ada2 also regulates Ste3a then it may additionally influence capsule via this pathway in MATa cells . Future studies of MATa ada2Δ mutants will be needed to address this possibility . Gat201 is a GATA family transcription factor reported to act as a positive regulator of capsule [25] . Interestingly , we observe a significant overlap in the genes that are directly activated by Ada2 ( as shown by ChIP-Seq ) and those that are direct targets of Gat201 ( by ChIP-chip [24] ) , including the antiphagocytic genes BLP1 and GAT204 . Since the SAGA complex typically works in concert with other transcription factors , this suggests that Ada2 may work with Gat201 to activate transcription . It may be that Gat201 recruits Ada2 in the context of SAGA for these purposes . Alternatively , another factor may recruit the SAGA complex , which then enables Gat201 to bind . To explore the interplay between regulatory pathways we considered two transcription factors , Cir1 [23] and Nrg1 [20] , which like Ada2 enhance both capsule and mating responses . In the set of genes regulated by Cir1 and Nrg1 ( Table S5 and Table S6 ) , we identified two that encode proteins in the HOG pathway , Hog1 and Pbs2; both Cir1 and Nrg1 transcriptionally repress Pbs2 and activate Hog1 . Cells lacking either Pbs2 or Hog1 show increased capsule formation and sexual development [26] . Furthermore , both ADA2 and GCN5 were shown in earlier work to be transcriptionally repressed by Hog1 under nutrient rich conditions [47] . This observation , in conjunction with our data , suggests that the HOG pathway may regulate capsule and mating via Ada2 ( Figure 10 ) . Our transcriptional analysis suggests that Nrg1 and Cir1 operate on the HOG pathway through a shared incoherent feed-forward loop , by transcriptionally activating Hog1 and simultaneously repressing Pbs2 . In nutrient rich conditions , Hog1 is constitutively phosphorylated by Pbs2 and represses mating and capsule . The logic of this circuit implies that in capsule inducing conditions Cir1 and Nrg1 repress transcription of PBS2; this leads to reduced levels of phosphorylated Hog1 , thus derepressing ADA2 transcription and enhancing capsule formation . Simultaneously transcription of HOG1 is increased , leading to an even greater abundance of unphosphorylated Hog1 . This increase in Pbs2 substrate may allow rapid restoration of the transcriptional repression of Ada2 once the environmental cues for capsule induction are no longer present . Although transcript levels of ADA2 were not significantly altered in the nrg1Δ and cir1Δ mutants at the 90-minute time point that we tested , ADA2 expression may be affected by these mutations at later time points . Future studies will also be needed to determine whether the influences of Cir1 and Nrg1 on PBS2 and HOG1 result from direct or indirect regulation , and to better characterize the exact structure and function of this hypothesized regulatory circuit . This model , rich in testable hypotheses , illustrates the power of combining RNA-Seq and ChIP-Seq data in an integrated analysis . Our identification of Ada2 in the capsule transcriptional network validates our strategy for probing capsule regulation and suggests that it may be valuable in studying the regulation of other processes that are important in microbial pathogenesis . These studies also lead in numerous exciting directions for the future . Our parallel comparison of multiple mutants in the same strain background and growth conditions has allowed us to identify previously unobserved relationships among capsule regulators , which we look forward to testing . Our analysis of the transcriptional signature of capsule induction also suggests multiple potential transcription factors that can be pursued to further probe the complex confluence of pathways that lead to capsule synthesis , and our implementation of ChIP-Seq in C . neoformans demonstrates a high-resolution way for differentiating direct from indirect regulatory relationships . Overall , our work highlights the power of integrative transcriptome analysis to dissect regulatory networks in C . neoformans and beyond .
All animal studies were reviewed and approved by the Animal Studies Committee of Washington University School of Medicine and conducted according to the National Institutes of Health guidelines for housing and care of laboratory animals . All chemicals were from Sigma , primers were from Invitrogen , and restriction enzymes were from New England Biolabs unless otherwise noted . All kits and enzymes were used according to manufacturer recommendations unless otherwise specified . All strains used in this study are capsule serotype A , which causes the majority of illness in immunocompromised patients [58] , and are listed in Table S3 . Microarray experiments to identify the transcriptional signature of capsule were performed with C . neoformans H99 and mutants were constructed in C . neoformans KN99 . All cells were grown with continuous shaking ( 230 rpm ) at 30°C in YPD medium ( 1% w/v yeast extract , 2% w/v peptone , 2% w/v glucose ) , or at 30°C on agar plates ( YPD medium with 2% w/v agar ) . As appropriate , media were supplemented with either 100 µg/ml of nourseothricin ( from Werner BioAgents ) or 100 µg/ml of Geneticin ( G418; from Invitrogen ) . Genetic crosses were performed at room temperature ( RT ) in the dark on V8 agar plates ( 5% v/v V8 juice , 0 . 05% w/v KH2PO4 pH 5 , 4% w/v agar ) as described [37] . To induce expression of genes involved in capsule formation , cells cultured overnight in YPD were collected by centrifugation , washed in DMEM , and adjusted to 4 × 107 cells/ml in DMEM . This cell suspension was first incubated at 30°C in room air for 2 hr , then shifted to 37°C with 5% CO2 for 1 . 5 hr . Conditions used for phenotypic testing of mutants are detailed in Text S1 . Approximately 2 × 108 cells were collected by centrifugation , suspended in TRIzol reagent ( from Invitrogen ) , and subjected to mechanical lysis by bead beating at 4°C with 0 . 5-mm glass beads for 1 min , followed by a 2-min rest , for a total of 5 cycles . Following lysis , total RNA was extracted according to the manufacturer's instructions . Residual DNA was removed from the RNA preparation by treatment with the Turbo DNA-free kit ( from Ambion ) according to the manufacturer's instructions . H99 cells were cultured overnight at 37°C in the following eight conditions: low iron medium with or without both 500 mM ethylenediaminetetraacetic acid ( EDTA ) and 10 mM bathophenanthroline disulfonate ( BPDS ) ; phosphate-buffered saline ( PBS ) with or without 10% v/v fetal bovine serum; Dulbecco's Modified Eagle's Medium ( from Sigma ) in room air or 5% CO2; and Littman's medium [13] with either 0 . 01 µg/ml or 1 µg/ml thiamine . All experiments were performed in triplicate . Total RNA was isolated from each culture and hybridized to a C . neoformans serotype A/D microarray against a shared reference pool of RNA as described [59] . Slides were scanned on a Perkin-Elmer ScanArray Express HT scanner to measure Cy3 and Cy5 fluorescence as described [59] . Normalization of the raw spot intensities was performed using LIMMA [60] . Normalization was performed using normexp with an offset of 50 followed by Loess and values for replicate probes on the array were averaged to represent expression of the associated gene . The correlation between gene expression and capsule radius ( which was measured for each sample at the time of RNA isolation ) was assessed using SAM [61] and statistical significance was calculated using a false discovery threshold of 5% . A hypergeometric test was applied to determine the enrichment of capsule-implicated genes ( genes whose mutation yields an alteration in capsule size or morphology; Table S1 ) in the positively and negatively correlating sets of genes . The complete array data set is available at GEO accession number GSE31911 . The C . neoformans H99 reference sequence was accessed through the Fungal Genome Initiative database at the Broad Institute of MIT and Harvard available at <http://www . broadinstitute . org/science/projects/projects> . Cryptococcal genomic DNA was isolated as described [62] and a split-marker approach [63] was used to replace each genomic coding sequence of interest with a nourseothricin resistance marker ( NAT ) by homologous recombination . Each mutant was also labeled with a unique signature tag by incorporating a 13-bp tag sequence ( see Table S8 ) and an 18-bp priming site ( 5′ - AGAGACCTCGTGGACATC - 3′ ) immediately downstream of NAT . We also used the split-marker gene replacement approach to introduce a single copy of the hemagglutinin ( HA ) epitope-tag sequence at the 3′ end of the ADA2 genomic coding sequence . Details of strain construction are provided in Text S1 , Table S9 and Table S10 . Cells cultured in YPD were washed extensively in DMEM , then adjusted to 106 cells/ml in DMEM and incubated for 24 hours at 37°C with 5% CO2 . Capsules were visualized by negative staining with India ink , and a minimum of 100 randomly chosen cells were imaged with identical acquisition settings on a Zeiss Axioskop 2 MOT Plus wide-field fluorescence microscope . Capsule radius was calculated as half the difference between the capsule diameter and the diameter of the cell body . Cells were cultured overnight in YPD , and the expression of genes involved in capsule formation was induced as described above . Cells were then collected by centrifugation , washed in PBS , adjusted to 3 × 108 cells/ml in 4% w/v formaldehyde buffered in PBS , and incubated for 1 hr with rotation . Fixed cells were collected by centrifugation ( 1 min , 400 × g ) , washed extensively in PBS , adjusted to 3 × 108 cells/ml in Lysis Buffer ( 50 mM sodium citrate pH 6 . 0 , 1 M sorbitol , 35 mM β-mercaptoethanol ) plus 25 mg/ml Lysing Enzymes ( from Trichoderma harzianum ) , and incubated for 1 hr at 30°C . Digested cells were collected by centrifugation ( 3 min , 400 × g ) , washed with HS Buffer ( 100 mM HEPES pH 7 . 5 , 1 M sorbitol ) , and resuspended in 100-200 µl of HS Buffer . The cell suspension was spotted in 20-µl aliquots on a glass microscope slide coated with 0 . 1% w/v poly-L-lysine and incubated for 20 min at RT . All subsequent treatments and washes were performed by the application of 20-µl volumes and incubation at RT , and were followed by aspiration . The slides were first treated with HS Buffer containing 1% v/v Triton X-100 and incubated for 10 min; they were then washed with PBS and treated with Blocking Buffer ( 5% v/v goat serum , 0 . 02% v/v Tween-20 in PBS ) for 1 hr . Cells were next labeled with either a high-affinity rat anti-HA monoclonal antibody ( 0 . 2 µg/ml in Blocking Buffer; from Roche ) , a rabbit anti-acetyl-Histone H3 polyclonal antibody ( 0 . 5 µg/ml in Blocking Buffer; from Millipore ) , a rabbit anti-acetyl-Histone H4 polyclonal antibody ( 1 µg/ml in Blocking Buffer; from Millipore ) , or Blocking Buffer alone overnight in a moist chamber at 4°C . Cells were then washed with Blocking Buffer , and treated for 1 hr in the dark with either Alexa Fluor 594 goat anti-rat IgG or Alexa Fluor 594 goat anti-rabbit IgG ( 2 µg/ml in Blocking Buffer; from Invitrogen ) . Next , cells were again washed with Blocking Buffer , counterstained with 4′ , 6-diamidino-2-phenylindole ( DAPI; 5 µg/ml in PBS ) for 20 min in the dark , washed with PBS , allowed to air-dry , and mounted in Prolong Gold ( from Invitrogen ) . Brightfield and fluorescence images were acquired simultaneously on a Zeiss Axioskop 2 MOT Plus wide-field fluorescence microscope . All samples were imaged with identical acquisition settings . Two types of animal studies were performed , both in compliance with all institutional guidelines for animal experimentation . For a short term model of fungal survival in the mouse lung , strains to be tested were cultured overnight in YPD medium , collected by centrifugation , washed in PBS , and diluted to 2 . 5 × 105 cells/ml in PBS . For each strain , eight 4–6 week-old female C57Bl/6 mice ( from Jackson Laboratories ) were anesthetized with a combination of ketaset-HCl and xylazine , and inoculated intranasally with 50 µl of the prepared yeast suspension . Three animals from each cohort were sacrificed at 1 hr post-inoculation; the remaining five were sacrificed after 7 days . Lungs were harvested following sacrifice , and homogenized in PBS . Serial dilutions of the homogenate were plated on YPD agar for determination of colony-forming units ( CFU ) . Initial inocula were also plated to confirm CFU . To assess longer-term affects of cryptococcal infection , each strain was cultured and prepared as above , with the exception that the cells were diluted to 2 × 106 cells/ml in PBS . Ten 4–6 week-old female A/Jcr mice ( from the National Cancer Institute ) were anesthetized as described above and inoculated intranasally with 100 µl of the prepared cell suspension . The animals were weighed within 1 hr post-inoculation , and subsequently on every other day . Mice were sacrificed if weight decreased to a value less than 80% of peak weight ( an outcome which in this protocol precedes any signs of disease ) or upon completion of the study . Initial inocula were plated to confirm CFUs . Cells were cultured overnight in YPD , and grown for 90 minutes in either capsule-inducing ( DMEM , 37°C , 5% CO2 ) or capsule non-inducing ( DMEM , 30°C , room air ) conditions prior to isolation of total RNA . A minimum of two biological replicates were performed for each mutant ( ada2Δ , nrg1Δ and cir1Δ ) and four for wild type . PolyA+ RNA was purified from total RNA using the Dynabeads mRNA Purification Kit according to the manufacturer's instructions ( from Invitrogen ) . Each sample was resuspended in 2 µl of 100 mM zinc acetate and heated at 60°C for 3 minutes to fragment the RNA by hydrolysis . The reaction was quenched by the addition of 2 µl volumes of 200 mM EDTA and purified with an Illustra Microspin G25 column ( from GE Healthcare ) . First strand cDNA was made using hexameric random primers and SuperScript III Reverse Transcriptase ( from Invitrogen ) according to the manufacturer recommendations , and the product was treated with E . coli DNA ligase , DNA polymerase I , and RNase H to prepare double stranded cDNA using standard methods . The cDNA libraries were end-repaired with a Quick Blunting kit ( from New England BioLabs ) and A-tailed using Klenow exo- and dATP . Illumina adapters with four base barcodes were ligated to the cDNA and fragments ranging from 150-250 bp in size were selected using gel electrophoresis as recommended by the manufacturer . The libraries were enriched in a 10-cycle PCR with Phusion Hot Start II High-Fidelity DNA Polymerase ( from Finnzymes Reagents ) and pooled in equimolar ratios for multiplex sequencing . Single read , 36-cycle runs were completed on the Illumina Genome Analyzer IIx . Sequenced reads were aligned to the C . neoformans H99 reference sequence [45] using Tophat [64] . Reads that aligned uniquely to the reference sequence were considered for gene expression quantification with Cufflinks [65] using the current genome annotation provided by the Broad institute . Gene expression was normalized using the Cufflinks provided option for quartile normalization . Differential expression analysis comparing mutant to wild type was performed with LIMMA [60] and ELNN [66] using a 5% false discovery rate . Genes whose expression was found to be significantly changed by either analysis method were counted as differentially expressed . RNA-Seq data is available at GEO accession number GSE32049 . GO enrichment analysis was performed by assigning GO categories to each gene according to the Broad Institute's PFAM annotations using the mapping provided by the Gene Ontology project ( http://www . geneontology . org/external2go/pfam2go ) . A hypergeometric test was applied for each GO category , the resulting p-values were corrected for multiple hypothesis testing , and a cutoff of 0 . 05 was used to determine significance . Wild type and ada2Δ cells were cultured in triplicate overnight in YPD , and grown in capsule-inducing conditions ( DMEM , 37 °C , 5% CO2 ) for 90 minutes . Cells were then fixed for 5 min in 1% ( v/v ) formaldehyde , and the reaction quenched with a final concentration of 125 mM glycine . Fixed cells were collected by centrifugation , washed with PBS , and resuspended in Buffer A ( 50 mM HEPES pH 7 . 5 , 140 mM NaCl , 1 mM EDTA , 1% v/v Triton X-100 , 0 . 1% w/v sodium deoxycholate ) supplemented with protease inhibitors and 20 mM sodium butyrate ( a histone deacetylase inhibitor ) . The cell suspension was subjected to mechanical bead-beating with 0 . 5-mm zirconium silicate beads for 2 min at 4°C , followed by a 2-min rest , for a total of 10 cycles . Chromatin was then sheared by sonicating the lysate for 30 sec at 40% power output , followed by a 1-min rest on ice , for a total of 40 cycles , and the lysate clarified by centrifugation . A fraction of the sheared chromatin was reserved as an input sample and the remainder was used for immunoprecipitation . Acetylated histone H3 was immunoprecipitated overnight with anti-acetyl-H3 ( K9 ) antibody ( from Millipore ) tethered to protein-A sepharose ( 10 ml in a total volume of 700 ml ) . The beads were next washed sequentially in Buffer A , Buffer B ( 50 mM HEPES pH 7 . 5 , 500 mM NaCl , 1 mM EDTA , 1% ( v/v ) Triton X-100 , 0 . 1% ( w/v ) sodium deoxycholate ) , Buffer C ( 10 mM Tris-HCl pH 8 . 0 , 250 mM LiCl , 1 mM EDTA , 0 . 5% ( v/v ) NP-40 , 0 . 5% w/v sodium deoxycholate ) , and Buffer D ( 10 mM Tris , 1 mM EDTA ) , and immunoprecipitated protein was eluted with Buffer E ( 50 mM Tris pH 8 . 0 , 10 mM EDTA , 1% ( w/v ) SDS ) . Crosslinked DNA from input and IP samples was released by incubating the eluate at 65°C overnight , and extracted with a solution of phenol/chloroform/isoamyl alcohol ( 25∶24∶1 ) prior to ethanol precipitation and resuspension in water . Mock IP reactions with no antibody yielded no measurable product ( not shown ) and were not quantified further . ChIP-DNA for input and IP samples was end-repaired with Klenow DNA Polymerase and the DNA was purified with AMPure XP System beads ( Beckman Coulter Genomics ) and modified with A-tails using Klenow exo- before ligation to adapters to incorporate 7-base index sequences using T4 DNA ligase ( Enzymatics ) . Adapter addition was confirmed on an Agilent 2100 bioanalyzer , and the DNA was PCR-amplified and then gel purified to remove adapter dimers and select sizes optimal for high-throughput sequencing ( 150 to 300 bp ) . Libraries were 12-way multiplexed on an individual lane of an Illumina Hi-Seq 2000 flow cell , resulting in approximately 7 million 42-bp single ended reads per sample . Reads generated from the input and IP samples were aligned to the C . neoformans serotype A reference sequence [45] using Bowtie [67] . Reads that mapped to multiple genomic loci were discarded . Peak calling was performed using MACS [68] with a significance threshold of 1 × 10-10 . To assess gross differences between the mutant and wild type , the average number of peaks over the three biological replicates of each strain was compared . Peaks were associated with specific genes if the peak center fell within 500 bp of the gene transcription start site according to the current annotation by the Broad Institute [45] . ( For genes with unannotated 5′-UTRs this may correspond to the translation start site . ) Ada2-dependent peak loss was identified by cases where a gene in two of the three wild type biological replicates possessed a neighboring peak and no peak was found to neighbor the gene in any of the three ada2Δ mutant replicates . ChIP-Seq data is available at GEO accession number GSE32075 .
|
Cryptococcus neoformans is a fungal pathogen that causes serious disease in immunocompromised individuals , killing over 600 , 000 people per year worldwide . A major factor in the ability of this microbe to cause disease is an extensive polysaccharide capsule that surrounds the cell and interferes with the host immune response to infection . This capsule expands dramatically in certain growth conditions , including those found in the mammalian host . We grew cells in multiple conditions and assessed gene expression and capsule size . This allowed us to identify a ‘transcriptional signature’ of genes whose expression correlates with capsule size; we speculated that a subset of these genes acts in capsule regulation . To test this hypothesis , we characterized one previously unstudied gene in this signature and found it to be a novel regulator of capsule expansion , fungal virulence , and mating . This gene encodes cryptococcal Ada2 , a well-conserved protein that regulates genes involved in stress response and development . We used phenotypic analysis , RNA sequencing , and chromatin-immunoprecipitation sequencing ( ChIP-Seq ) to situate Ada2 in the complex network of genes that regulate capsule and other cryptococcal virulence factors . This approach , which yielded insights into the regulation of a critical fungal virulence factor , is applicable to similar questions in other pathogens .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome",
"expression",
"analysis",
"cellular",
"stress",
"responses",
"genetic",
"networks",
"gene",
"regulation",
"regulatory",
"proteins",
"dna-binding",
"proteins",
"microbiology",
"histone",
"modification",
"genome",
"analysis",
"tools",
"molecular",
"genetics",
"chromatin",
"mycology",
"proteins",
"gene",
"expression",
"microbial",
"pathogens",
"regulatory",
"networks",
"biology",
"pathogenesis",
"molecular",
"biology",
"transcriptomes",
"microarrays",
"biochemistry",
"systems",
"biology",
"cell",
"biology",
"genomics",
"molecular",
"cell",
"biology",
"computational",
"biology",
"genetics",
"and",
"genomics"
] |
2011
|
Toward an Integrated Model of Capsule Regulation in Cryptococcus neoformans
|
The continued northwards spread of Rhodesian sleeping sickness or Human African Trypanosomiasis ( HAT ) within Uganda is raising concerns of overlap with the Gambian form of the disease . Disease convergence would result in compromised diagnosis and treatment for HAT . Spatial determinants for HAT are poorly understood across small areas . This study examines the relationships between Rhodesian HAT and several environmental , climatic and social factors in two newly affected districts , Kaberamaido and Dokolo . A one-step logistic regression analysis of HAT prevalence and a two-step logistic regression method permitted separate analysis of both HAT occurrence and HAT prevalence . Both the occurrence and prevalence of HAT were negatively correlated with distance to the closest livestock market in all models . The significance of distance to the closest livestock market strongly indicates that HAT may have been introduced to this previously unaffected area via the movement of infected , untreated livestock from endemic areas . This illustrates the importance of the animal reservoir in disease transmission , and highlights the need for trypanosomiasis control in livestock and the stringent implementation of regulations requiring the treatment of cattle prior to sale at livestock markets to prevent any further spread of Rhodesian HAT within Uganda .
Human African trypanosomiasis ( HAT ) , or sleeping sickness , is caused by two sub species of a hemoflagellate parasite that are transmitted by tsetse flies . Trypanosoma brucei rhodesiense causes an acute disease in eastern sub-Saharan Africa and has a reservoir in wild and domestic animals while Trypanosoma brucei gambiense causes a chronic form of the disease in western and central sub-Saharan Africa . Uganda has had the misfortune to sustain active transmission of both types of the disease: T . b . gambiense in the north west and T . b . rhodesiense in the south east [1] . To date , however , Rhodesian and Gambian HAT have not co-existed in any area of Uganda , which is fortunate since the two forms of HAT are diagnosed and treated differently and geographical location forms the basis of diagnostic tool selection for the confirmation of diagnosis [2] . Uganda has experienced a resurgence of HAT in the past two decades . Since HAT ( caused by T . b . rhodesiense ) was introduced into Tororo District in 1987 , the disease has persistently spread northwards into previously unaffected areas of Uganda [3] , [4] . Since the disease imparts a considerable burden on the health systems of the poor , rural communities that it affects , the expansion of the T . b . rhodesiense focus is a persistent concern . The Northwards spread of disease has narrowed the area between the active foci of Rhodesian and Gambian HAT , with an estimated 150 km now separating the two forms of the disease [3] . Evidence suggests that the introduction of Rhodesian HAT into Soroti district could be attributed to the movement of untreated cattle from endemic areas through the local livestock market [5] , [6] . The further spread into Kaberamaido , Dokolo , Lira and Amolotar districts raised the possibility of the potential overlap of the two types of the disease and stimulated the creation of a Public Private Partnership , Stamp Out Sleeping Sickness , to control the disease spread by treating the animal reservoir of infection [7] . It is essential that the dynamics of disease spread are understood if HAT is to be controlled in Uganda . A comprehensive understanding of the factors involved in the disease's spatial distribution and movements will enable more effective targeting of control efforts . The spatial distribution of HAT is driven by complex interactions of many factors . The occurrence of disease in an area is dependent on the establishment of disease transmission , which in turn is reliant on the suitability of an area for the disease . Within affected areas , a spatially varying intensity of transmission can result in the heterogeneous village level prevalence of disease . These two processes giving rise to i ) the establishment of HAT transmission and ii ) the heterogeneous prevalence of HAT in an area are likely to be driven by different environmental , climatic and social factors associated with the presence and density of tsetse flies [8]–[11] , the introduction of the parasite , the presence of reservoir host species and the frequency of human-fly contact [12] . Spatial analysis and geographic information systems ( GIS ) have been applied increasingly to infectious disease epidemiology in recent years , including to the analysis of HAT [6] , [12]–[15] , animal trypanosomiasis [16]–[18] and tsetse distribution data [19] , [20] . However , the factors that control the heterogeneous distribution of HAT within small areas are poorly understood , though this knowledge would be of practical use for the targeting of control efforts and the prevention of further spread . Previous studies have linked the distribution of Rhodesian HAT in Uganda with proximity to areas of swamp and low population densities [14] , [15] . Distance to the local HAT treatment centre has also been found to have a confounding effect due to issues of health care accessibility [14] . In addition , several studies have examined the distribution of the tsetse fly vector , with a number of environmental variables found to have significant correlations with their distribution , including the normalised difference vegetation index ( NDVI – a measure of the amount of green vegetation ) , humidity [21] , temperature , rainfall [22] and elevation [23] , utilising a variety of data sources , including remotely sensed data . The spatial distribution of T . b . rhodesiense HAT in two newly affected districts of Uganda ( Kaberamaido and Dokolo ) was examined in relation to several environmental , climatic and social variables . Prevalence of HAT was then predicted spatially to highlight areas with the potential for high prevalence and to enable the targeting of future control efforts . The utilities of two different methodologies were compared: a two-step regression method and a traditional one-step regression method . The two-step regression was used to allow the separate analysis of factors governing the occurrence and prevalence of HAT . The prevalence analysis in the two-step regression model was conducted solely on areas that had a high predicted probability of occurrence . This was anticipated to provide an increase in predictive accuracy ( for predicted prevalence ) due to the exclusion of large areas with little or no HAT transmission .
A handheld global positioning system ( GPS: Garmin , E-trex ) was used to geo-reference the central point of all villages within the study area with guidance from local government staff . Coordinates were taken in the WGS84 geographical coordinate system in decimal degrees ( data were re-projected to Universal Transverse Mercator for the calculation of distances ) . Comprehensive HAT hospital records were collected in collaboration with the Ugandan Ministry of Health from the two HAT treatment centres serving the study area; Lwala Hospital ( Kaberamaido district ) and Serere Health Centre IV ( Soroti district ) . To maintain anonymity of subjects and patient confidentiality and to adhere to the International Ethical Guidelines for Biomedical Research Involving Human Subjects , no patient names were recorded within the database or as part of the data collection process . The hospital records were matched with the geo-referenced villages by cross-referencing each case's village of residence with the names from the geo-referenced villages . This resulted in a spatially referenced dataset of all patients residing within the study area who had received a diagnosis of HAT ( normally using light microscopy ) . Cases occurring from February 2004 ( when the first cases were reported ) to December 2006 were included in the analysis . Cases diagnosed later than December 2006 were excluded because a control programme was instigated in September 2006 that involved the mass treatment of cattle in the study area and adjoining districts . By decreasing the prevalence of human infective T . b . rhodesiense in the reservoir , the control programme resulted in an altered epidemiology of HAT within the study area in the subsequent year and so may have affected the results of the regression analyses . The geo-referenced HAT case data were visualised using ArcMap 9 . 1 ( ESRI , Redlands , CA ) . External covariate datasets as listed in Table 1 were collected and linked with the HAT case data by village . Several temporal Fourier-processed indices were obtained from Advanced Very High Resolution Radiometer ( AVHRR ) imagery: land surface temperature ( LST ) , NDVI and middle-infrared ( MIR , AVHRR channel 3 ) . NDVI is a measure of the amount of green vegetation [25] and reflectance in the MIR band has also been linked to vegetation cover [26] . Both vegetation cover ( in terms of suitable tsetse habitat ) and temperature have been shown to influence the distribution of tsetse [22] . Temporal Fourier processing reduces the number of data to be processed by eliminating redundancy and characterising seasonality . The minimum , mean , maximum , phase ( the timing of the cycle ) and amplitude ( the amount of variation around the mean ) of the annual and biannual cycles were used for each of LST , NDVI and MIR . Full details regarding these data can be found in Hay et al [27] . NDVI was also calculated using the red and near-infrared wavebands of a Landsat ETM+ image ( which has a finer spatial resolution than AVHRR imagery ) [28] using the following formula: NDVI = ( near-infrared−red ) / ( near-infrared + red ) [25] . Predicted tsetse suitability maps were obtained from the Food and Agricultural Organization [29] . This dataset was the result of a predictive model ( using tsetse fly distribution data with environmental , climatic and demographic covariates ) , and its reliability for the study area depends on the availability of training data from this area during the model development . Elevation [30] , population density [31] and nighttime lights data [32] ( which has been demonstrated to be a proxy for poverty [33] ) were also obtained for use in the analyses . Distances to physical features ( in km ) were calculated . Land cover data [34] were used to calculate distance to gazetted land , rivers , bush , woodland , swamps , permanently wet land and seasonally wet land . Several of these variables ( bush , woodland , swamps and seasonally wet land ) were the result of a quantitative interpretation of remotely sensed images along with ground data and supplementary data layers and , thus , their accuracy may be variable . These covariates were selected as potential tsetse habitats to investigate the effect of proximity of villages to these types of landcover on HAT occurrence and prevalence . In addition , distances to the closest livestock market and health centre ( of any type ) were calculated using the coordinates of each of these features that were obtained during fieldwork . The distance to the closest health centre ( of any type i . e . not necessarily trained or equipped to diagnose or treat HAT ) was used to deal with the confounding effect of access to health care . The distance to the closest livestock market was included to investigate the possibility that cattle movements in this area may have caused or contributed to the introduction and establishment of HAT transmission , as was found in a neighbouring district [6] . The distance to the HAT treatment centre was not used as there was only one treatment centre within the study area and an additional treatment centre in the neighbouring district serving the study population , which would affect the final predictions and prevent extrapolation over a larger area . The covariates used are listed in Table 1: all were continuous variables . In addition , village population data from the most recent national census were obtained from the Uganda Bureau of Statistics [35] . Exploratory analysis was conducted for each of the covariates: i ) scatter plots to examine relationships with HAT prevalence; ii ) box and whisker plots to examine the distributions of covariate data in villages which have had cases of HAT compared to villages which have not and iii ) visualisation of the geographical distributions of the outcome variables in relation to the external covariates . Seventeen covariates were selected for use in the regression analyses ( Table 1 ) based on observed relationships with HAT occurrence and prevalence and previous knowledge of significant variables from published research . The statistical modelling was carried out using logistic regression: a generalised linear model used for the analysis of binomial data such as disease occurrence ( outcome variable can take one of two possible values ) or disease prevalence ( where the outcome is bounded between zero and one ) [36] . The modelling process describes the variability in the response variable as a function of the explanatory variables . Odds ratios ( ORs ) are calculated by exponentiating the regression parameters associated with each covariate; these illustrate the strength and direction of associations between the explanatory and outcome variables . An OR of one indicates no association , an OR greater than one indicates a positive association with the odds of disease and an OR less than one indicates a negative association [36] . The size of the OR signifies the strength of the association; for example an OR of 0 . 5 would mean that every increase of one unit in the explanatory variable relates to a 50% reduction in the odds of disease . Likewise , an OR of 1 . 5 would show a 50% increase in the odds of disease for an increase of one unit for the explanatory variable . The intercept term can be interpreted as the odds of disease when all the explanatory variables are ( hypothetically ) zero . Statistical significance was judged at the 95% level in all analyses . All statistical analyses were carried out using R statistical software [37] , and the main steps are summarised in Figure 2 . This methodology comprised two logistic regression models applied sequentially ( first analysis , Figure 2 ) . An initial model was fitted that predicted probability of HAT occurrence using the HAT status of all villages in the study area as the outcome of interest . Villages for which at least one case of HAT was reported during the study period were classified as case villages , while villages for which no cases were reported were treated as controls ( giving a binary outcome ) . The two-step model was developed to test its predictive capability against a traditional regression analysis and to investigate aspects of the underlying epidemiology affecting the spatial heterogeneity in disease occurrence ( which villages had been affected by HAT ) as well as prevalence ( how intense was the transmission within affected areas ) which are confounded in a one-step approach . Forwards stepwise addition beginning with the null model ( no explanatory variables ) was used in the model fitting . At each step the variable resulting in the greatest reduction in deviance was selected . A Chi-squared likelihood ratio test was used to compare models , and additional explanatory variables were accepted only if this test was significant and the covariate was significant within the model . Any variables that lost significance in subsequent steps were removed from the model . The stepwise addition of plausible interaction terms ( if interaction is present the effect of one variable on odds of disease changes in relation to the effect of another variable ) was then carried out in the same manner after the variables were centred ( variable mean was subtracted from each value ) . The sensitivity ( true positive rate ) and specificity ( true negative rate ) of the fitted model were calculated for a variety of cut-off points ( the value of the predicted probability of occurrence above which a location would be defined as a case village ) using the predicted and observed values , and plotted against the cut-off points . The cut-off point where the sensitivity and specificity crossed was selected as a suitable cut-off point for the classification of case and non-case villages: this point maximises both the specificity and the sensitivity of the classification of locations . A 10-fold cross-validation ( where predicted values are compared with observed values ) was performed using ten random sub-divisions of the dataset . The area under the receiver operator characteristic curve ( AUC ) was calculated; this value gives a measure of the overall performance of the model in classifying villages . An AUC of 1 indicates perfect discrimination between case and control villages , and an AUC of 0 . 5 illustrates a model that is in effect worthless for discrimination purposes . The resulting regression equation ( probability of occurrence as a function of the explanatory variables ) was used to predict probability of occurrence of HAT across a grid with an area of 30 , 000 km2 ( including the study region ) and a 1 . 1 km cell size ( this was the minimum spatial resolution from the covariate datasets ) . All villages within the study area lying within an area of high predicted probability of occurrence ( probability of occurrence above the selected cut-off value ) were extracted for use in the second step of the analysis . The outcome variable for the second step of the two-step regression was defined as prevalence of HAT ( number of cases divided by village population ) . Prevalence data from all villages within areas of high predicted probability of occurrence were included in the model , including those with no reported cases ( i . e . a reported prevalence of zero ) . Forwards stepwise addition was used in the model fitting procedure , as for the first step . For this section of the analysis , the distance to health centre variable was forced into the model ( regardless of it's significance ) to ensure that access to health care was controlled for in the final results . The fitted model was used to predict the prevalence of Rhodesian HAT across the same area as was used in the first step . For the one-step analysis ( second analysis , Figure 2 ) , the same methodology was used as the second step of the two-step regression , using prevalence data from all villages .
Four covariates were found to influence significantly the occurrence of HAT across the study area ( p<0 . 05 ) as shown in Table 2 . Occurrence of HAT was negatively correlated with distance to the closest livestock market , with a 21% reduction in odds of disease for every kilometre increase in distance when accounting for the additional variables . This was found to interact ( the effect of one variable on odds of disease changes in relation to the effect of another variable ) with maximum NDVI , which also demonstrated a negative correlation with HAT occurrence . In addition , occurrence was positively correlated with minimum LST and negatively correlated with distance to the closest health centre . For prediction purposes , the selected probability cut-off point for the prediction of areas suitable for transmission was 0 . 2 , and model diagnostics indicated that the model provided a reasonable fit to the data , and reliable predictions ( AUC: 0 . 87 , 10-fold cross-validation estimate of accuracy: 85% ) . The predicted suitability for transmission across the study area using the specified model is illustrated in Figure 4 . The prediction was used to create a mask over the study area; all areas with a predicted probability of occurrence less than 0 . 2 were excluded . 279 villages lay within the area defined as having a high probability of occurrence . However , seven of those villages had no population data and so were excluded from the remaining analysis leaving 272 villages . The results from the second ( prevalence ) model are shown in Table 3 . HAT prevalence was significantly correlated with nine variables in addition to distance to the closest health centre that was negatively correlated and of borderline significance ( p = 0 . 05 , variable forced into the model ) . Prevalence was negatively correlated with distance to the closest livestock market with every additional kilometre resulting in a 20% decrease in odds of disease . This was shown to interact with distance to the closest area of woodland , which in turn showed a positive correlation with prevalence . In addition , HAT prevalence was negatively correlated with distance to the closest area of bush and maximum NDVI and positively correlated with NDVI phase of annual cycle , NDVI annual amplitude , LST phase of annual cycle , LST annual amplitude and minimum LST . The two-step regression analysis resulted in a correlation between observed and predicted prevalence of 0 . 57 ( a value of 1 indicates perfect correlation and 0 no correlation ) . The model had a small tendency to over predict prevalence with a median error of 0 . 05% ( error calculations are based on prevalence per 100 population and so are expressed as a percentage ) . The mean absolute error for the predicted prevalence per 100 population was 0 . 24% . The scatter plot of predicted prevalence against observed prevalence ( Figure 5 ) shows a tendency for over-prediction of prevalence in villages with an observed prevalence of zero . The predicted prevalence from the two-step analysis is shown in Figure 6 . Nine variables were shown to be significantly associated with prevalence of HAT across the study area using the one-step regression , as shown in Table 4 . HAT prevalence was negatively correlated with distance to the closest livestock market with a 21% reduction in odds of disease for every kilometre increase in distance . This was shown to interact significantly with both NDVI phase of annual cycle and distance to the closest area of woodland , both of which were also negatively correlated with prevalence . Additionally , prevalence was negatively correlated with maximum NDVI , mean LST and distance to the closest health centre . HAT prevalence was positively correlated with minimum LST , LST phase of annual cycle and LST annual amplitude . The correlation between predicted and observed prevalence values was 0 . 58 indicating a modest linear association . The model was slightly biased with a very small tendency to over-predict prevalence ( median error = 0 . 02% ) and the mean absolute error was 0 . 13% ( calculated based on prevalence per 100 population and so expressed as a percentage ) . The scatter plot of predicted prevalence against observed prevalence values ( Figure 7 ) illustrates that many of the errors are associated with over-prediction for villages with observed prevalence of zero . Figure 8 shows the predicted prevalence across the study area using the final prevalence model . To allow a direct comparison of the predictive accuracy of the two methodologies , the one-step model was used to calculate predicted prevalence for the villages with high predicted probabilities of occurrence from the two-step analysis ( i . e . excluding areas with a predicted probability of occurrence of less than 0 . 2 ) . The correlation between predicted and observed prevalence was 0 . 50 , lower than that for the two-step regression method ( 0 . 57 ) . Again , the model was shown to have a tendency to over predict prevalence , with a median error of 0 . 05% ( calculated using prevalence per 100 population ) . The mean absolute error was 0 . 24% , equal to the mean absolute error from the two-step regression methodology .
Spatial determinants for HAT are poorly understood across small areas . This study examined the relationships between Rhodesian HAT and several environmental , climatic and social factors in two newly affected districts , Kaberamaido and Dokolo . The application of a two-step regression approach for the prediction of HAT prevalence in a newly affected area of Uganda allowed the investigation of factors influencing the occurrence and prevalence of HAT separately , and overall resulted in a slight increase in predictive accuracy when compared to a one-step analysis in areas with high predicted probability of occurrence . Each of the models has illustrated an increased risk of HAT in villages closer to livestock markets than in villages further away , suggesting the persistent spread of Rhodesian HAT in Uganda may have resulted from the continued movement of untreated cattle . The two-step regression model gave a slight increase in predictive accuracy in comparison with the one-step analysis with a correlation between fitted and observed prevalence values of 0 . 57 for the two-step regression and 0 . 50 for the one-step regression analysis ( when looking only at areas with a high predicted probability of occurrence ) . Both models tended to predict higher prevalence than was observed , particularly in villages of zero prevalence , with a median error of 0 . 05% for both models . The mean absolute error was equal for the two methods ( 0 . 24% ) . The difference in predicted prevalence of HAT from the two methods was small over the majority of the prediction area , with divergences mainly occurring in areas of high predicted prevalence outside of the study area ( see Figure 9 ) . There were only two health centres trained and equipped to diagnose and treat HAT serving the study population during the study period . It has been shown previously that levels of geographical accessibility to treatment facilities can have an effect on the observed spatial distribution of HAT , with smaller numbers of cases reported from areas which are further from the treatment centres [13] . However , an added complication arises in this study as the choice of site for the main HAT treatment facility in the area ( at Lwala Hospital ) was driven in part by its location within the focus of new cases of HAT in 2004 . Moreover , this facility is close to one of the major livestock markets in the study area ( 7 . 5 km away ) making their separate influences on observed prevalence difficult to distinguish . Distance to the closest livestock market was an important predictor in the one-step regression and in both steps of the two-step regression , with decreasing odds of infection at increasing distances . Previous research has confirmed the introduction of HAT to a previously unaffected area via the introduction of untreated , infected livestock [6] . These results suggest that despite reinforced policy regarding the treatment of livestock for trypanosomes prior to movement from endemic areas [38] , the ongoing spread of HAT into Kaberamaido and Dokolo may have been facilitated by the movement of infected cattle through one or more of the local livestock markets . The main cattle trading routes within this part of Uganda run from T . b . rhodesiense endemic areas in the south east , through the study area and neighbouring districts , to the T . b . gambiense endemic areas in the far north west of Uganda towards southern Sudan . Clearly , this increases the risk of overlap of the two subspecies , particularly if the regulations regarding the treatment of cattle being moved from T . b . rhodesiense endemic areas continue to be broken . The stringent implementation of regulations requiring the treatment of cattle prior to sale at livestock markets should be a priority for the Ugandan Government and tsetse control efforts may be more efficiently targeted to areas surrounding livestock markets to prevent the establishment of transmission in previously unaffected areas as occurred in Soroti district in the late 1990s and Kaberamaido and Dokolo districts in 2004 . Other variables that were also significantly correlated with HAT prevalence and/or occurrence included distance to the nearest health centre , maximum Normalised Difference Vegetation Index ( NDVI ) , NDVI phase of annual variation , NDVI annual amplitude , minimum Land Surface Temperature ( LST ) , LST phase of annual variation , LST annual amplitude , mean LST , distance to the closest area of woodland and distance to the closest area of bush . The significance of these variables highlights the importance of climatic and environmental conditions for HAT transmission . Distance to the closest health centre was also a significant factor in each model , with decreasing prevalence observed at increasing distances . This suggests a confounding relationship due to accessibility of health services as has been previously reported [13] . Each of the regression models ( the one-step regression model and each step of the two-step regression models ) included maximum NDVI ( negative association ) and minimum LST ( positive association ) as significant predictors . These are likely to relate to the habitat and environmental requirements of the tsetse fly vector of disease . The additional variables found to be significantly correlated with HAT prevalence in each analysis are probably also linked to the suitability of an area for the tsetse fly vector ( due to their preferred habitat and also climatic requirements ) , and so will influence the intensity of transmission and observed prevalence of HAT . Analysis of the residual variation ( after accounting for the covariate effects ) indicated that there was some spatial autocorrelation in the residuals from the one-step regression and the probability of occurrence analysis ( first step of the two-step regression analysis ) . For the two-step regression , the probability of occurrence regression was carried out partially to provide a mask over areas with low predicted probability of occurrence to enable the focusing of the prevalence analysis , and so the small amount of spatial autocorrelation in the residuals is not seen as problematic as it would have a negligible effect on the final prevalence model . However , for the one-step regression , the small amount of spatial autocorrelation in the residuals may lead to inflated statistical significance for some of the covariates . Further research is underway to address this autocorrelation in the residuals and to assess any increase in the predictive accuracy using a model-based geostatistics approach [39] . From these and previous findings [6] , it is thought to be likely that the movement of T . b . rhodesiense infected livestock from endemic areas through livestock markets within the study area occurs periodically . A complex interaction of factors is involved in the establishment of transmission following such an occurrence . In addition to the variables included in the current analysis , tsetse and livestock densities , human-cattle-tsetse contact and also to a large degree , chance , may play roles . Further research is planned to build upon these findings , incorporating detailed livestock market data and cattle trading networks to give a more thorough understanding of the spatial and temporal dynamics of HAT within Uganda .
|
Human African Trypanosomiasis ( HAT ) or sleeping sickness is a parasitic disease of humans , transmitted by the tsetse fly . There are two different forms of HAT: Rhodesian ( in eastern sub-Saharan Africa ) , which also affects wild and domestic animals , and Gambian ( in western and central sub-Saharan Africa ) . Diagnosis and treatment of the two diseases differ , and disease characterisation is based on prior knowledge of known geographical disease distributions . Presently , the two forms of HAT do not overlap in any area: Uganda is the only country which sustains active transmission of both types . In recent years , Rhodesian HAT has spread into areas of Uganda that had not previously been affected , thus narrowing the gap between areas of Rhodesian and Gambian HAT transmission . This spread has raised concerns of a potential overlap of the two types of the disease , which would severely complicate their diagnosis and treatment . Earlier work indicated that Rhodesian HAT was introduced to Soroti district due to the movement of untreated cattle from affected areas . Here we show that the continued spread of HAT in Uganda ( to a further 2 districts ) may also have occurred due to cattle movements , despite legal requirements to treat livestock from affected areas prior to sale at markets . These findings can assist in the targeting of HAT control efforts in Uganda and show that the stringent implementation of animal treatments at livestock markets should be a priority .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/protozoal",
"infections",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"infectious",
"diseases/epidemiology",
"and",
"control",
"of",
"infectious",
"diseases"
] |
2009
|
Spatial Predictions of Rhodesian Human African Trypanosomiasis (Sleeping Sickness) Prevalence in Kaberamaido and Dokolo, Two Newly Affected Districts of Uganda
|
Linkage analysis is useful in investigating disease transmission dynamics and the effect of interventions on them , but estimates of probabilities of linkage between infected people from observed data can be biased downward when missingness is informative . We investigate variation in the rates at which subjects' viral genotypes link across groups defined by viral load ( low/high ) and antiretroviral treatment ( ART ) status using blood samples from household surveys in the Northeast sector of Mochudi , Botswana . The probability of obtaining a sequence from a sample varies with viral load; samples with low viral load are harder to amplify . Pairwise genetic distances were estimated from aligned nucleotide sequences of HIV-1C env gp120 . It is first shown that the probability that randomly selected sequences are linked can be estimated consistently from observed data . This is then used to develop estimates of the probability that a sequence from one group links to at least one sequence from another group under the assumption of independence across pairs . Furthermore , a resampling approach is developed that accounts for the presence of correlation across pairs , with diagnostics for assessing the reliability of the method . Sequences were obtained for 65% of subjects with high viral load ( HVL , n = 117 ) , 54% of subjects with low viral load but not on ART ( LVL , n = 180 ) , and 45% of subjects on ART ( ART , n = 126 ) . The probability of linkage between two individuals is highest if both have HVL , and lowest if one has LVL and the other has LVL or is on ART . Linkage across groups is high for HVL and lower for LVL and ART . Adjustment for missing data increases the group-wise linkage rates by 40–100% , and changes the relative rates between groups . Bias in inferences regarding HIV viral linkage that arise from differential ability to genotype samples can be reduced by appropriate methods for accommodating missing data .
Interest has been growing in the use of viral linkage analysis to investigate disease transmission dynamics and the effect of interventions on them [1]–[8] . To optimize interventions intended to control the HIV epidemic , it will be useful to identify host characteristics ( e . g . disease status and demographics ) that are associated with high rates of clustered or genetically-linked infections . Many studies attempt to make inferences about linkage patterns in a larger population than that represented by the set of observed viral genetic sequences without considering the effect of sampling or missing data ( see , e . g . [5] , [9] ) . However , estimates of probabilities of linkage that ignore the impact of missing data ( henceforth referred to as unadjusted estimators ) can be biased downward . In order to estimate the amount of linkage in communities or compare rates of linkage across groups we must properly account for the presence of missing data . The work presented here arose from a desire to compare linkage rates between demographic groups found via a household survey from the Mochudi study , an HIV prevention program for Mochudi , Botswana ( R01 AI083036; PI: M . Essex; www . aids . harvard . edu/news/spotlight/archives/v6i3_mochudi_project . html ) . Young males were found to be severely underrepresented , making inferences about linkage involving this group unreliable . As information regarding the size of this subpopulation is available , it is possible to leverage it to improve inferences . This household survey is part of a pilot project leading to a large community-randomized trial , also in Botswana , of a combination HIV prevention intervention , the Botswana Combination Prevention Project ( BCPP; U01 GH000447; PIs: M . Essex & V . De Gruttola ) [10] , [11] . One of the goals of the BCPP study is to leverage viral linkage to understand the patterns of mixing across communities and the relative contributions of within-community and outside-community sources to new infections . This paper develops estimators for linkage probabilities under the assumption that unobserved sequences are missing at random conditional on observed information . We consider analyses in which linkage is defined by a threshold on the pairwise distance between viral sequences . The choice of the threshold is an important scientific question in the analysis of viral genetic data , but the methods developed here apply regardless of the particular value of the threshold chosen , or can be applied to range of thresholds of interest . We first show that the probability that randomly selected sequences are linked can be estimated unbiasedly from observed data . We then derive an estimate of conditional probabilities of linkage between groups given the existence of a link , and consider estimation of group-level probabilities of linkage . We first develop estimators under the assumption that indicators of linkage are independent across pairs of individuals who may be linked – an assumption that could be appropriate in situations with either a very sparse graph or sparse sampling in the population . We then develop a bootstrap resampling approach that is approximately correct under general assumptions about the structure of correlations of linkage indicators across pairs . Finally , we propose a diagnostic approach for assessing the reliability of the method . We apply the methods developed to analyses of viral sequences from the northeast sector of the village of Mochudi in Botswana , the site of a pilot study intended to determine the feasibility of testing for HIV infection in a household setting and linking infected subjects to care . Our investigation focuses on assessment of whether rates at which subjects' HIV genotypes link with others depends on ART treatment status and viral load levels ( low/high ) among the untreated . Such clustering reflects underlying HIV transmission dynamics; a tendency for subjects with high viral load to link more frequently with others might suggest an increased role of subjects with elevated levels of viral replication in HIV transmission . This is also consistent with high viremia in early infection; the contribution of those with elevated viral load to onward spread is difficult to assess in samples of prevalent cases due to the fact that a subject's category varies over time . With high prevalence , however , it is unlikely that a high proportion of subjects in the sample are newly infected; nonetheless , this approach will be particularly useful in the analysis of data from the BCPP , which will identify incident cases and permit comparison of their linkage rates with the groups discussed here .
Our first goal is to estimate the probability of linkage between viral sequences from two individuals selected at random from their respective groups , . Under the assumption of missingness at random ( MAR ) conditional on group membership , the hosts for whom viral sequences are available represent a random sample of the total population of their group , and by extension the observed linkage indicators are a random sample of the linkage indicators for the full population . Thus , the Law of Large Numbers tells us that the sample average , , converges to the population mean , . As a result , under the assumption of MAR conditional on group , it is possible to obtain an unbiased estimate of the probability that a pair of sequences are linked without adjustment for missing data . One quantity of interest in the analysis of a community randomized trial such as the BCPP is the relative probability that a new infection arises from contact with an infected person from within a community versus from outside the community . Therefore we may wish to estimate the conditional probability , , that a pair of sequences are from groups , given that are linked . If missingness is completely at random ( unconditional on group ) , then we can use the observed proportions of links in each group pair , , with , to estimate the conditional probabilities . If missingness is MAR conditional on group , as we assume , this estimate requires adjustment for the differing missingness rates between groups . In a population of size , there are total possible pairs . The probability of linkage for a randomly selected pair is given byThe probability that a randomly selected pair is from groups and is linked isThus , the conditional probability we desire isWe substitute into these formulas to obtain a plug-in estimator of . Note that this derivation does not require an assumption of independence , so we can consistently estimate the conditional probability of linkage regardless of the underlying correlations of linkage indicators across pairs . Based on the results above , we now focus on estimation of the probability that a randomly selected sequence from group links with at least one sequence from group ( excluding itself if ) , . In this case the unadjusted estimate of the probability of linkage between groups will be an underestimate of the true rate: any sequence that does not link with any other in the observed data may in fact link with sequence ( s ) from the community that were not observed . Thus , the proportion of observed sequences in group that do not link with any sequence in group will be higher than the proportion in the population . For the purposes of exposition , we begin with an assumption of independence among linkage estimators , but we extend to a case with individual-by-group random effects driving the correlations among indicators . This flexible model accounts for correlations due to individual factors – biology , behavior , network position – as well as differential interactions of individuals with different groups .
As a first step in validating the performance of the approach , we perform a simulation study applying our methods to data simulated from an evolutionary model . To implement the simulation , we used SeqGen v1 . 3 . 2 [17] . We obtained the tree required as input to the program by fitting a maximum likelihood tree to the 423 observed sequences from Mochudi , and parameterized the evolutionary model by fitting the general time-reversible model with gamma distributed rate heterogeneity to those sequences and using the estimated parameters ( both using MEGA version 5 [18] ) . Each node maintained the group assignment it had in the Mochudi data . The simulation proceeds as follows: The threshold ranged from 0 . 17 to 0 . 24 , which corresponds roughly to the 0 . 04th to 0 . 41st percentiles of the distance distribution . The expected number of links per sequence ranged from 0 . 04 to 0 . 79 . We simulated 100 sets of sequences , and for each set , we simulated 100 different observed data sets for each threshold , for a total of 10 , 000 simulations per threshold . Figure 1 plots the mean relative bias ( |estimate–truth|/truth ) of the unadjusted and adjusted estimators across the range of thresholds . The unadjusted estimator has uniformly higher bias than the adjusted , and the differences in the degree of bias is often large; averaged across subpopulations ( weighting by their size ) and thresholds , the relative bias of 25 . 7% in the unadjusted analyses is reduced to 6 . 5% in adjusted analyses . For higher thresholds , the adjustment reduces the bias to under 5% in the majority of cases and to under 10% in all . For the lower thresholds ( where linkage rates are lower ) , the bias in unadjusted analyses is generally greater than for higher thresholds-exceeding 35% in some cases . By contrast the bias in the adjusted analyses is below 10% in the majority of cases and below 20% in all but one . The worst performance for the adjusted analyses ( low thresholds for LVL to LVL ) still shows a considerable reduction in bias . In the analysis of the Mochudi household survey data , we consider three groups: HVL , LVL and ART . We observe sequences for out of individuals in each group , yielding . We use p-distance as our distance measure: the proportion of compared sites at which two sequences differ . Viral linkage in this analysis is defined by a p-distance below a specified value . We present the results in two ways: first , using a range of thresholds from 0 . 085 to 0 . 12 ( corresponding to the 0 . 03rd to 0 . 54th percentiles ) , and second , focusing on a threshold of 0 . 1 for more detailed consideration . This latter threshold yields an overall rate of linkage of 18% within the observed sample . Using the results for the probability of linkage between individuals , we find the given in Figure 2 and Table 1 . As one would expect , the overall probability of linkage increases with the more generous thresholds , but the pattern of relative probabilities appears to be maintained . In the table , we can see more clearly that linkage is most likely with the HVL group for all groups , while the LVL group demonstrates less linkage overall . We now move to estimation of the conditional probability that a linked pair are from groups , . First , we examine the performance of the estimator via simulation from real data . Treating the 423 observed sequences in the Mochudi data as a full population , we sample with probability ( 0 . 7 , 0 . 6 , 0 . 8 ) from the ( HVL , LVL , ART ) groups . We can then record the true conditional probabilities from the full data and the unadjusted and adjusted estimates from the sampled data . Figure 3 gives the distribution of estimates of the conditional probabilities , compared against the probabilities observed in the full sample . The MLE is quite accurate , as we would expect given the generality of the results in Methods for conditional probabilities of linkage . The adjusted estimates of the conditional probabilities for the full sample are given in Figure 4 and Table 2 . The relative probabilities vary more with the threshold in this case than in the individual-to-individual case , likely because the probabilities of linkage are extremely small ( particularly when involving the ART group ) and thus minor differences in the distribution of distances by group pair could lead to widely varying conditional probability estimates . It does appear to be most likely that a given link occurs between HVL and ART or LVL , and it is least likely to be between two LVL individuals . Before we proceed to estimate group-wise linkage rates for the Mochudi data , it is useful to examine the estimated correlation under the exchangeable model , which we will consider in development of a diagnostic tool for assessing the reliability of our methods . For the Mochudi data , we obtain a population-wide estimate of ; group-specific estimates are given in Table 3 . Most are close to the population-level estimate , but there is some variability . We can also see how the realized values of change with the sampling fraction . Figure 5 show boxplots of these realized values for subsamples of 5 to 95% of the Mochudi data . Each boxplot represents 500 samples . The estimates become increasingly variable as the sample gets smaller , but remain centered about the value of from the full data ( red line ) until the sample size falls below 40% , at which point the estimates decline sharply . This is likely due to an increased probability of obtaining a sample with very few observed links between the two groups . In the extreme case when no links are observed , this yields and , and we can expect the estimated correlation to be extremely small in cases with only a handful of links as well . This decline in the estimated correlation , , for very small samples has implications for bootstrap bias correction . We propose as a diagnostic creation of a plot similar to Figure 5 from the observed data by group . If the estimated median of appears to remain fairly constant over a range of sampling fractions including the size of the appropriate subsample , the estimated is likely to be similar to the true value , and the bootstrap bias correction should work well , since the estimated adjustment ratio described in the Methods depends upon being similar across the population , observed sample , and bootstrap subsample . We first assess the performance of the bootstrap estimator via simulation . Figures 6 and 7 show the resulting estimates for 70% and 30% samples of the observed data , respectively , including one modification: if the adjustment reduces the unadjusted estimate , we take the unadjusted value rather than the bias-adjusted value . This rule follows from knowing that the unadjusted estimate is an underestimate , implying that any reductions are likely due to very small bootstrap sample sizes or disparate correlations in the observed data and the subsample . This restriction has no effect on the 70% sample , but does impact the 30% sample substantially; we see that the bias-corrected results are not very different from the unadjusted estimates . We can now compare the estimates of for each pair of groups using the unadjusted estimator based on observed data and using the bootstrap adjustment method . The estimated probabilities of inclusion are 65 , 54 , and 45% for HVL , LVL and ART , respectively . Based on the decline of in Figure 8 , we would expect that the observed correlation in the subsamples is likely to be different from the sample correlation for the LVL group , despite the theoretical possibility of obtaining the 8% subsample needed for interval subsampling . Therefore , we use interval subsampling for the HVL group , but use proportionate subsampling for the LVL and ART groups . We calculate a confidence interval for the bootstrap adjusted estimate using a bootstrap quantile interval . Because the adjustment is made by taking the inverse of the bootstrap samples , the upper ( lower ) bound of the interval will be given by taking the ( ) quantile of the bootstrap distribution of the ratio of the unadjusted estimate to the bootstrapped value ( raised to the power of if using proportionate subsampling ) and calculating with this quantile . Simulation results show that the coverage of this interval is likely to be good as long as the sampling percentage is at least 65% and may be anti-conservative if the percentage is lower . Intervals for the unadjusted estimator are found using a traditional binomial interval . We first present the results of applying the adjustment across a range of possible thresholds . As can be seen in Figure 9 , the adjusted estimates are consistently higher than the unadjusted , regardless of threshold . In many cases , particularly for higher thresholds , the confidence interval for the unadjusted estimator excludes the point estimate from the bootstrap adjustment . The bootstrap quantile interval is consistently narrower than the unadjusted interval in the cases where interval subsampling was used ( column 1 ) . To see the effects of adjustment in more detail , we focus on a single cutoff of 0 . 1 in Table 4 , where we see both estimates for the probability that a member of group A ( rows ) is linked with at least one member of group B ( columns ) . The adjusted estimates range from 40 to 100% larger than those of the unadjusted estimator . The relative values of the probabilities change as a result - for example , using the unadjusted estimator , it appears that someone with high viral load is nearly twice as likely to cluster within the high viral load group as to cluster with anyone on ART . After adjustment , an HVL individual is only half again as likely to cluster within group as with ART . In this case , qualitative comparisons - specifically , the ranking of the prevalence of various combinations - remain unchanged , although it is possible in other applications that this would not be the case . Unlike the sequence-to-sequence or conditional probabilities of linkage , the probability of linkage from group to group is not equal to the probability of linkage from group to group . This is a function both of the sizes of the groups ( if group is much bigger , there are more chances for someone in group to have a link with than the other way around ) and of the distribution of links . Consider , for example HVL/LVL linkage; it is much more likely for someone with high viral load to link with the low viral load group than for someone with LVL to link with the HVL group . This could arise simply due to the size of the LVL group , but it is also possible that several individuals in the LVL group link with multiple individuals in the HVL group . For each such configuration , only one person in LVL is counted as having link ( s ) to HVL , but multiple people in HVL are counted as having link ( s ) to LVL . The results for group-wise linkage rates suggest that individuals with high viral load have more links; by extension , this suggests they are involved in more transmissions or more recent transmissions . This group should include both individuals in the chronic phase of infection with poor viral suppression and recent infections [21] . Of 75 HVL with a CD4 count available , 44% had CD4 count below 250 cells/mm3 , while only 19% of the 122 LVL individuals with a CD4 count measurement were below 250 . Individuals on ART are most likely to link within their own group , suggesting links from older transmissions in which both individuals have progressed to the point of needing treatment . Individuals in the low viral load group link relatively little . Without information on the relative timing of infections ( such as will be available from the BCPP ) , we cannot make inference about transmission contributions . However , taken together , these results are at least consistent with the hypothesis that those with low viral load , either from natural suppression or treatment , are not transmitting as efficiently as those with high viral load . Data on prevalence and incident cases collected over time will permit more formal testing of this hypothesis . Many different methods are available for calculating distances between genetic sequences [22] . As the definition of linkage is a simple threshold on the distance between sequences , distance models that give different results could result in rate estimates that vary widely . We compared four different distance calculation methods on the set of 423 sequences from the first year household survey in Mochudi , Botswana . The methods compared were: Codon positions included were 1st+2nd+3rd+Noncoding . All positions with less than 95% site coverage were eliminated . That is , fewer than 5% alignment gaps , missing data , and ambiguous bases were allowed at any position . There were a total of 1050 positions in the final dataset . Analyses of uncorrected pairwise distances and corrected by different evolutionary models were conducted in MEGA version 5 [18] . The scale or mean of the distance distribution might be expected to vary over the methods used to calculate the distances . To reduce this source of variability , we treat the threshold for linkage as a quantile of the distribution ( i . e . , the bottom 10% of distances cluster ) , thereby ensuring that measures that maintain the same ranking of distances provide equivalent results . Table 5 gives the Spearman rank correlation matrix of the five methods listed above . As the Spearman correlation considers only ranks , any measures that are equivalent up to a monotonic increasing transformation will have a correlation of 1 . All four methods used have nearly perfect correlation , indicating that applying the analysis methods described here with a quantile-based cutoff will result in nearly identical results regardless of the distance model used .
Genetic linkage analyses have been useful in making inferences about important HIV epidemic drivers , including the impact of acutely or recently infected subjects [1] , [2] , [4] . Application of these methods to community randomized trials of HIV prevention interventions such as the cluster randomized trial of HIV prevention in Botswana [10] , [23] may be useful not only for this purpose but also to provide information regarding the subpopulations in which these interventions are succeeding or failing . For example if newly infected subjects in communities randomized to the intervention cluster only with viruses that infect people living outside the community , this knowledge would imply that the intervention is succeeding in stopping transmission within communities . The implications for the success of the intervention are very different from the setting in which newly infected cases are in fact being infected with viruses circulating within communities . For the latter , it is important to know what subgroups contribute most to onward transmission of virus , whether these subgroups be defined by plasma HIV levels , ART treatment status , or demographic or behavioral factors . All such analyses , however , are very much impacted by potentially informatively missing data . This paper proposes methods to adjust for such biases . Our methods adjust for the presence of missing viral sequences in estimates of viral linkage rates under the assumption that sequences are missing at random conditional on group membership . We show that we can consistently estimate the probability that two sequences are linked without adjustment for missing data , and can consistently estimate conditional probabilities of linkage between two sequences from a pair of groups given the existence of a link via a minor adjustment using the ( known ) sizes of the groups in the population . In settings where it is reasonable to assume that the linkage status of pairs of sequences are all independent conditional on group , the estimator presented for estimation of group-wise linkage probabilities under independence is in fact the MLE and provides an exact solution . This assumption might be reasonable in investigations of airborne pathogens , or in settings with sparse sampling . For settings in which the assumption of independence is not reasonable , we propose a bootstrap resampling approach to adjust for the bias in the unadjusted estimator . If linkage indicators are exchangeably correlated or if their correlations can reasonably be modeled as functions only of individual effects ( a random-effects type model ) and we can use interval subsampling , then the resampling method can adequately adjust for bias . When using proportionate subsampling under the random-effects model the bootstrap may under-correct , but the resulting estimates are still preferable to those provided by unadjusted estimators . We note that departures from the assumption of a random effects structure in the correlation would arise if the probability of linkage depended not only on the individual characteristics of sequences and the people infected with them , but also , in unspecified ways , on the interactions between these characteristics . In such cases , unbiased adjustment for missing data is not possible , because such departures would imply that unobserved linkages followed a different process from those that are observed . Even in this case , however , it would be useful to employ our methods , because they at least provide estimates that are valid under much broader assumptions than in the case for unadjusted analyses and they demonstrate the effect of the broadening of assumptions on results . Large changes in estimates provide caution against overinterpretation of results . Furthermore , our simulation results using the Mochudi data suggest that the adjustment may be adequate in some realistic settings where the assumption of the random effects structure may not hold perfectly . To provide guidance on appropriate usage of the method , we propose a diagnostic tool that provides assessment of the likely reliability of the bootstrap resampling approach to adjust estimates of clustering rates . The choice of the threshold defining linkage will vary broadly with the goal of analysis and methods of data collection . This choice is critical to any linkage analysis , and sensitivity to the choice of threshold should be examined . The methods developed here can be applied to any threshold or range of thresholds in order to obtain linkage rate estimates that are adjusted for the presence of missing data . Considering adjusted results for a range of thresholds will permit more reliable comparisons between groups and between thresholds . Although the groups of interest for linkage and those of relevance for the missingness model were the same in our example , this condition is not required . A more general missingness model could be formed by creating a partition into subgroups such that pairs of observations are missing at random given subgroup membership . Our method would then proceed by first estimating linkage rates for each of these subgroups , and then aggregating across them to obtain the estimates for the groups of interest ( as suggested in Methods ) . As an example , to address our fundamental goal of estimating the relative contributions of within-community and outside-community partners to new infections , we would include community as one of the variables that defines our groups . We might , for example , define groups as community by sex by age category , for example . Given age- and sex-specific prevalence estimates for each community , we can adjust for missing data within these categories , and then aggregate to the level of community , yielding estimates of the proportion of individuals in community 1 who cluster with community 2 and vice versa , as well as the proportion who cluster within their own communities . Such an analysis will provide an indication of the relative force of infection from within versus outside the community , especially if we have separate groups for incident infections . The methods can also be extended to allow the model for missingness to depend on continuous-valued variables . The approach discussed here is not restricted to linkage indicators defined by a pairwise distance cutoff . The rate of occurrence of any feature of interest that can be coded as an indicator variable for each pair of sequences can also be estimated with adjustment for missing data . Beyond the change in the definition of a link , the application of the method is identical . The bootstrap method described here is similar in spirit to inverse-probability weighting in that adjustment for bias makes use of information on the probability of observation to estimate a scaling factor . In our setting , however , it is not possible to express the weight in closed form because of the complex correlation structure induced by the vagaries of HIV evolution and of patterns of viral transmission .
|
The analysis of viral genomes has great potential for investigating transmission of disease , including the identification of risk factors and transmission clusters , and can thereby aid in targeting interventions . To make use of genetic data in this way , it is necessary to make inferences about population-level patterns of viral linkage . As with any rigorous statistical inference from sampled data to a population , it is important to consider the effect of the sampling strategy and the occurrence of missing data on the final inferences made . In this paper we highlight the effects of missing data on the resulting estimates of population level linkage rates and develop methods for adjusting for the presence of missing data . As an example , we consider comparing the rates of linkage of HIV sequences from subjects with high viral load , low viral load , or on antiretroviral treatment , and show that comparative inferences are compromised when adjustment is not made for missing sequences and bias in inferences can be reduced with proper adjustment .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"sequence",
"analysis",
"epidemiological",
"methods",
"mathematics",
"infectious",
"disease",
"epidemiology",
"epidemiology",
"statistics",
"genetics",
"population",
"biology",
"biology",
"computational",
"biology",
"biostatistics",
"genetics",
"of",
"disease",
"statistical",
"methods"
] |
2014
|
Linkage of Viral Sequences among HIV-Infected Village Residents in Botswana: Estimation of Linkage Rates in the Presence of Missing Data
|
ESCRT-III proteins catalyze membrane fission during multi vesicular body biogenesis , budding of some enveloped viruses and cell division . We suggest and analyze a novel mechanism of membrane fission by the mammalian ESCRT-III subunits CHMP2 and CHMP3 . We propose that the CHMP2-CHMP3 complexes self-assemble into hemi-spherical dome-like structures within the necks of the initial membrane buds generated by CHMP4 filaments . The dome formation is accompanied by the membrane attachment to the dome surface , which drives narrowing of the membrane neck and accumulation of the elastic stresses leading , ultimately , to the neck fission . Based on the bending elastic model of lipid bilayers , we determine the degree of the membrane attachment to the dome enabling the neck fission and compute the required values of the protein-membrane binding energy . We estimate the feasible values of this energy and predict a high efficiency for the CHMP2-CHMP3 complexes in mediating membrane fission . We support the computational model by electron tomography imaging of CHMP2-CHMP3 assemblies in vitro . We predict a high efficiency for the CHMP2-CHMP3 complexes in mediating membrane fission .
Membrane fission leading to division of one continuous membrane into two separate ones is ubiquitous in cell physiology . It is one of the crucial events in generation of transport intermediates from plasma membranes and intracellular organelles; steady-state dynamics of the endoplasmic reticulum , mitochondria and Golgi complex; virus budding , cytokinesis and other fundamental phenomena ( see for review e . g . [1]–[3] ) . In the process of fission , a membrane changes its shape and undergoes a topological transformation which includes transient perturbations of the membrane continuity . To overcome the membrane resistance to shaping and remodeling , a substantial energy has to be invested into the system , which requires action of specialized proteins ( see for review [2] , [3] ) . Identification of proteins which shape and remodel membranes in the course of diverse intracellular processes has become a hot topic of cell biology [1] , [3] , [4] . The major advance has been achieved in discovering proteins generating and/or sensing the membrane curvature . The list of such proteins is constantly expanding and the mechanisms of their action are being elaborated [1] , [4] , [5] . Less progress has been made in understanding how proteins drive the membrane fission per se . While several protein types such as the dynamin-family proteins ( see e . g . [6]–[10] ) , CtBP1/BARS [11] and PKD [12] have been implicated in fission of cell membranes , until recently , the ability to split membranes was unambiguously demonstrated for , perhaps , only one protein , dynamin-1 [9] , [13]–[15] . Whereas different versions of the mechanism of membrane fission by dynamin-1 were suggested ( see for review [10] ) , the idea unifying the majority of these proposals is that dynamin self-assembles on the membrane surface into helical oligomers constricting the membrane underneath into thin tubes . Strong mechanical stresses induced by dynamin in the tubulated membrane upon GTP hydrolysis can relax as a result of membrane division and , therefore , drive membrane fission . Accumulating evidence suggests that the ESCRT ( Endosmal Sorting Complexes Required for Transport ) complexes [16] – are able to catalyze the membrane budding and fission processes . The ESCRT machinery consists of five different complexes - theVps27complex ( ESCRT-0 ) , ESCRT-I , -II , and -III , and the Vps4 complex - whose coordinated action sorts trans-membrane proteins into intralumenal vesicles ( ILV ) , which bud off from the limiting membranes of endosomes and transform endosomes into multivesicular bodies ( MVB ) [16]–[19] . In addition to the MVB generation , the combined action of ESCRT-III and VPS4 complexes are required for the budding of some enveloped viruses including HIV-1 [20]and during late steps in cytokinesis [21]–[24] . It is thus most likely that ESCRT-III and VPS4 catalyze membrane fission reactions , common to all three biological processes [21]–[25] . The ESCRT-III complex in yeast consists of four core subunits Vps20 , Snf7 , Vps24 , and Vps2 [26] whose mammalian analogues are the charged multivesicular body proteins CHMP6 , CHMP4 , CHMP3 and CHMP2 , respectively . The subunits are consecutively recruited to the membrane in the order of Vps20/CHMP6 , Snf7/CHMP4 , Vps24/CHMP3 and Vps2/CHMP2 [27]–[29] and their assembly into higher order complexes was suggested to drive the inward membrane budding in vitro [28] . Moreover , these four proteins are able to act as minimal budding machinery as was confirmed by demonstration that their sequential addition to giant unilamellar vesicles ( GUV ) generated membrane invagination and abscission of the inward vesicles [29] . Specifically , formation of membrane buds connected by open necks to the initial membrane was shown to depend , critically , on the Snf7 ( CHMP4 ) and Vps20 ( CHMP6 ) subunits , while the neck fission proved to require the Vps24 ( CHMP3 ) subunits [29] . Three different albeit similar models for ESCRT-III catalyzed budding have been suggested [30] . First , Snf7 ( CHMP4 ) circular filaments or flat spirals lying in the membrane plane [31] start at the center of a newly formed membrane bud and catalyze membrane bending as the bud grows [31] . A second model suggests that a circular ESCRT-III filament with asymmetric ends delineates a membrane patch containing cargo molecules and constricts the neck of an evolving membrane bud via the disassembly action of Vps4 [27] . A third model , similar to the second one , proposes that an ESCRT-III spiral surrounds and constricts a cargo containing membrane domain leading to membrane budding and fission [29] . However , spiral polymers of ESCRT-III have only been observed for hSnf7 ( CHMP4 ) in vivo [31] and in vitro [32] , whereas the detachment of the forming vesicle including fission of a membrane neck was shown to be crucially dependent on Vps24 ( CHMP3 ) [29] . Therefore , in addition to the Snf7 ( CHMP4 ) filaments , the structures formed by self-assembly of Vps24 ( CHMP3 ) must play an indispensable role in the ESCRT-III mediated membrane budding and fission . CHMP3 ( Vps24 ) and CHMP2A ( Vps2 ) form heterodimers [26] , [33] that assemble into tubular nano-structures which display a variety of end-cap shapes including nearly hemispherical dome-like end-caps ( [34] and the section “Experimental support for the model” below ) . The external and internal radii of these structures are approximately 52 and 43nm , respectively [34] . In vitro , the AAA ATPase VPS4 binds to the inside of the CHMP2-CHMP3 polymers and leads to their disassembly in the presence of ATP [34] . The external surface of a CHMP2-CHMP3 nano-structure has a considerable affinity to membranes containing acidic lipids [34] . Therefore , in the process of self-assembly , the CHMP2-CHMP3 complex must be able to attract a lipid bilayer , hence , scaffolding the bilayer into a strongly curved shape , a process that might drive membrane fission reactions [34] . In spite of the apparent similarities between the dynamin-I and CHMP2-CHMP3 assemblies such as ( i ) the ability to scaffold membranes into cylindrical shapes , and ( ii ) the energy input by nucleotide hydrolysis , CHMPs cannot employ any of the mechanisms of membrane fission suggested for the dynamin action . Indeed , topologically , the fission reactions mediated by dynamin and ESCRT-III are directed differently: dynamin and its partners drive membrane budding and abscission towards the cytosol , while ESCRT-III mediates membrane abscission away from the cytosol and towards the lumen of an endosome . Structurally , a membrane portion tubulated by a dynamin oligomer is situated within the protein scaffold and , hence , could undergo further thinning upon detachment from dynamin and divide by self-fusion within the protein framework [14] . In contrast , the membrane wrapped around a CHMP2-CHMP3 structure is attached to the outside surface of the protein scaffold and , hence , the scaffold hinders the membrane sterically from direct thinning and self-fusion . Thus , the character of membrane deformation leading to fission driven by CHMP2-CHMP3 structure must differ essentially from that generated by dynamin and the mechanics of the fission reaction must be dissimilar in the two cases . Here , we suggest and integrate the current structural knowledge on ESCRT-III complexes to elaborate on a novel mechanism of membrane fission by dome-like assemblies formed by the CHMP2-CHMP3 subunits of ESCRT-III . The essence of our proposal is that , in contrast to the fission mechanisms suggested for the dynamin action ( see for review [10] ) , the site of membrane fission driven by ESCRT-III is not co-localized with the protein scaffold but rather emerges aside of it within a membrane neck which forms in the course of membrane wrapping around the ESCRT-III dome . The major energy for the fission reaction comes from the energy of membrane attachment to the surface of the ESCRT-III complex . We discuss a possibility for a reinforcement of the ESCRT-III based mechanism by the Vps4 binding . Our calculations predict that ESCRT-III domes can serve as effective mediators of membrane fission resulting in generation of vesicles of biologically relevant dimensions .
We consider a hemi-spherical protein dome of radius serving as a scaffold for attachment of a membrane fragment of a total area ( Fig . 2a ) . While , in reality , the membrane attachment to the dome proceeds concomitantly with the dome assembly , for the calculation purposes we will regard the dome to be completed . This is based on a plausible assumption that the attractive interaction between the subunits of the CHMP2-CHMP3 structure must be much stronger than all other relevant interactions characterizing the system . Therefore , the protein self-assembly proceeds irrespectively of the membrane attachment , while the latter follows the dome building and its extent is determined by the interplay between the membrane bending energy and the membrane affinity to the protein surface . The absolute value of the energy of the membrane interaction with the dome surface per unit area of the membrane-protein interface will be referred to as the membrane affinity and denoted by . Since the membrane-protein interaction is attractive its energy is negative and its value per unit area is . Note that , according to our definition , the affinity accounts only for the direct ( probably , electrostatic ) interaction between the protein and the lipid polar groups and does not include the energy of membrane bending , which accompanies the membrane binding to the protein dome and contributes to the total energy of this process . Therefore , the value of is not supposed to depend on curvature of the protein surface . In this respect , the notion of the affinity we are using differs from the total energy of the membrane attachment to the protein complex , which includes the bending contribution and is commonly used to characterize interaction of proteins with bent membranes ( see e . g . [1] , [4] , [39] , [40] ) . In our approach the curvature effects are considered separately from the direct membrane-protein interaction . The membrane adopts a curved shape of a bud characterized at each point by the total curvature and the Gaussian curvature [41] . The radius of the narrowest cross-section of the bud neck will be referred to as the neck radius , ( Fig . 2a ) . The membrane bending energy per unit area of the membrane mid plane , , is given by [42] , [43] , ( 1 ) where is the bilayer bending modulus ( see e . g . [44] ) , and is the bilayer modulus of Gaussian curvature whose values were not directly measured but estimated to be negative ( see e . g . [45] , [46] ) . We analyze two alternative states of the system: the fore-fission state where the membrane bud is connected by a membrane neck to the membrane portion attached to the protein dome ( Fig . 2a ) , and the post-fission state represented by a separate spherical vesicle and the protein dome completely covered by the membrane ( Fig . 2b ) . Our goals are ( i ) to compute the energies of the two states and to find , by their comparison , the affinity values at which the membrane fission event is energetically favorable , and ( ii ) to determine at which the membrane neck in the fore-fission state becomes as small as guaranteeing fast fission [35] . In the fore-fission state , the extent of the membrane attachment to the protein dome will be characterized by the angle referred below to as the attachment angle which indicates the position of the upper border of the attached area ( Fig . 2a ) . The total energy of the system in the fore-fission state , , is the sum of two contributions . First , the total attachment energy found by integration of the attachment energy density , , over the attached area . Second , the total bending energy of the membrane , , determined by integration of over the whole area of the membrane including and the area of the bud . Taking into account Eq . 1 and the system geometry ( Fig . 2a ) , the total energy of the fore-fission state can be expressed as ( 2 ) The first contribution to the Eq . 2 represents the sum of the attachment energy and the bending energy of the attached membrane portion whose total curvature , , is related to the dome radius , , by . The second contribution is the bending energy of the bud , which depends on the curvature distribution along the bud surface . The third contribution is the energy of the Gaussian curvature , which does not depend on the system configuration . The energy ( Eq . 2 ) has to be minimized with respect to the attachment angle and the distribution of the total curvature along the surface of the bud for any given value of the affinity . This will give the equilibrium values for and the corresponding attached area , determine the equilibrium shape of the membrane bud including its neck radius , and provide the equilibrium total energy of the fore-fission state . Because of a complex shape of the membrane bud , minimization of Eq . 2 will be performed numerically by the standard method of finite elements using the COMSOL Multiphysics software . In the post-fission state , consisting of a spherical vesicle and the hemi-spherical dome covered completely by the membrane ( Fig . 2b ) the total energy is ( 3 ) In the following , we can skip the Gaussian curvature contribution to the fore-fission energy , and account for the addition of to the energy of the post-fission state . CHMP2A/CHMP3 polymers were assembled and analyzed by negative staining electron microscopy as described [34] . CHMP2A/CHMP3 polymers were applied to a holey carbon grid and plunge frozen in liquid ethane . The samples were examined in an FEI F30 Polara microscope , equipped with a Gatan GIF post-column energy filter [47] . Tilt series were acquired over an angular range of 120 degrees , at a nominal magnification of 27 , 500 times , which corresponded to a pixel size of 0 . 49nm , and at a defocus of 5 to 7 microns . Tomograms were generated from these tilt series using the IMOD software package [48] and visualized in Amira ( Visage Imaging ) .
A typical computed shape of the membrane bud corresponding to a certain attachment angle , is presented in Fig . 2a and can be described as a sphere-like cap connected to the attached membrane by a funnel-like neck . The larger the angle , the smaller the neck radius ( Fig . 3 ) . At the attachment angle the neck radius becomes smaller than the threshold value , , which fulfills the condition of the fast fission [35] . Therefore , we limited the considered range of the attachment angles by . Generally , the computation could be stretched to higher attachment angles corresponding to even narrower necks . This would require , however , including in the elastic energy model additional terms of higher order in the curvature of the internal monolayer of the neck , and taking into account the energy of the short range hydration repulsion through the neck lumen between the elements of the internal surface of the neck . Such sophistication of the model would complicate considerably the computation without significant changes of the model predictions on the neck fission . The character of the dependence of the system energy on the attachment angle is determined by the affinity ( Fig . 4 ) . According to the first term in Eq . 2 , the membrane binding to the protein dome will occur only if the affinity exceeds a certain value , , which is the least affinity needed for compensation of the energy penalty of membrane bending accompanying the attachment to the dome surface . At each particular affinity value larger than , the system can reside in a stable or quasi-stable configuration described by the values of corresponding to the energy minima ( Fig . 4 ) . There are four different ranges of the affinity determining different regimes of the possible system configurations . Transitions between these regimes are determined by the three characteristic values of the affinity denoted by , and and presented in Fig . 5 . The first regime corresponds to the affinities smaller than the first characteristic value , . Here , the energy has one minimum at small values , , of the attachment angle ( Fig . 4 ) , meaning that the stable configuration of the system is a bud with a neck whose radius is somewhat smaller than but comparable with the radius of the protein dome . We will refer to this configuration as the broad neck configuration . In the second regime , the affinity varies between the first and the second characteristic values , . In this range , a second energy minimum emerges at the largest possible attachment angle within the considered range , ( Fig . 4 ) , corresponding to a bud with a neck of radius ( Fig . 3 ) . This configuration will be called the narrow neck configuration . The total energy in the second minimum is higher than in the first one , , which means that the narrow neck is a quasi-stable while the broad neck is a stable configuration . It has to be noted that , in contrast to the first energy minimum , the second one is not characterized by a vanishing first derivative of the energy function and represents the minimal energy value found in the considered range of the attachment angle . This feature of the second minimum does not influence , however , the conclusions of the analysis of the membrane fission conditions . In the third regime , the affinity is in the range between the second and third characteristic values , . Under these conditions , the narrow neck is energetically more favorable ( Fig . 4 ) and , hence , becomes stable whereas the broad neck turns quasi-stable . Finally , in the fourth regime the affinity is larger than the third characteristic value , . Here , the energy minimum corresponding to the broad neck vanishes and the only stable state of the system is that of the narrow neck . The three characteristic affinity values , , and , and the geometrical characteristics of the membrane bud in the four regimes of configurations are illustrated in the phase diagrams ( Fig . 5a , b , c ) . The first two phase diagrams represents the total energies ( Fig . 5a ) and the corresponding attachment angels ( Fig . 5b ) of the broad and narrow neck configurations for a specific value of the membrane area . The third phase diagram ( Fig . 5c ) shows how , and depend on the membrane area and , hence , on the area of a vesicle which would form if fission occurs . All the three characteristic affinities decrease with the membrane area which means that the larger the membrane , the lower affinities are needed for generation of buds with narrow necks . Recall that we analyze two requirements for membrane fission . According to the first requirement , the fission reaction has to be energetically favorable meaning that the total system energy in the post-fission state must be lower than in the fore-fission state , . Upon this condition , the fission reaction may be slow because of the existence of kinetic barriers . According to the second requirement , the energy barriers of the fission reaction must , practically , vanish , which guarantees fast rates of the membrane splitting . Particularly , the membrane neck has to narrow up to the threshold value , which guarantees that not just the overall fission reaction but also the intermediate hemi-fission stage is energetically favorable and does not limit the fission rate [35] . The computed system energies in the fore- and post- fission states for different values of the affinity and different moduli of the Gaussian curvature are presented in Fig . 6 . According to these results the first requirement is always satisfied in the narrow neck configuration confirming the previous works . Also for the broad neck configurations the fission reaction may be energetically favorable . To this end the affinity has to be larger than a certain value varying in the range between 0 . 27mN/m and 0 . 37mN/m for feasible values of the Gaussian curvature modulus ( Fig . 7 ) . The more negative is , the looser are the fission conditions , i . e . the lower affinity is needed for fission to be energetically favorable . However , to undergo fission from the broad neck configuration , the system has to overcome a substantial energy barrier and , in practical terms , the membrane splitting will not occur . The requirement of fast fission can be fulfilled if the system reaches the narrow neck configuration . However , to achieve this state in the course of the membrane attachment to the protein dome , the system has to proceed through the whole range of the attachment angles beginning from and up to . This means that the system has to move along one of the energy profiles represented in Fig . 4 . According to Fig . 4 , if the affinity value is smaller than , there is an energy barrier and the system has to overcome to reach the narrow neck configurations . This means that for the membrane fission will be restricted kinetically . At the larger affinity values , , evolution of the membrane bud up to the narrow neck configuration is accompanied by a monotonous decrease of the energy and , hence , proceeds without kinetic restrictions . Summarizing , the condition for the fast fission is . To support the model , we studied the structures resulting from the CHMP2-CHMP3 self-assembly by negative staining [34] and cryo electron tomography ( see Materials and Methods ) . We observed assembly of open tubes , tubes with flat closures , tubes with hemispherical almost closed ends ( defects in closure ) and closed tubular structures with hemi-spherical end-caps ( Fig . 8 ) . The presence of closure defects observed in the structures assembled in vitro might be due to fact that they have been assembled in the absence of membranes . In the current model we propose that these structures assemble directly on membranes . Formation of the closed hemi-spherically capped tubes substantiates the existence of the protein domes which play the central role in the model . These structures should represent the final stage of CHMP2-CHMP3 polymerization and our model suggests that they are physiologically relevant .
According to our computations , the affinity required to drive fission of the membrane neck depends considerably on the area of the membrane fragment undergoing budding and , hence , on the dimension of the vesicle generated in the result of fission ( Fig . 5c ) . The ESCRT-III proteins have been implicated in generation of multivesicular bodies ( MVBs ) consisting of vesicles with characteristic diameters between 20 and 100 nm [19] , [53] and in budding of enveloped viruses with diameters varying up to about 100 nm . Therefore , we performed calculations for the areas of the membrane bud between and corresponding to the relevant range of the vesicle diameters . The largest affinity denoted as is needed to drive a kinetically unconstructed formation of a bud with a narrow neck of radius less which enables fast fission . The affinity ( as well as two other characteristic affinities , and , determining conditions for slower fission processes ) , decreases with increasing membrane area . The maximum value of is needed for generation of the small 20 nm vesicles of MVBs . According to our results ( Fig . 5c ) , the required affinity is . The feasible values of the membrane affinity to the protein dome can be estimated based on a thermodynamic analysis of the kinetic measurements of the CHMP2A and CMHP3 monomer binding to the DOPS-SOPC bilayers [34] . According to these measurements , the CHMP2A and CHMP3 monomers dissociate from lipid with a dissociation rate constant ( koff ) of 0 . 08 s−1 and 0 . 3 s−1 respectively [34] . The association to lipid for both , CHMP2A and CHMP3 , was found to be diffusion controlled thereby putting a lower limit on the association rate constant ( kon ) of 1×106 M−1 s−1 . The condition of equilibrium between the lipid-bound and free protein monomers resulting from the equality of the rates of their association to and dissociation from the lipid can be expressed by the equation ( 4 ) where is the number of the lipid-bound protein monomers , is the number of the lipid molecules and is the volume concentration of the free protein monomers . On the other hand , thermodynamically , the same equilibrium condition can be expressed through the equality of chemical potentials of the lipid-bound and free protein monomers , ( 5 ) where and are the so called standard chemical potentials of the free and lipid-bound protein monomers accounting for the free energy of the direct monomer interaction with the surrounding , and are the contributions of the free and lipid-bound protein monomers from the translational entropy in the solution and on the membrane surface , respectively , is the molar concentration of water molecules . Eq . 5 takes into account that the whole lipid is organized into one or few extended membranes whose translational entropy has a vanishing effect on the chemical potentials . The protein-membrane binding energy per protein monomer is related to the standard chemical potentials by , so that the affinity which represents , according to the definition above , an absolute value of the binding energy related to the unit area of the protein-membrane interface , is given by ( 6 ) where is the area of a CHMP monomer exposed to interaction with the membrane . Combining Eqs . 4–6 we obtain for the affinity . Given the kinetic constants above , and the estimation for the monomer contact area [33] we determine the membrane affinities of CHMP2A and CHMP3 to be and . Taking into account that the protein dome consists of the CHMP2A-CHMP3 heterodimers , the average affinity should be about , which exceeds almost by a factor of six the above estimation of for the affinity required for fast fission of the vesicles . Fission of larger vesicles requires even lesser affinities . Hence , the binding energy provided by the CHMP-membrane interaction must be excessively large and guarantees fast membrane budding and fission under all biologically relevant conditions . The suggested mechanism of membrane fission by the ESCRT-III proteins CHMP2A-CHMP3 and the related calculations demonstrate that dome-like assemblies of these proteins could scaffold membrane necks into strongly curved shapes and favor membrane fission . Since , in contrast to the proteins of the dynamin family , the ESCRT protein complexes attach the membrane to their external surfaces , the fission site emerges within a free membrane fragment aside of the zone of protein-lipid interaction . The task of the CHMP4 and CHMP6 subunits , which are recruited to the membrane upstream of the CHMP2 and CHMP3 recruitment , is to generate an initial membrane bud with a fixed membrane area whose neck has to undergo fission to complete the vesicle formation . A role for Vps4 , in addition to its recycling function , can be in reinforcing the wall of the ESCRT-dome which facilitates membrane bending and fission . It is conceivable that the suggested mechanism is not limited by the action of ESCRT-III proteins but rather has a more general character .
|
Membrane fission is a key step of fundamental intracellular processes such as endocytosis , membrane trafficking , cytokinesis and virus budding . The fission reaction requires substantial energy inputs provided by specialized proteins . Recently , the ESCRT-III proteins have been implicated in membrane budding and fission involved in multivesicular body formation , cytokinesis and virus budding . The ESCRT-III proteins self-assemble into circular filaments and flat spirals in the membrane plane and generate tubular structures with dome-like end caps . We suggest and elaborate computationally on a mechanism by which the ESCRT-III complexes can drive membrane fission . The essence of the mechanism is in generation in the course of membrane attachment to the dome-like surface of an ESCRT-III assembly of a thin membrane neck accumulating large elastic stresses . Relaxation of these stresses can drive the neck fission and formation of separate vesicles of biologically relevant sizes . Estimations of the membrane affinity to the protein surface required for the neck fission to occur and comparison of these values with the experimentally expected values justify quantitatively the proposed mechanism and demonstrate that ESCRT-III assemblies must be highly effective in promoting membrane fission .
|
[
"Abstract",
"Introduction",
"Model",
"Results",
"Discussion"
] |
[
"biophysics/cell",
"signaling",
"and",
"trafficking",
"structures",
"biophysics/theory",
"and",
"simulation"
] |
2009
|
Computational Model of Membrane Fission Catalyzed by ESCRT-III
|
Human bocavirus 1 ( HBoV1 ) belongs to the genus Bocaparvovirus of the Parvoviridae family , and is an emerging human pathogenic respiratory virus . In vitro , HBoV1 infects well-differentiated/polarized primary human airway epithelium ( HAE ) cultured at an air-liquid interface ( HAE-ALI ) . Although it is well known that autonomous parvovirus replication depends on the S phase of the host cells , we demonstrate here that the HBoV1 genome amplifies efficiently in mitotically quiescent airway epithelial cells of HAE-ALI cultures . Analysis of HBoV1 DNA in infected HAE-ALI revealed that HBoV1 amplifies its ssDNA genome following a typical parvovirus rolling-hairpin DNA replication mechanism . Notably , HBoV1 infection of HAE-ALI initiates a DNA damage response ( DDR ) with activation of all three phosphatidylinositol 3-kinase–related kinases ( PI3KKs ) . We found that the activation of the three PI3KKs is required for HBoV1 genome amplification; and , more importantly , we identified that two Y-family DNA polymerases , Pol η and Pol κ , are involved in HBoV1 genome amplification . Overall , we have provided an example of de novo DNA synthesis ( genome amplification ) of an autonomous parvovirus in non-dividing cells , which is dependent on the cellular DNA damage and repair pathways .
Human bocavirus 1 ( HBoV1 ) belongs to the Bocaparvovirus genus in the Parvoviridae family [1 , 2] . HBoV1 is one of a group of etiological respiratory viruses that cause acute respiratory tract infections in young children . Wheezing is one of the most common symptoms of the virus infection [3 , 4] . Acute HBoV1 infection , diagnosed by detection of HBoV1-specific IgM/an increased HBoV1-specific IgG antibody in serum , a virus load higher than 1 × 104 viral genome copy numbers ( gc ) /ml , or HBoV1 mRNA in nasopharyngeal aspirates , or diagnosed HBoV1 viremia , results in respiratory illness [3 , 5–10] . Life-threatening HBoV1 infections in pediatric patients have been reported [11] . Studies of children with pneumonia , acute wheezing , asthma , and/or bronchiolitis suggest that HBoV1 infects the lower respiratory airways down to the bronchioles [3 , 5] . In vitro , HBoV1 infects well-differentiated or polarized human primary airway epithelium ( HAE ) cultured at an air-liquid interface ( HAE-ALI ) [12] . The in vitro model of HAE-ALI , which is derived from primary human bronchial epithelial cells , is a novel system that has provided new insights into the infection characteristics of human respiratory RNA viruses [13 , 14] , as well as respiratory DNA viruses [15] . We have demonstrated that HBoV1 infection of HAE-ALI is long-lasting , persistent , and productive , causing a remarkable loss of epithelial integrity [16 , 17] , which is consistent with the prolonged primary shedding events of HBoV1 for up to a year in patients with respiratory illness [18] . In general , autonomous parvovirus replication is dependent on the S phase of the infected cells because the incoming single-stranded genome of the parvovirus does not support transcription and relies on the host cell DNA replication machinery [19–22] . Except for HBoV1 infection of HAE-ALI , there have been no reports to date of productive infection or viral DNA replication of autonomous parvoviruses in mitotically quiescent cells . Dependoparvovirus adeno-associated virus ( AAV ) of the Parvoviridae family , on the other hand , depends on a helper virus , e . g . , adenovirus or herpes simplex virus , or DNA damaging agents [23] , for its genome replication . These helper viruses induce a cellular environment conducive to AAV replication . AAV DNA replication has been studied extensively in culture of dividing cells; however , how AAV replicates in the context of the non-dividing cells of the host remains elusive [23] . In this report , we studied the mechanism underlying genome amplification of human parvovirus HBoV1 in well-differentiated ( non-dividing ) airway epithelial cells of the HAE-ALI culture . We demonstrated that HBoV1 infection of HAE-ALI induces a DNA damage response ( DDR ) that facilitates viral genome amplification . Importantly , we provide evidence that Y-family DNA repair polymerases , Pol η and Pol κ , are involved in HBoV1 genome amplification . To our knowledge , this is the first report to show that parvovirus DNA replicates in non-dividing cells autonomously .
We examined the cell cycle status of HBoV1-infected cells of HAE-ALI . We used polarized HAE-ALI cultures that had a transepithelial electrical resistance ( TEER ) of >1 . 5 KΩ for infection . We found that the HAE cells of the ALI cultures were well differentiated with p27 expression , which is a marker of G0 phase [24] , but without expression of proliferating cell nuclear antigen ( PCNA ) , which is a marker of cellular DNA replication [25] , or expression of Ki67 , which marks all phases of the cell cycle including S phase [26] . ( S1A , S1B and S1C Fig ) . Therefore , polarized HAE-ALI cultures are largely composed of non-dividing cells . HBoV1 infected p27-expressing cells , as shown by co-immunostaining of anti-p27 and anti-HBoV1 NS1C antibodies ( Fig 1A , p27 ) . The anti-NS1C antibody recognizes both the large and small viral non-structural proteins ( NS ) expressed during HBoV1 infection [27] . HBoV1-infected cells also did not express Ki67 ( Fig 1A , Ki67 ) . Proliferating primary human airway epithelial cells in monolayer culture , for which over half of the cells are proliferating in S phase and do not support HBoV1 DNA replication [16] , were used as negative and positive controls for staining with anti-p27 and anti-Ki67 , respectively , and were not infected by HBoV1 ( S1D Fig ) . We next used a BrdU incorporation assay to pulse-chase viral genome amplification . In this assay , denaturation is necessary to detect the BrdU incorporated in double-stranded DNA ( dsDNA ) , but not single-stranded DNA ( ssDNA ) [28] . In the absence of HCl treatment ( no denaturation ) , NS-expressing cells incorporated BrdU into viral ssDNA , as shown by co-immunostaining of the anti-NS1C and anti-BrdU antibodies ( Fig 1B ) , indicating viral ssDNA synthesis . Notably , under the denaturation condition , mock-infected HAE cells did not incorporate BrdU ( Fig 1C ) , confirming that there was no obvious cellular DNA synthesis in HAE cells as also supported by the lack of Ki67 staining ( Fig 1A ) , which marks all phases of the cell cycle . In infected HAE-ALI , HBoV1 virions were released daily from the apical side , and reached a level of > 1010 gc/well at 16–23 days post-infection ( dpi ) ( Fig 1D , 108 gc/μl ) . Viral ssDNA genome amplification in infected HAE-ALI was confirmed ( Fig 1E and 1F ) , which undergoes intermediates of double and mono replicative forms ( dRF and mRF , respectively ) , a procedure similar to the DNA replication of minute virus of mice ( MVM ) , an autonomous parvovirus [29] . We observed a roughly linear increase in the ssDNA synthesis vs . a several log increase in progeny virion release over time ( Fig 1D and 1E ) . We speculate that the synthesized viral ssDNA genomes are rapidly packaged into capsids , and the matured virions are rapidly released from the cells . Collectively , these results confirmed that HBoV1 amplifies its ssDNA genome in non-dividing airway epithelial cells of the HAE-AL culture and produces progeny virions over the course of infection . Key factors of DNA replication , such as proliferating cell nuclear antigen ( PCNA ) and DNA polymerase ( Pol ) δ , are not expressed in non-dividing cells [30]; therefore , how HBoV1 genome amplifies in infected HAE-ALI without these proteins remains an enigma . We therefore looked into the DNA damage and repair pathways in HBoV1-infected HAE cells . In NS-expressing cells , both RPA32 ( replication protein A 32 ) and histone variant H2AX ( H2A histone family , member X ) were phosphorylated , as detected using antibodies against p-RPA32 ( the RPA32 phosphorylated on serine 33 ) and γH2AX ( the H2AX phosphorylated on serine 139 ) , over the course of infection ( Fig 2A and 2B ) , suggesting that HBoV1 infection induces a DDR . It is thought that three phosphatidylinositol 3-kinase-like kinases ( PI3KKs ) are responsible for the DDR [31 , 32] , we next looked at the activation status of the three PI3KKs . We found that all three PI3KKs , ATM ( Ataxia telangiectasia mutated ) , ATR ( ATM- and RAD3-related ) , and DNA-PKcs ( DNA-dependent protein kinase catalytic subunit ) , were activated in infected cells and colocalized with NS , as assessed by immunofluorescence ( IF ) analysis using antibodies against the specifically phosphorylated site of each kinase ( Fig 2C ) . Phosphorylation of RPA32 , H2AX , ATM , ATR , and DNA-PKcs was also confirmed by Western blotting ( Fig 2D and 2E ) . ATM , ATR and DNA-PKcs are phosphorylated at serine 1981 , threonine 1989 , and serine 2056 , respectively , which are all functional phosphorylation sites that are required for DDR signaling [33–35] . As a control , treatment with hydroxyurea ( HU ) also induced phosphorylation of these proteins ( Fig 2D and 2E ) . To functionally interrogate the requirement for PI3KK activation in mediating viral genome amplification , we applied ATM- , ATR- , or DNA-PKcs-pharmacological inhibitors , which specially inhibit phosphorylation of their respective kinases , to HAE-ALI cultures and evaluated their effects on viral genomes released from the apical surface . Application of an ATM-specific inhibitor , KU60019 [36] , at a concentration of 40 μM , decreased apical virion release by 4–5 log starting at 4 dpi ( Fig 3A ) . Application of the KU60019 also prevented infection-dependent barrier dysfunction , as demonstrated by the lack of a decline in TEER ( Fig 3B ) , no significant dissociation of the tight junction protein Zona occludens-1 ( ZO-1 ) [37] , and no loss of cilia ( β-tubulin IV expression ) ( Fig 3C ) , which were all observed in the vehicle ( DMSO ) treated HBoV1-infected group ( Fig 3B and 3C ) . Application of KU60019 also effectively reduced phosphorylation of ATM in HBoV1-infected HAE-ALI cultures to a level observed in mock-infected cells ( Fig 3D ) . Similarly , we examined an ATR-specific inhibitor AZ20 [38] . At 20 μM , AZ20 inhibited apical virus release by 4 log over the course of 6–23 dpi ( Fig 4A ) , and prevented airway epithelial damage , which was marked by disruption of the TEER ( Fig 4B ) and the dissociation of ZO-1 and no expression of β-tubulin IV ( Fig 4C ) , which were observed in the vehicle-treated HBoV1 infected group ( Fig 4B and 4C ) . Application of AZ20 reduced ATR phosphorylation of HBoV1-infected cells to near background levels observed in mock-infected cells ( Fig 4D ) . Inhibition of apical virus release using the DNA-PKcs-specific inhibitor NU7441 [39] was also substantial and gave results strikingly similar to that of KU60019 . At a concentration of 20 μM , NU7441 decreased apical virion release by 4–5 log over a period of 5–23 dpi ( Fig 5A ) , and prevented the epithelial barrier damage caused by virus infection ( Fig 5B and 5C ) . Applying NU7441 nearly abolished DNA-PKcs phosphorylation in HBoV1-infected cells ( Fig 5D ) . Applying KU60019 , AZ20 and NU7441 alone at the concentrations used did not alter epithelial barrier function . The TEER remained >1 . 6K Ω ( Figs 3B , 4B and 5B , compare Mock/KU , AZ or NU with Mock ) , and cell viability , which was assessed by cellular ATPase activity , was unchanged ( S2 Fig ) . However , the three inhibitors reduced the phosphorylation of their respective kinases in HBoV1-infected cells to a background level of mock-infected cells ( Figs 3D , 4D and 5D ) . Taken together , these results demonstrate that the HBoV1 infection-dependent phosphorylation of ATR , ATM , and DNA-PKcs is critical for HBoV1 genome amplification . To confirm the function of the three PI3KKs in HBoV1 genome amplification , we used ATR- , ATM- , or DNA-PKcs-specific shRNA . We generated lentiviral vectors that co-expressed each shRNA with an mCherry reporter to transduce monolayer cultures of proliferating airway epithelial cells , prior to seeding for ALI cultures . Stable and efficient transduction was evidenced by the expression of mCherry reporter in virtually all the cells of well-differentiated ALI cultures at 4 weeks post-transduction ( S3A Fig ) . At this time , the ATM- , ATR- , and DNA-PKcs-specific shRNA-expressing HAE-ALI cultures demonstrated decreased expression of ATM , ATR and DNA-PKsc , respectively ( by >4-fold ) , but the reduction of PI3KK expression was not observed in the shScram-expressing HAE-ALI ( S3B Fig ) . We then infected these shRNA-expressing ALI cultures with HBoV1 , and analyzed viral DNA replication in them . Viral DNA of either the mRF or ssDNA form in shATM- , shATR- , and shDNA-PKcs-expressing ALI cultures decreased dramatically at both 7 and 22 days , compared to those in shScram-expressing cultures ( Fig 6A ) . Correspondently , apical virus release decreased by 3–4 log from 7 to 22 dpi , in shATM , shATR , and shDNA-PKcs-expressing HAE-ALI , but not in shScram-expressing HAE-ALI ( Fig 6B ) . At 22 dpi , significantly decreased phosphorylation of ATM , ATR and DNA-PKcs was confirmed in their respective shRNA-expressing HAE-ALI , but the HBoV1 infection-dependent phosphorylation in shScram-expressing HAE-ALI remained at the same level as high as that in the HBoV1-infected HAE-ALI ( Fig 6C , 6D and 6E ) . In response to the reduced HBoV1 infection , the shATM , shATR , and shDNA-PKcs-expressing HAE-ALI showed neither a significant decrease in TEER ( Fig 6F ) , nor an obvious dissociation of the tight junction protein ZO-1 , nor a total loss of β-tubulin IV-expressing cilia , which were otherwise observed in infected shScram-applied HAE-ALI ( Fig 6G ) . As controls , shRNA expression alone did not affect the barrier function ( TEER ) ( Fig 6F ) . Taken together , the above results confirmed that all three PI3KKs ( ATM , ATR and DNA-PKs ) play an important role in HBoV1 genome amplification in HAE-ALI . Parvovirus DNA replication follows a rolling-hairpin model of DNA replication , in which , DNA replication factors , i . e . , PCNA , RPA32 and Pol δ , are required [40 , 41] . However , the key DNA replication factors PCNA and Pol δ are not expressed in non-dividing HAE-ALI , as determined by IF analysis and Western blotting ( S1A and S1C Fig , Fig 7A and S4 Fig ) . Similarly , primase Pol α and the leading strand synthesis Pol ε are also not expressed in HAE-ALI , as determined by IF analysis ( Fig 7B and 7C ) and Western blotting ( S4 Fig ) . Thus , we hypothesized that the DNA polymerases utilized in DNA repair might be involved in HBoV1 DNA replication within non-dividing airway epithelial cells . We next examined the Y-family DNA repair polymerase Pol η , Pol ι and Pol κ , B-family polymerase DNA Pol ζ and Pol Rev 1 , and the X-family polymerase Pol β , Pol λ , and Pol μ , which are important DNA polymerases in DNA repair [42] . Notably , Pol η and Pol κ were expressed in non-dividing HAE cells of the ALI cultures ( Fig 7D and 7F ) , while Pol ι ( Fig 7E ) , Pol Rev1 and Pol ζ ( Fig 7G and 7H ) , Pol β , Pol μ , and Pol λ ( Fig 8 ) were not , as determined by IF analysis and also confirmed by Western blotting ( S4 Fig ) . We next visualized interactions between Pol η and Pol κ with nascent replicating viral DNA that was pulse-labeled with BrdU in HBoV1-infected cells using a proximity ligation assay ( PLA ) . We observed clearly positive fluorescent foci in HBoV1-infected HAE cells stained with anti-Pol η and anti-BrdU antibodies , as well as with anti-Pol κ and anti-BrdU antibodies , but not in mock-infected cells ( Fig 9 ) , suggesting a direct interaction of Pol η and Pol κ with the replicating viral genomes . We then sought to knock down Pol η and Pol κ and directly interrogate their functions in HBoV1 genome amplification in HAE-ALI , using lentiviral vectors that expressed Pol η- and Pol κ-specific shRNAs ( shPol η- and shPol κ ) . The lentiviral vector transduction and polarization of the transduced airway cells were conducted in the same manner as the PI3KK shRNA study described above . After HBoV1 infection , the shPol η-expressing HAE-ALI had decreases in apical virion release of 2–3 log and >3 log at 4–8 dpi and 9–18 dpi , respectively; while shPol κ-expressing HAE-ALI showed a decrease of 1 log at 3–4 dpi and of >2 log at 5–18 dpi in apical virus release , compared with the shScram-expressing HAE-ALI ( Fig 10A ) . At 18 dpi , Southern blot analysis of viral DNA replication showed that there was 10-fold and 3-fold reductions in the level of viral ssDNA in shPol η- and shPol κ-expressing cells , respectively , compared with the shScram controls ( Fig 10B ) . Western blotting showed that shPol η and shPol κ knocked down Pol η- and Pol κ , respectively , by 3 . 2-fold and 2 . 5-fold in HAE-ALI ( S3C and S3D Fig ) . In infected HAE-ALI , shPol η-expressing HAE-ALI demonstrated protection from HBoV1 infection-dependent decrease in TEER , while shPol κ and shScram-expressing HAE-ALI did not ( Fig 10C ) . However , both shPol η and shPol κ protected the infected HAE from HBoV1 infection-dependent loss of cilia ( β-tubulin IV expression ) and dissociation of the tight junction protein ZO-1 , to various extents , compared to the shScram-controls ( Fig 10D and 10E ) . The expression of shRNAs alone did not have an obviously deleterious effect on the HAE-ALI , as indicated by the TEER ( Fig 10C ) . Taken together , our results provide evidence that the DNA repair polymerases Pol η- and Pol κ are involved in HBoV1 genome amplification in non-dividing HAE cells .
Autonomous parvovirus DNA replication is thought in general to rely on the activity of host DNA replication machinery of the cells at S phase of the cell cycle during cell proliferation . However , we demonstrate for the first time that genome amplification of a member of autonomous parvovirus occurs in non-dividing cells . We confirmed that productive infection of HBoV1 in non-dividing airway epithelial cells employs the cellular DNA damage and repair machinery to amplify the viral genome . This innovative finding solves the puzzle of how HBoV1 amplifies its genome in terminally differentiated airway epithelial cells and causes structural lesions in the airway . Notably , all three PI3KKs ( ATM , ATR , and DNA-PKcs ) are phosphorylated at sites ( ATM at serine1981 , ATR at threonine1989 , and DNA-PKcs serine 2056 ) that are functionally required to transduce DDR signaling [33–35] . Importantly , we found clues that the Y-family DNA polymerases Pol η and Pol κ play a role in HBoV1 genome amplification . Thus , our study provides direct evidence that the concomitant DDR induced by virus infection recruits cellular DNA repair polymerases , which can be utilized for viral genome amplification . The group of small DNA viruses , including dsDNA papillomavirus and polyomavirus , and ssDNA parvovirus and circovirus , do not encode viral DNA polymerase , and , therefore , they must employ host DNA polymerases for their genome amplification . Most of these small DNA viruses use a strategy of either replicating in dividing cells [40] or inducing infected cells to enter the S phase of the cell cycle by expressing an oncogenic viral protein [43] . There are only a few exceptions of small DNA viruses that replicate in differentiated cells , e . g . , human papillomaviruses ( HPV ) . HPV productive infection is tightly associated with epithelial differentiation [44] , and its active genome amplification is dependent on the activation of the ATM-mediated DNA repair pathway [45 , 46] . However , it remains unclear how HPV employs the DNA repair machinery for viral genome amplification in differentiated epithelial cells [44] . In dividing cells , infection of autonomous parvoviruses induces a DDR with at least one of the PI3KKs activated [47–52] . Activation of ATM is critical to the replication of the Protoparvovirus MVM [49 , 50] and Bocaparvovirus minute virus of canines ( MVC ) [47]; whereas activation of ATR and DNA-PKcs plays a key role in Erythroparvovirus B19 DNA replication [48] . Notably , an apparent cell cycle arrest at S or late S/G2 phase is always accompanied with the DDR induced by infections of these parvoviruses [28 , 51 , 53] . However , on the other hand , autonomous parvovirus replication relies on the host cell DNA replication machinery , and is dependent on the S phase of the infected cells [19–22 , 51] . Therefore , the DDR-facilitating parvovirus DNA replication during infections of MVM , MVC , and B19 is likely a result of cell cycle arrest at the S/late S phase induced by ATM or ATR activation . As for Dependoparvovirus , AAV2 infection in the presence of helper virus induces activation of DNA-PKcs and ATM [54 , 55] . Without helper viruses and any AAV2 protein expression , the infection of UV-inactivated AAV2 infection activates ATR by mimicking a stalled replication fork , and induces G2/M arrest and apoptosis [56] . Unfortunately , all these studies were performed in dividing cells . Of note , recombinant AAV ( rAAV ) vector transduces non-dividing cells efficiently [57] , which is one of the advantages in using rAAV vector in human gene therapy [41] . DNA-damaging agents have been reported to greatly increase the transduction of non-dividing cells by rAAV [58] . Nevertheless , how the ssDNA genome of rAAV converts to transcription-capable dsDNA form in non-dividing cells is still elusive . Additionally , AAV replication can also be stimulated in the absence of helper viruses by treatments that cause cellular genotoxic stress [59] . These agents include hydroxyurea , topoisomerase inhibitors , and UV irradiation . Exactly how these treatments create a favorable environment for AAV replication remains unclear . The involvement of the Y-family DNA polymerase Pol η and Pol κ in HBoV1 genome amplification suggests a DNA repair model of HBoV1 genome amplification . The HBoV1 genome contains heterogeneous terminal repeats at two ends , a DNA molecule similar to ssDNA break that normally induces ATR activation [60] . The hairpinned HBoV1 genome may be recognized by ATR and repaired by the Y-family polymerases following a DNA repair mechanism . DNA polymerase Pol η is a eukaryotic DNA polymerase involved in the DNA repair by translesion synthesis ( TLS ) [42] . It is particularly important in allowing accurate translesion synthesis of DNA damage resulting from ultraviolet radiation , and in some cases , Pol η can perform DNA repair at high fidelity [61] . DNA polymerase Pol κ that is specifically involved in DNA repair , also plays an important role in translesion synthesis , where the normal high-fidelity DNA polymerases cannot proceed and DNA synthesis stalls [62] . The translesion DNA synthesis by the Y-family DNA polymerases is mediated via interaction with mono-ubiquitination of PCNA [63] . However , that fact that PCNA is not expressed in non-dividing airway epithelial cells suggests that HBoV1 genome amplification follows a PCNA-independent DNA repair pathway . The activation of ATM and the importance of its signaling suggest that the homologous recombinational repair ( HRR ) pathway plays a role in HBoV1 genome amplification of differentiated epithelial cells . The high fidelity HRR has been implicated in HPV late gene amplification in differentiated epithelial cells [46 , 64] . Studies have revealed that the majority of the rAAV vector genomes persist as circular episomes of monomers or concatemers in tissues [65 , 66] , and have identified host DNA repair factors in ATM and DNA-PK pathways involved in processing AAV genomes in non-dividing cells [67 , 68] . In addition , during replication of the Protoparvovirus MVM , the 3’ end of the newly synthesized complementary strand is ligated to the right-end hairpin of the viral genome , resulting in the formation of a covalently closed RF DNA , which is the major conversion product [69] . It is impossible that this ligation is carried out by the DNA-PK complex . In fact , DNA-PK has been proved to play a role in AAV replication using both in vivo and in vitro replication assays [70] . The detailed mechanism underlying how ATR , ATM , and DNA-PKcs , functioning either independently or synergistically , mediate HBoV1 genome amplification in non-dividing cells , especially the involvement of the DNA repairing factors , warrants further investigation . Although AAV2 is a Dependoparvovirus , it was observed that the AAV2 genome replicates autonomously in a skin raft model of differentiated keratinocytes [71] . We believe that a differentiation ( DNA repair ) -dependent viral DNA replication probably exists as a general mechanism of parvovirus DNA replication in non-dividing cells . We find HBoV1 infects only cells of polarized primary airway epithelium , and thus , we could not examine the DDR-supported viral DNA replication in other types of non-dividing cells . The non-dividing HAE infection model is essential in understanding the mechanism underlying the role of parvovirus infection-induced DDR signal transduction in facilitating viral DNA synthesis , without disturbing the cell cycle as in dividing cells . We speculate that it is highly likely that other small DNA viruses , in addition to HBoV1 , utilize cellular DNA repair factors , in particular the Y-family DNA polymerases , for their genome amplification in non-dividing cells . Viruses have evolved in a way to utilize various host DNA polymerases depending on which ones are available in the host cells . A concrete understanding of these pathways may enhance the development of anti-viral therapies and also may improve the utility of recombinant vectors that utilizes these viruses for gene therapy .
Primary human airway ( tracheobronchial ) epithelial cells were isolated from the lungs of healthy human donors at Cell Culture Core of the Center for Gene Therapy , University of Iowa , under IRB approval by the Institutional Review Board of the University of Iowa ( IRB ID No . 9507432 ) . We obtained the well differentiated ( polarized ) human airway epithelium ( HAE ) ALI cultures from at the Cell Culture Core without any identification information on them , and , therefore , an IRB review was waived . Primary human airway ( tracheobronchial ) epithelial cells were cultured on collagen-coated , semipermeable membrane inserts ( 0 . 6 cm2 , Millicell-PCF; EMD-Millipore , Billerica , MA; or 0 . 33 cm2 , Transwell , Corning , Tewksbury , MA ) , and then were differentiated at an ALI for 3–4 weeks [16] . This procedure was carried out at the Tissue and Cell Culture Core of the Center for Gene Therapy , University of Iowa . In some circumstances , primary human airway epithelial cells of HAE-ALI cultures were propagated within a fibroblast feeder cell system in F medium , in which cells were co-cultivated with irradiated 3T3 fibroblast ( J2 strain ) with the addition of ROCK inhibitor Y-276322 [15 , 72] , and then were transferred into a Transwell insert ( 0 . 33 cm2 , Transwell ) for ALI differentiation . Briefly , 2×104 airway cells were seeded onto the Transwell insert . In the first 2 to 3 days , F medium was fed in both the apical and basolateral chambers of the insert . Then , F medium was aspirated from both chambers , and the cells were fed only with 500 μl of PneumaCult-ALI medium ( StemCell , Vancouver , BC , Canada ) in the basolateral chamber . The medium was changed every 3–4 days , and the ALI-cultured HAE took 3–4 weeks for full differentiation . We chose the cultures with a transepithelial electrical resistance ( TEER ) of over 1 , 500 Ω∙cm2 , as determined with an epithelial Ohm-voltmeter ( Millicell-ERS; EMD-Millipore ) , for subsequent HBoV1 infection . HBoV1 virions were collected from apical washes of the HBoV1-infected HAE-ALI and were used for infection at a multiplicity of infection ( MOI ) of 1 viral genome copy number ( gc ) /cell , as described previously [17] . At various time points , 100 μl aliquots of phosphate buffered saline , pH7 . 4 ( PBS ) were added to the apical chamber of the HAE-ALI culture , and were harvested as apical washes . All the washes were stored at 4°C for quantification of viral genome copy numbers using a quantitative PCR ( qPCR ) with HBoV1-specific primers and probe , essentially following the method described previously [16] . Hydroxyurea ( HU; Calbiochem , EMD Millipore ) was dissolved in deionized water to make a 200 mM stock solution . The following pharmacological inhibitors were used in this study: KU60019 ( an ATM-specific inhibitor , Tocris Bioscience , Bristol , UK ) , AZ20 ( an ATR-specific inhibitor , Selleckchem , Houston , TX ) , and NU7441 ( a DNA-PKcs-specific inhibitor , Tocris Bioscience ) . All inhibitors were dissolved in dimethyl sulfoxide ( DMSO ) to make stock solutions at 10 mM . Inhibitors were applied 2 days prior to infection , and were included in the ALI medium throughout the experimental period , which was refreshed every three days . Differentiated airway epithelial cells on the ALI membrane support of the HAE-ALI cultures were treated with 0 . 05% trypsin for 5 min , washed once with PBS , and collected in 200 μl of PBS ( ~5 × 104 cells ) . The proliferating primary human airway epithelial cells or the differentiated airway epithelial cells isolated from the ALI membrane were cytospun onto a slide at 2 , 000 rpm for 5 min . The cells were then air-dried for 1 hr at room temperature . For analysis of β-tubulin IV and ZO-1 expression in differentiated cells on the ALI membrane , we directly stained the ALI culture . IF analysis was essentially followed using a method described previously [16] , with antibodies against proteins as indicated in the figures . Confocal images were taken with an Eclipse C1 Plus confocal microscope ( Nikon ) controlled by Nikon EZ-C1 software . DAPI ( 4’ , 6-diamidino-2-phenylindole ) was used to stain the nucleus . Differentiated airway epithelial cells were treated with 5 mM EDTA for 5 min and then trypsinized off the insert of the infected HAE-ALI . Approximately 1 × 105 cells were resuspended in 1 ml of the PneumaCult-ALI medium ( StemCell ) with BrdU ( Sigma , St Louis , MO ) at a final concentration of 30 μM and incubated for 20 min . Next , cells were cytospun onto slides for IF analysis with anti-BrdU and anti-HBoV1 NS1C antibodies . For the detection of cellular DNA replication , BrdU-incorporated cells were further treated with 1 M HCl for 30 min to denature chromosome DNA [28] . PLA was performed using the Duolink PLA Kit ( Sigma ) according to the manufacturer’s instructions . HAE cells were collected from the Transwell insert and were labeled with BrdU as described above . At room temperature , the cells were fixed with 3 . 7% paraformaldehyde for 15 min , permeabilized with 0 . 2% Triton X-100 for 5 min , and blocked with Duolink Blocking Buffer for 30 min . Then , the cells were incubated with primary antibodies , mouse anti-BrdU and rabbit anti-Pol η or with mouse anti-BrdU and rabbit anti-Pol κ , for 1 hr . Two diluted PLA probes , which are specific to mouse and rabbit IgG , respectively , were applied to the cells and incubated for 60 min at 37°C . The hybridized oligonucleotides were ligated in the Ligation Solution at 37°C for 30 min and amplified in Amplification Solution for 100 min . Finally , the cells were washed and mounted with Duolink In Situ Mounting Medium with DAPI and visualized under a Nikon Eclipse C1 Plus confocal microscope . For Western blotting , the HAE cells on the insert of the ALI culture were lysed in 200 μl of 1 × SDS-loading buffer . Lysed samples were loaded for SDS-polyacrylamide gel electrophoresis ( PAGE ) , transferred , and blotted with antibodies as indicated in the figures , as previously described [16] . Images were developed under the imager FUJIFILM LAS 4000 ( FUJIFILM Life Sciences ) and quantified with Multi Gauge V2 . 3 software ( FUJIFILM Life Sciences ) . For Southern blotting , HAE cells were trypsinized off the insert of the ALI culture , washed and collected for extraction of low molecular weight ( Hirt ) DNA [73 , 74] . Southern blotting was performed using an HBoV1 NS and Cap gene probe , as previously described [27] . A mitochondrial DNA probe was used as a control for the recovery of the Hirt DNA [75] . Images were developed with a Typhoon FLA 9000 phosphor imager and quantified using ImageQuant TL 8 . 1 ( GE Healthcare ) . pLKO-mCherry backbone vector was constructed by inserting a CMV-driven mCherry gene into pLKO . 1 vector ( Addgene , Inc . , Cambridge , MA ) through the BstB I and Mfe I sites . shRNAs sequences , which are generated by annealing oligonucleotides synthesized at Integrated DNA Technologies ( IDT; Coralville , IA ) , were cloned into the pLKO-mCherry using the Age I and EcoR I sites . Lentiviral vectors were produced and purified as previously described [76] . To generate shRNA-expressing HAE-ALI cultures , proliferating airway epithelial cells cultured as monolayer were infected with lentiviral vector at an MOI of ~10 . After 1 day , transduced cells were transferred into Transwell inserts . After 2–3 days , PneumaCult–ALI medium ( StemCell ) was used to establish an ALI for polarization , as described above . Cell viability was quantified using a Cytotoxicity Assay kit ( Promega , Madison , WI ) following the manufacturer’s instructions . Briefly , HAE-ALI cultures were treated with KU60019 ( 40 μM ) , AZ20 ( 20 μM ) , and NU7441 ( 20 μM ) for 23 days . Staurosporine was used as a positive control in different final concentrations ( 2 , 20 , and 200 μM ) for 2 days . DMSO at 0 . 1% was used as a vehicle control . At the end of treatment , HAE cells were collected from the Transwell inserts and seeded into a 96-well plate , followed by addition of the cytotoxicity assay reagents . After incubations , luminescence was measured by a Synergy H1 microplate reader ( BioTek U . S . , Winooski , VT ) . Then , lysis reagent was added into the mixtures , after incubation luminescence was measured again . The dead cell numbers and total cell numbers were determined from the first luminescence and second luminescence results , respectively . The cell viabilities were normalized to the “Untreated” group . Rat anti-HBoV1 NS1C antibody was produced previously [77] . The following antibodies were purchased: anti-p27 , anti-PCNA , anti-Ki67 , and anti-BrdU from BD Biosciences ( San Jose , CA ) , anti-phosphorylated H2AX ( γ-H2AX ) from Millipore , anti-phosphorylated replication protein A32 ( p-RPA32 on serine 33 ) , anti-p-ATR ( Thr1989 ) , anti-Pol ι , and anti-Pol ε from GeneTex ( Irvine , CA ) , anti-p-ATM ( Ser1981 ) , anti-p-DNA-PKcs ( Ser2056 ) , anti-ATR , anti-Pol κ , anti-Pol ζ , and anti-Pol η from Abcam ( Cambridge , MA ) , anti-ATM from Cell Signaling Inc . ( Danvers , MA ) , anti-DNA-PKcs from Biolegend , anti-Pol α , anti-Pol δ , anti-Rev1 and anti-Pol β from Santa Cruz ( Dallas , Texas ) , and anti-β-actin from Sigma . An anti-Pol ι antibody from Bethyl Laboratories , Inc . ( Montgomery , TX ) was used for Western blotting . The following shRNA sequences were chosen to target the genes of interest: shRNA specific to ATM ( shATM ) , 5’- CCG GGA TTT GCG TAT TAC TCA GTC TCG AGA CTG AGT AAT ACG CAA ATC CTT TTT G-3’ [78]; shRNA specific to ATR ( shATR ) , 5’- CCG GGG CGT CGT CTC AGC TCG TCT CCT CGA GGA GAC GAG CTG AGA CGA CGC CTT TTT G-3’; shRNA specific to DNA-PKcs ( shDNA-PKcs ) [78] , 5’-CCG GGA TCG CAC CTT ACT CTG TTC TCG AGA ACA GAG TAA GGT GCG ATC TTT TTG-3’ [79]; shRNA specific to Pol η ( shPol η ) , 5’-CCG GCC CGC TAT GAT GCT CAC AAG ACT CGA GTC TTG TGA GCA TCA TAG CGG GTT TTT G-3’ [80]; shRNA specific to DNA Pol κ ( shPol κ ) , 5’-CCG GGC CAT TGC TAA GGA ATT GCT ACT CGA GTA GCA ATT CCT TAG CAA TGG CTT TTT G-3’ ( Sigma , TRCN0000115999 ) . The following scrambled shRNA ( shScram ) was used as an shRNA control: 5’-CCG GCC TAA GGT TAA GTC GCC CTC GCT CGA GCG AGG GCG ACT TAA CCT TAG GTT TTT G-3’ [48] .
|
Parvovirus is unique among DNA viruses . It has a single stranded DNA genome of ~5 . 5 kb in length . Autonomous parvoviruses , which replicate autonomously in cells , rely on the S phase cell cycle for genome amplification . In the current study , we demonstrated that human bocavirus 1 ( HBoV1 ) , an autonomous human Bocaparvovirus , replicates its genome in well-differentiated ( non-dividing ) primary human airway epithelial cells . HBoV1 infection of non-dividing human airway epithelial cells induces a DNA damage response . We provide evidence that HBoV1 genome amplification in non-dividing airway epithelial cells is facilitated by the DNA damage response-mediated signaling pathways . Importantly , we discovered that two Y-family DNA repair polymerases , but not cellular DNA replication polymerases , are directly involved in HBoV1 genome amplification . Therefore , our study is innovative because it is the first to show that an autonomous parvovirus amplifies its genome in non-dividing cells , and that the DNA repair polymerases are involved in viral genome amplification .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2016
|
Replication of an Autonomous Human Parvovirus in Non-dividing Human Airway Epithelium Is Facilitated through the DNA Damage and Repair Pathways
|
Cripto , the founding member of the EGF-CFC genes , plays an essential role in embryo development and is involved in cancer progression . Cripto is a GPI-anchored protein that can interact with various components of multiple signaling pathways , such as TGF-β , Wnt and MAPK , driving different processes , among them epithelial-mesenchymal transition , cell proliferation , and stem cell renewal . Cripto protein can also be cleaved and released outside the cell in a soluble and still active form . Cripto is not significantly expressed in adult somatic tissues and its re-expression has been observed associated to pathological conditions , mainly cancer . Accordingly , CRIPTO has been detected at very low levels in the plasma of healthy volunteers , whereas its levels are significantly higher in patients with breast , colon or glioblastoma tumors . These data suggest that CRIPTO levels in human plasma or serum may have clinical significance . However , very little is known about the variability of serum levels of CRIPTO at a population level and the genetic contribution underlying this variability remains unknown . Here , we report the first genome-wide association study of CRIPTO serum levels in isolated populations ( n = 1 , 054 ) from Cilento area in South Italy . The most associated SNPs ( p-value<5*10-8 ) were all located on chromosome 3p22 . 1-3p21 . 3 , in the CRIPTO gene region . Overall six CRIPTO associated loci were replicated in an independent sample ( n = 535 ) . Pathway analysis identified a main network including two other genes , besides CRIPTO , in the associated regions , involved in cell movement and proliferation . The replicated loci explain more than 87% of the CRIPTO variance , with 85% explained by the most associated SNP . Moreover , the functional analysis of the main associated locus identified a causal variant in the 5’UTR of CRIPTO gene which is able to strongly modulate CRIPTO expression through an AP-1-mediate transcriptional regulation .
Cripto , also known as Teratocarcinoma-derived growth factor 1 ( TDGF1 ) , is the original member of the Epidermal Growth Factor-Cripto/Fibroblast growth factor Receptor Ligand 1/Cryptic ( EGF-CFC ) family of vertebrate proteins involved in embryo development [1–3] . Cripto has been isolated in human and mouse [3] and is a GPI-anchored membrane protein [4] that can function in both membrane anchored and soluble form [5 , 6] . It is involved in multiple signaling pathways , such as TGF-β , Wnt and MAPK/ERK pathways , it regulates essential steps in early embryogenesis and it is also involved in processes such as cell migration , epithelial-mesenchymal transition ( EMT ) , stem cell maintenance , all processes which are implicated in cancerogenesis [7–14] . Cripto has also a role in angiogenesis , being able to enhance the proliferation , migration and invasion of human umbilical endothelial cells , to stimulate their differentiation into vascular-like structures in Matrigel and is also able to induce tumor neovascularization in vivo [13] . Cripto is expressed at very low levels in different adult tissue types and organs , among them a higher expression is detected in colon , skeletal muscle , heart , cortex of adrenal gland and cerebellum ( http://www . biogps . org/ [15] , http://www . proteinatlas . org/ [16] ) . Pathological re-expression is seen in a number of solid cancers . Most studies focused on the role of Cripto in breast and colorectal cancer [17–20] , in inflammatory conditions and also in a macaque model of neuroAIDS [21] . Numerous studies have demonstrated correlation between high expression levels of CRIPTO and malignant transformation , tumor invasiveness , metastatic spreading and hence poor prognosis [17 , 22–27] . In vitro and in vivo functional studies confirm a strong involvement of Cripto in cancer development and indicate that its effect on tumorigenesis might strictly depend on the cellular context in which it acts [28–33] . Moreover , many data indicate Cripto as a promising target for cancer therapy . Adkins et coll . demonstrated that the block of Cripto signaling with an anti-CFC domain antibody determined a strong inhibition of tumor cell growth in vivo [34] . Ever since , different approaches based on the use of oligonucleotides , vaccines or antibodies have been successfully applied to target Cripto by inhibiting its activity and/or expression in tumors and in neurodegenerative diseases [35] . Cripto inhibition by different approaches always resulted in inhibition of cancer cell proliferation in vitro and of tumor growth in vivo [36] . CRIPTO has been also detected at very low levels in the plasma of healthy controls , whereas significantly higher concentrations have been found in patients with breast , colon or cerebral tumors [19 , 37] . In both studies , tumor tissues and patient-matched blood samples have been analyzed , showing that CRIPTO high levels in the plasma correspond to re-expression of CRIPTO in tumor tissues [19] . These data suggest that CRIPTO , as the carcinoembryonic antigen ( CEA ) ( another GPI anchored protein and a widely used tumor marker ) , is able to reach the bloodstream , being potentially released by tumor cells through GPI anchor cleavage . All together these data indicate that CRIPTO represents both a promising biomarker and a valid target for therapeutic intervention in cancer and that blood CRIPTO levels in humans may have clinical significance . The two studies on the measure of circulating CRIPTO in the plasma published so far were both conducted on a very small group of individuals: the study of Bianco and coworkers analysed 21 healthy donors , 33 patients with colon carcinoma and 75 patients with breast carcinoma or benign breast lesions while the study of Pilgaard and coworkers included 28 Glioblastoma Multiforme ( GBM ) patients , 4 low-grade glioma patients and 8 healthy controls . In the first case no statistically significant correlations were observed between CRIPTO plasma concentration and various clinicopathologic variables , including tumor size , lymph node involvement and proliferative index and the degree of positivity for CRIPTO in tumor sections [19] . In the second case higher levels of the protein correlated with a shorter overall survival [37] . In the present study , CRIPTO serum levels were measured in a population-based sample from three isolated villages of the Cilento area , South Italy and a very high heritability ( >80% ) was estimated underscoring the importance of its genetic determinants . We present the first genome-wide association study ( GWAS ) for CRIPTO protein levels aiming at identifying genetic variants associated with the levels of circulating protein in the serum and we report the functional characterization of the main associated locus .
A quantile-quantile plot for the 6 , 222 , 455 investigated autosomal SNPs in discovery GWAS revealed many more SNPs with low observed p-values than expected ( S1 Fig . ) . 455 SNPs associated with CRIPTO serum concentration at p-value<5*10–8 in the discovery stage were located on chromosome 3 in a region spanning 6 . 4 Mb ( 3p22 . 11–3p21 . 31 ) , with the most associated SNP ( rs3806702 , p-value = 1 . 03*10–159 ) located in the CRIPTO gene region . ( Figs . 1 and S2 ) . The C allele of this SNP was associated with higher levels of CRIPTO ( CC = 673 . 8±29 . 0pg/ml; CT = 310 . 5±6 . 9pg/ml; TT = 46 . 7±4 . 4pg/ml ) ( Table 2 ) . When running a conditional GWAS adjusting for the most associated SNP on chromosome 3 , no loci remain associated at the genome-wide significance level , but 700 SNPs were associated with p-value<1*10–4 on the genome ( S1 Table ) . Of those , linkage disequilibrium ( LD ) -based independent variants were defined if the pair-wise LD ( r2 ) was less than 0 . 01 and if they were separated by at least 1 Mb . In this way , 95 LD-based independent loci were identified that included SNPs associated with CRIPTO serum concentration at a p-value<1*10–4 . For each of the independent loci , the SNP with the lowest p-value was carried forward to replication . Criteria for replication were defined as a p-value<0 . 05 in the independent replication sample , the effect in the same direction between discovery and replication , and a p-value in the meta-analysis of discovery and replication samples lower than that obtained in the discovery sample ( see Methods ) . Table 2 shows the summary results for the main association and the additional 5 replicated loci for CRIPTO serum levels . Regional association plots provide a detailed overview of those loci ( S2 Fig . ) . The proportion of serum CRIPTO variance explained by the replicated loci in the discovery sample was 87 . 0% , with 84 . 9% explained by rs3806702 alone and 2 . 1% explained by the remaining 5 associated SNPs . Similar results were obtained in the replication sample , with 87 . 1% of CRIPTO variance explained by the 6 associated variants . To explore the functional relationship between CRIPTO and the associated genes we used the Ingenuity Pathway Analysis software ( IPA ) . CRIPTO gene and genes closest to the replicated SNPs were included in the analysis ( see Methods ) . Those 10 seed genes were reported in the S2 Table . The IPA analysis identified a significant ( p-value = 1*10–6 ) network of 35 molecules , 3 of which corresponded to CRIPTO associated loci ( CRIPTO itself , GAS-7 and TNS-1 ) ( Fig . 2 ) . Function categories assigned to the whole gene network by IPA were: “cell to cell signaling and interaction , hematological system development and function , and immune cell trafficking” . In depth , function annotation sub-categories significantly enriched in genes involved in the network and including CRIPTO , were: cell migration , proliferation of tumor cells , cell differentiation , and blood vessel development ( S3 Table ) . To identify functional elements in the associated loci , ENCODE data related to chromatin modifications and hypersensitivity DNAse elements were analyzed in 3 cell lines ( NTERA-2 , HepG2 , H1-hESC ) selected as expressing CRIPTO mRNA . Among the replicated SNPs and variants in LD with them ( r2>0 . 5 ) , 3 variants ( rs3791936 in the intronic region of TNS1 , rs112481213 in the 5’UTR of CRIPTO , and rs61117007 in the intronic region of NTRK2 ) were located in promoter or enhancer histone marks and 1 of those ( rs112481213 ) was also located in DNA hypersensitivity elements in the selected cell lines , suggesting a potential functional role of those associated variants . The top 8 associated SNPs in the CRIPTO gene region were all included in a unique LD block ( S3A–S3B Fig . ) . Among them the rs112481213 variant ( LD with rs3806702 , r2 = 0 . 99 , p-value of association = 1 . 53*10–158 ) , reported to be in a regulative region from ENCODE data ( see the above paragraph ) , was also predicted to create an AP-1 binding site by bioinformatics analysis with MatInspector program and TRANSFAC database ( S3C Fig . ) . To explore the possibility that this SNP , located in the 5’UTR of CRIPTO gene , might influence CRIPTO transcription we tested the SNP allele effect on the transcriptional activity in the NTERA2 teratocarcinoma cell line , expressing high levels of CRIPTO [38] . Two constructs , both containing a 1 , 051 bp region upstream the CRIPTO ATG start codon ( including almost all the 5’UTR and also 342 bp upstream the transcription start site ) , but differing for the rs112481213 allele ( -222A/luc and-222T/luc , respectively ) were transfected in NTERA2 cells . Overall , 15 SNPs were included in that region . As one of those SNPs , the rs3806703 variant was in linkage disequilibrium with rs112481213 ( r2 = 0 . 72 ) , two additional constructs ( -222T→A/luc , -222A→T/luc ) were also produced by site-directed mutagenesis to discriminate the effect of each of the two variants on the transcription efficiency ( Fig . 3A ) . The construct containing the rs112481213 A allele ( -222A/luc ) produced an about 5-fold increase ( p-value = 9 . 2*10–6 ) of the luciferase activity compared to the construct containing the T allele ( -222T/luc ) indicating that rs112481213 SNP affects the transcription ( Fig . 3B ) . Also , the luciferase activity was reported to high levels in the site-mutated-222T→A/luc construct demonstrating that the main effect on the transcription can be attributed to the rs112481213 variant . Interestingly , the activity produced by the-222T→A/luc is significantly higher than that observed for the-222A/luc ( p-value = 2 . 9*10–2 ) and inverse results were obtained for-222A→T/luc and-222T/luc ( p-value = 5 . 8*10–4 ) , indicating that rs3806703 might also have an effect , although modest , on the transcription . A cooperative role of rs112481213 and rs3806703 was also statistically supported by an interaction model tested for the two variants ( βrs112481213 = -1 . 18 , SErs112481213 = 0 . 04 , βrs3806703 = 0 . 27 , SErs3806703 = 0 . 06 , βinter = -0 . 15 , SEinter = 0 . 03 , p-valueinter = 7 . 08*10–8 ) . Therefore , the promoter activity data demonstrated that rs112481213 is a functional regulatory element of CRIPTO transcription which effect might be modulated by rs3806703 . We next verified that the mechanism through which rs112481213 variant might influence CRIPTO transcription is the creation of an AP-1 consensus binding site , as predicted by bioinformatics analysis . Indeed , Electrophoretic Mobility Shift Assay ( EMSA ) using an AP-1 containing PC3 nuclear extract suggested that the A allele-containing oligonucleotide probe might robustly bind AP-1 complex while an oligonucleotide probe containing the T-allele weakly bound the complex ( Fig . 4A ) . The AP-1 binding was confirmed by addition of specific antibodies for components of AP-1 complex , revealing a visibly supershifted band , representing a DNA-protein-antibody complex ( Fig . 4B ) . All together , these data demonstrate the specificity of the observed DNA-protein interaction as well as a differential interaction of the AP-1 complex with the rs112481213 alleles .
Cripto is a typical example of an oncodevelopmental gene having key functions in early embryogenesis , and being re-expressed in the adult during tumorigenesis . Cripto is a GPI-anchored membrane protein , that can also be cleaved and released in the medium and is able to induce cellular proliferation , EMT , migration , and invasion , as well as to stimulate tumor angiogenesis both in vitro and in vivo [36] . Cripto promotes oncogenesis via modulation of TGF-β ligand signaling and through mechanisms that are independent of TGF-β ligands and their signaling receptors [39] . The effect of TGF-β ligands and Cripto on tumorigenesis is also dependent on the cellular context [33 , 39] . Interestingly , Cripto protein is an obligatory co-receptor for some TGF-β family members such as Nodal , enabling them to bind to Activin receptorial complexes and activate Smad cascade and is also able to antagonize the signaling of other members of the TGF-β family , ( i . e . , Activins and TGF-β ) , inhibiting their antioncogenic effect [34 , 35 , 40] . Moreover , Cripto acts via separate , non-overlapping mechanisms to enhance the canonical Wnt/β-catenin signaling pathway by binding to low-density lipoprotein receptor-related protein ( LRP ) 5 and LRP6 [41] and to activate ras/raf/MAPK and PI3K/Akt pathways via c-Src [28] . More recently , novel Cripto-interacting proteins , also involved in cancer , have been identified including the chaperonin glucose regulated protein-78 ( Grp78 ) and Notch1 [39] . In humans , CRIPTO is expressed at very low levels in both normal tissues and plasma , while its expression was found increased in patients with cancer ( also in both tumor tissues and plasma ) , suggesting that CRIPTO blood levels might have a great clinical relevance [19 , 37] . However in these two articles only a small number of healthy volunteers was analysed and no data on a general population sample were reported . Our study is the first GWAS of circulating CRIPTO levels . It was undertaken in 1 , 589 individuals from three population isolates of the Cilento area , South Italy and represents the largest survey of CRIPTO measurement in a population-based sample . In our study , the variability of the circulating protein , according to the heritability estimation , was found to be highly determined by genetic factors . The GWAS identified the strongest association on chromosome 3 at rs3806702 located in the CRIPTO gene region . GWAS conditional to that variant showed that this locus represents the main genetic contribution to the modulation of CRIPTO in the serum . Indeed , 85% of the inherited component of circulating CRIPTO levels is explained by the rs3806702 variant . In accordance with this finding , all individuals with CRIPTO protein levels below the detection threshold ( 34% of the entire sample ) were found homozygous for the T allele of the rs3806702 variant . The lower levels of CRIPTO in the discovery sample compared to the replication sample can also be explained by the difference in allele frequency ( for T allele 0 . 77 and 0 . 67 respectively ) . A LD block included rs3806702 as well as the top 7 associated SNPs in the CRIPTO gene region . Some of those were reported by bioinformatics analysis as potential candidates affecting transcription factor binding sites . Among those variants , rs112481213 , located in the 5’UTR of CRIPTO gene , was identified by functional experiments as a causal SNP for CRIPTO transcriptional regulation . Regulation of Cripto expression during embryogenesis and tumorigenesis was still incompletely defined . So far different binding sites have been found in the promoter region of the Cripto gene: for Smad-proteins [42] , the T-cell factor/lymphoid enhancer factor ( Tcf/Lef ) [43] , the Hypoxia-Inducible Factor 1 ( HIF-1 ) [44] , the Nkx2–5 early cardiac transcription factor [45] , and the orphan nuclear Liver Receptor Homolog-1 ( LRH-1 ) [46] . HIF-1 and Nkx2–5 are able to transcriptionally activate Cripto during cardiac differentiation , HIF-1 also activates CRIPTO expression in human embryonal carcinoma cells , following hypoxic conditions [44 , 46] . Conversely , CRIPTO is directly repressed by the orphan nuclear receptor germ cell nuclear factor ( GCNF ) which binds to the promoter during retinoic acid-induced differentiation of human embryonic carcinoma cells and by the miR-15a/16 cluster which bind to the 3’UTR of CRIPTO mRNA [47 , 48] . However , this is the first time that an AP-1 transcriptional activation of CRIPTO has been described . We have indeed demonstrated that CRIPTO expression is regulated by AP-1 transcription factor and that this regulation depends on rs112481213 genotype . Transcriptional activity data also suggest that in addition to rs112481213 , rs3806703 , another SNP present in that region , may have a role in modulating the CRIPTO protein levels , possibly through the involvement of GATA binding transcriptional factors . In support to this hypothesis , a statistical interaction between rs112481213 and rs3806703 was also found . Due to the complexity of Cripto gene regulation and its dependency on the specific biological context , additional regulatory mechanisms might occur in the case of cancer-related cell dysfunction . Further , five additional loci were associated to CRIPTO serum levels at p-values<1*10–4 independently from the main signal . Although these associations did not reach genome-wide significance in the discovery , likely because of lack of power of our study , they were replicated in an independent sample and might represent good candidate loci as modulators of the circulating CRIPTO . Moreover , three of the replicated variants were located in regions involved in regulative processes associated to chromatin accessibility . A single network of 35 molecules including CRIPTO and two other associated loci , the growth-arrest-specific gene7 ( GAS-7 ) and the Tensin1 ( TNS-1 ) , was identified by IPA analysis . The network included 30 genes implicated in cell migration , 18 genes involved in tumor cell proliferation , 25 genes in cell differentiation and , interestingly , 17 genes implicated in blood vessel development . The analysis of CRIPTO associated loci showed that these are mainly linked to the MAPK/ERK signaling pathway with ERK1/2 as one of the principal players of the network . Aberrant regulation of MAPK cascades is known to strongly contribute to cancer and other human diseases . Like CRIPTO , TNS-1 and GAS-7 , the other two genes in the associated regions present in the network , are both involved in breast and colon cancer [49–52] . TNS-1 binds to actin filaments [53] and serves as a link between signal transduction pathways and the actin cytoskeleton by forming a structural platform that regulates the assembly of focal adhesion components , phosphoproteins , and signaling molecules for processes such as cell migration [54] . TNS-1 is expressed in normal tissues [55] while its expression is greatly reduced in human breast , prostate , head and neck squamous cell carcinomas , and melanoma suggesting a role as tumor suppressor [56] as well as in the maintenance of cell polarization , and the suppression of invasion that are involved in metastasis [57] . The TNS-1 phosphotyrosine binding ( PTB ) 1 domain binds the cytoplasmic tail of beta-integrin , presumed to be the basis for focal adhesion localization . Interestingly , overexpression of Cripto in vitro and in vivo has been associated with increased expression of fibronectin and various integrins and with increased activation of focal adhesion kinase [58] . Similarly to TNS-1 , GAS-7 binds actin and participates in cytoskeleton dynamics , executing different functions in different cellular processes , such as vesicle trafficking , cell migration and morphological differentiation [59 , 60] . GAS-7 hypermethylation has been found in breast and colon cancers whereas increased expression has been detected in medulloblastoma [61] . GAS-7 expression is regulated by ERK signaling pathway [62] , in which Cripto is also involved . In the same CRIPTO associated locus on chromosome 17 , besides GAS-7 , is also located the G-protein-coupled receptor ( GLP-2R ) gene [63–65] . Interestingly , GLP-2R activation also induces ERK1/2 MAP kinase activation and is able to both stimulate the expression of the immediate early genes c-Fos , c-Jun , JunB and Egr-1 and to activate AP1-driven gene transcription in a PKA-dependent manner [66 , 67] . Two other genes included in the associated loci , Myosin VA ( MYOVA ) gene on chromosome 15 and Neurotrophin tyrosine kinase receptor 2 ( NTRK2 ) on chromosome 9 , are both overexpressed in cancer . In particular , MYOVA is highly expressed in a number of highly metastatic cancer cell lines and metastatic colorectal cancer tissues and is able to interfere with metastatic capabilities by influencing cell migration . MYOVA expression is upregulated by the transcription factor Snail , one of the molecular switches for the EMT program involved in cancer metastasis [68] . NTRK2 has been found frequently overexpressed in human cancers , including pancreatic and prostate carcinoma , Wilms’ tumor and neuroblastomas , particularly those with aggressive behavior and poor prognosis . As Cripto , NTRK2 activates both phosphatidylinositol-3-kinase ( PI3K ) and MAPK/ERK signaling . In summary , our data showed that CRIPTO protein is measurable in the serum of the majority of individuals in a population-based sample . Further , we identified the largest genetic contribution to the CRIPTO variability and demonstrated that a functional variant located in the 5’UTR of CRIPTO gene is able to modulate CRIPTO expression through an AP-1-mediated transcriptional regulation . We also provided support for additional associated loci that will need to be confirmed in larger samples . Nevertheless , many of those associations converge in cancer phenotypes , mainly in cell movement and proliferation functions . As any association has been detected at these CRIPTO associated loci in large-scale cancer GWAS , further studies looking at CRIPTO variability in serum together with genotyping of the functional variant in specific cohorts of patients , focusing the analysis on specific cancer phenotypes , as metastasis formation , aggressiveness , prognosis , would be useful to better investigate possible associations between variants modulating CRIPTO protein levels and cancer features .
The discovery sample includes 1 , 054 individuals recruited through a population-based sampling strategy in two small isolated villages of the Cilento region , South Italy ( Gioi and Cardile ) [69] . In silico replication was performed in additional 535 subjects from another village ( Campora ) of the same region [70] . The study design was approved by the ethics committee of Azienda Sanitaria Locale Napoli 1 . The study was conducted according to the criteria set by the declaration of Helsinki and each subject signed an informed consent before participating to the study . Blood samples were collected in the morning after the participants had been fasting for at least 12 h . Aliquots of serum were immediately prepared and stored at -80°C , and were subsequently used for the assessment of CRIPTO levels . CRIPTO ( pg/ml ) was measured using an enzyme-linked immunosorbent assay , according to the manufacturer’s instructions ( DRG Instruments GmbH , Germany ) . An intra-assay coefficient of variation of the CRIPTO measure of 7 . 74% was obtained from 10 times measurements of 10 serum samples . Individuals with CRIPTO levels below the detection threshold were included in the study and a value of 0pg/ml was assigned to them . Mann-Whitney U test was used to compare median CRIPTO serum levels among the samples . A normal quantile transformation was applied to the trait and the transformed trait was used in all statistical analyses . The heritability of CRIPTO serum levels was estimated by SOLAR software [71] using extended genealogies of discovery and replication population samples and adjusting the phenotype for gender and age . Genotyping was performed with 370K and Omniexpress Illumina chips , phasing and imputation were conducted separately by platform with the MaCH [72] and minimac ( http://genome . sph . umich . edu/wiki/Minimac ) software respectively , using 1000G v3 data as reference . SNP allele frequencies in Cilento samples versus 1000 Genomes reference allele frequencies for all genotyped SNPs were reported in the S4 Fig . Quality control filters applied before imputation were call rate >95% for SNPs and samples and minor allele frequency ( MAF ) >0 . 01 . GWAS was carried out through a mixed model linear regression where the variance/covariance matrix is the genomic kinship to account for relatedness between individuals . Age and gender were used as covariates and an additive genetic model was considered . The analysis was performed with GenABEL package [73] for genotyped SNPs and ProbABEL [74] for imputed data . SNPs with imputation quality ( Rsq in MACH ) <0 . 4 or MAF <0 . 05 were excluded . Conditional analysis was carried out in the discovery and replication samples adding the additive effect of rs3806702 as covariate in the association model . To select linkage disequilibrium ( LD ) -based independent association signals among the CRIPTO associated SNPs , we conducted the clumping procedure implemented in PLINK [75] and picked the index SNPs with the most significant association p-value from each clumped association region based on the GWAS . The 1000G v3 genotypes were used as reference panel; the physical threshold for clumping was 1 Mb , and the r2 threshold for clumping was 0 . 01 . To assess evidence for replication , test-statistics of discovery and in silico replication samples were meta-analysed using a fixed effect model weighted by inverse variance , using Metal [76] . SNPs were considered replicated if the SNP p-value was <0 . 05 in the replication sample alone , the effect was in the same direction between discovery and replication , and the p-value in the meta-analysis was lower than in the discovery sample . The percentage of the variance of the CRIPTO levels explained by the replicated SNPs was calculated both in the discovery and replication samples . Three linear mixed effects models were fitted , in which the CRIPTO was regressed , respectively , on: 1 ) gender and age ( basic model ) ; 2 ) gender , age , additive effect of a single SNP ( single SNP model ) ; 3 ) gender , age , additive effect of each of the replicated SNPs ( multiple SNP model ) . The variance explained by each SNP was calculated as the difference between the variance explained by the single SNP model and that explained by the basic model . Similarly , the variance explained by the replicated SNPs all together was estimated as the difference between the variance explained by the multiple SNP model and that explained by the basic model . The lmekin function ( R package ) , which uses the genomic kinship matrix to correct for relatedness between individuals , was applied . To test for interaction between rs112481213 and rs3806703 , a linear mixed effect model , ( including gender , age , additive effect of the two SNPs and the interaction between the two SNPs effects ) implemented in the lmekin function ( R package ) was used . The variance inflation factor ( VIF ) was checked to be below ten to exclude collinearity problems [77] . For each replicated SNP , seed genes were selected as located within a region of 100 kb upstream and 100 kb downstream the region delimited by SNPs in LD ( r2>0 . 5 ) with it . For the main associated locus on chromosome 3 , CRIPTO gene was included as seed . Overall , 12seed genes were analyzed with Ingenuity Pathway Analysis software ( IPA , Ingenuity Systems , www . ingenuity . com ) to explore the functional relationship between the proteins encoded by those genes . IPA tests a set of genes for enrichment in defined canonical pathways or functions and generates de novo networks of interacting genes or gene products . IPA computes a p-value , based on a Fisher’s exact test , that represents the likelihood of the core genes in a network and biological function being found together due to random chance . Direct and indirect interactions , a high confidence ( experimentally observed or highly predicted ) and a maximum size of 35 genes/proteins per network were used as parameters in the analysis . The associated loci were investigated for presence of chromatin histone marks and hypersensitive DNAse elements using data from ENCODE included in Haploreg software ( http://www . broadinstitute . org/mammals/haploreg/ ) [78] . The replicated SNPs and variants in LD with them ( r2>0 . 5 ) were analyzed . Cell lines where the CRIPTO mRNA was reported to be expressed in ENCODE database ( www . genome . gov/encode/ ) were selected for the analysis . A 1051-bp fragment upstream the ATG of the CRIPTO gene ( position-1071/-20 , RefSeq Gene NG_017049 . 1 ) was amplified by PCR from genomic DNA of individuals homozygous for either the A allele or the T allele of the rs112481213 at position-222 using as primers 5’-CGACGCGTCAAGCGGCACATCAGAGTC-3’ and 5’-GAAGATCTGAAAAGAGGCGTTAGCATCG-3’ . The PCR products were digested with MluI/BglII and directionally cloned into the MluI and BglII sites of the luciferase reporter pGL3-basic vector ( Promega , Madison , WI ) to obtain-222A/luc and-222T/luc reporter constructs . The integrity of constructs was confirmed by DNA sequencing . Site-directed mutagenesis of both-222A/luc and-222T/luc constructs was performed using GeneArt Site-Directed Mutagenesis System ( Invitrogen ) according to the manufacturer’s protocol using as primers 5’-GAATCCCCGGAAAGGCTGAGTCACCAGCTCAAGGTCAAAACGTCC-3’ and 5’-GGACGTTTTGACCTTGAGCTGGTGACTCAGCCTTTCCGGGGATTC-3’ to perform the mutagenesis of-222T/luc ( -222T→A/luc ) and 5’-GAATCCCCGGAAAGGCTGAGTCTCCAGCTCAAGGTCAAAACGTCC-3’ and 5’-GGACGTTTTGACCTTGAGCTGGAGACTCAGCCTTTCCGGGGATTC-3’ for-222A/luc ( -222A→T/luc ) . The constructs were then sequenced to confirm the sequence changes . NTERA2 cells , cultured in Dulbecco’s modified Eagle’s medium F-12 ( Gibco-Invitrogen ) supplemented with 10% FBS at 37°C and 5% CO2 , were transiently transfected using JetPRIME transfection reagent ( PolyPlus Transfection ) following the manufacturer’s protocol . Briefly , 0 . 5 μg of either pGL3-Basic vector , -222T/luc , -222A/luc , -222T→A/luc , -222A→T/luc together with 10 ng of Renilla luciferase reporter plasmid ( Promega ) were cotransfected . Luciferase activity was assayed at 48h using the Dual-Luciferase Reporter Assay System ( Promega ) according to the manufacturer’s protocol . Measurement of the firefly luciferase activity was normalized relative to the activity of the Renilla luciferase . Each construct was tested in triplicate in at least 3 independent experiments . All reactions included double-stranded , biotin-labeled oligonucleotide probe at 40fmol concentration . EMSAs were performed by using the LightShift Chemiluminescent EMSA kit ( Pierce Biotechnology ) according to the manufacturer’s protocol . PC3 cell line was used because of high expression of AP-1 components [79] . Nuclear extracts were prepared using Subcellular Protein Fractionation Kit for Cultured Cells ( Pierce Biotechnology ) according to the manufacturer’s protocol . Nuclear extracts ( 5μg ) were incubated at room temperature either with the biotin-labeled probe alone or with the biotin-labeled probe and 50 or 200-fold molar excess unlabeled competitor probe for 20 min , before loading on a 4% nondenaturing acrylamide gel and subjected to autoradiography . The following double-stranded biotin-labeled oligonucleotides were used as probe: AP-1 control 5’-biotin-CGCTTGATGACTCAGCCGGAA-3’; -222T probe 5’-biotin-GAAAGGCTGAGTCTCCAGCTC-3’; -222A probe 5’-biotin-GAAAGGCTGAGTCACCAGCTC-3’ . Also , a scrambled oligonucleotide ( 5’-GAAAGGCTTGACGACCAGCTC-3’ ) was used for competition at 200-fold molar excess . Supershift assays were performed identically except for the addition of 3μg of antibody for 3h in ice before the addition of-222A probe . Antibodies used were anti-c-Jun , anti-JunB , anti-JunD , anti-Fra1 and anti-Fra2 [79] from Santa Cruz Biotechnology . HA-probe antibody against the influenza hemagglutinin ( HA ) protein ( Santa Cruz Biotechnology ) , was used as negative control .
|
Cripto gene has a fundamental role in embryo development and is also involved in cancer . The protein is bound to the cell membrane through an anchor , that can be cleaved , causing the secretion of the protein , in a still active form . In the adult , CRIPTO is detected at very low levels in normal tissues and in the blood , while its increase in both tissues and blood is associated to pathological conditions , mainly cancer . As other GPI linked proteins such as the carcinoembryonic antigen ( CEA ) , one of the most used tumor markers , CRIPTO is able to reach the bloodstream . Therefore , CRIPTO represents a new promising biomarker and potential therapeutic target , and blood CRIPTO levels might be associated to clinical features . Here we examined the variability of blood CRIPTO levels at a population level ( population isolates from the Cilento region in South Italy ) and we investigated the genetic architecture underlying this variability . We reported the association of common genetic variants with the levels of CRIPTO protein in the blood and we identified a main locus on chromosome 3 and additional five associated loci . Moreover , through functional analyses , we were able to uncover the mechanism responsible for the variation in CRIPTO levels , which is a regulation mediated by the transcriptional factor AP-1 .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
Genetic Variants Modulating CRIPTO Serum Levels Identified by Genome-Wide Association Study in Cilento Isolates
|
The recent discovery of a new class of 30-nucleotide long RNAs in mammalian testes , called PIWI-interacting RNA ( piRNA ) , with similarities to microRNAs and repeat-associated small interfering RNAs ( rasiRNAs ) , has raised puzzling questions regarding their biogenesis and function . We report a comparative analysis of currently available piRNA sequence data from the pachytene stage of mouse spermatogenesis that sheds light on their sequence diversity and mechanism of biogenesis . We conclude that ( i ) there are at least four times as many piRNAs in mouse testes than currently known; ( ii ) piRNAs , which originate from long precursor transcripts , are generated by quasi-random enzymatic processing that is guided by a weak sequence signature at the piRNA 5′ends resulting in a large number of distinct sequences; and ( iii ) many of the piRNA clusters contain inverted repeats segments capable of forming double-strand RNA fold-back segments that may initiate piRNA processing analogous to transposon silencing .
Five groups reported the discovery of small RNAs expressed exclusively in mammalian testes ( mouse , rat , and human ) that bind MIWI ( murine PIWI ) or MILI proteins [1–5] . Here , we focus on the three largest datasets ( A–C , listed in decreasing number of piRNA sequences identified in [1–3] ) each with thousands of distinct piRNA sequences ( a recent fourth comprehensive dataset of MILI-bound piRNAs identified in the pre-pachytene stage of spermatogenesis [6] is not included in this analysis ) . The number of unique piRNA sequences ranges from 3 , 482 to 40 , 102 ( Table S1 ) , as a result of the different methods used to identify the sequences . Overall , the length distributions of piRNAs peak at 29–31 nucleotides . However , the MILI-bound piRNAs ( dataset C ) [3] are generally shorter ( 26–28 nt ) than the MIWI-bound piRNAs ( 29–31 nt ) [1 , 2] , possibly due to differences in binding modes of the two proteins . The short length of piRNAs and the structural homology between PIWI and Argonaute proteins are suggestive of functional similarities between piRNAs and microRNAs . However , the combined evidence indicates that both the biogenesis and function of these two classes of RNA are distinct ( Table 1 ) . Primary differences are in genomic organization , sequence conservation , and in the number of unique sequences—among which are hundreds of microRNAs and tens of thousands of piRNAs . The majority of the identified piRNAs have a preference for a uridine base at the first position ( 78%–94% ) . Similar 5′ bias was observed in other types of small RNAs such as microRNAs and siRNAs , although to a lesser extent . The 5′ U is reminiscent of processing by RNase III enzymes [17 , 18] but may also reflect preferential binding to the Argonaute-like proteins . Although microRNAs and piRNAs share similar 5′ termini , other aspects of their biogenesis pathways are noticeably distinct: ( i ) piRNAs undergo 2′-O-methylation at their 3′ end [22–26] , which animal microRNAs do not; ( ii ) microRNA precursors are characterized by a distinct hairpin structure whereas piRNA precursors have no apparent secondary structure; and ( iii ) in contrast to microRNAs , piRNA maturation is independent of Dicer enzymes [16] . The majority of piRNAs ( 81%–96% ) is organized in clusters ( Figure S1 ) with distinct strand preference that ranges from 1 to 127 kb in size and are found predominantly in autosomes . Some of the clusters are organized in a bipartite arrangement with a stretch of piRNAs on one strand adjacent to a second stretch of piRNAs on the opposing strand . This organization is consistent with bi-directional transcription—for a minority of the clusters—from a common origin that generates two RNA precursors . The organization of piRNAs into clusters is common to mouse , human , and rat with significant conservation of the cluster genomic locations ( synteny ) [2 , 3] . In contrast , there is very little conservation at the level of individual piRNA sequences ( unpublished data and previously reported by [1–3 , 6] ) . Most reported piRNAs are in un-annotated intergenic regions and only a small fraction appears to be derived from mRNAs ( 5 . 7%–12% ) or is coincident with other classes of RNAs such as snoRNAs , tRNAs , rRNAs , or miRNAs ( 0 . 2%–3 . 5% ) [1–3] . piRNAs bind MILI and MIWI proteins , which are members of the PIWI protein family , a subclass of the Argonaute family . In eukaryotes , Argonaute proteins are key components of the interfering RNA pathway in which they bind mature microRNAs or siRNAs to form the RNA-induced silencing complex ( RISC ) [27] . All three murine PIWI members ( MIWI , MILI , and MIWI2 ) are required for spermatogenesis as determined by knockout experiments and are predominantly expressed in testes in partially overlapping time intervals [28–31] . Recent reports link mammalian MIWI protein to chromatoid bodies ( also known as nuages in Drosophila ) [32] . These are cytoplasmic structures found in all mammalian spermatogenic cells that physically associate with the nuclear membrane during spermatogenesis and contain an RNA helicase protein ( VASA ) . The function of chromatoid bodies is unknown but they are presumed to be the site of post-transcriptional processing and storage of mRNAs analogous to processing bodies in somatic cells ( P-bodies ) [33] . It is unknown if the co-localization of MIWI proteins to chromatoid bodies is linked in any way to their function with piRNAs . rasiRNAs are a class of interfering RNA with a size distribution of 23–28 nucleotides that were identified in a number of organisms [17] . They originate from repeat sequences related to transposable elements and heterochromatic regions [15] , and evidence supports their involvement in transposon silencing [13–21] . rasiRNAs are found in both female and male germline where they bind members of the PIWI family ( Piwi , Aub , and Ago3 in Drosophila ) [13 , 20 , 34] . There are two distinct types of Drosophila rasiRNAs ( there is evidence that similar classes exist in Zebrafish [21] ) ; the first type bind Piwi or Aub proteins , are mostly antisense to transposable elements , and enriched for 5′ uridine . The second type bind Ago3 proteins , are mostly sense to the transposable elements , and enriched in adenosine at position 10 . The different strand-specificity and the U and A enrichments led to the hypothesis that the biogenesis of the two types of rasiRNAs is coupled [13 , 20] . In this model the Piwi/Aub-associated rasiRNAs guide the 5′ cleavage of the Ago3-associated rasiRNAs by hybridization to the sense transcript . Similarly , the Ago3-bound rasiRNAs direct the 5′ cleavage of the Piwi/Aub-bound rasiRNAs by hybridization to the anti-sense transcripts . Thus , the two rasiRNA types are engaged in a mutual amplification loop that facilitates the silencing of multiple transposon copies . The length characteristics , testis-specific expression , PIWI interaction , genomic organization , and 5′ uridine enrichment suggest that piRNAs may be the mammalian equivalent of rasiRNAs . This would support the idea that mammalian piRNAs might be involved in silencing transposable elements . However , at present , there are a number of differences that cast doubt on this functional analogy . First , genomic annotation of piRNAs indicates that only 12%–20% are repeat derived [1–3 , 5] , which is smaller than the frequency of repeat sequences in the mouse genome ( 37 . 5% ) [35] , while Drosophila rasiRNAs originate preferentially from repeat regions . Second , mammalian piRNAs originate from one strand or the other forming clusters with continuous strand bias whereas rasiRNAs originate from both strands of the clusters with positional enrichment for “U” and “A . ” We explored the analogy between rasiRNAs and piRNAs , but did not find significant 5′ partial complementarity between piRNA sequences as found in rasiRNAs [13 , 20] . However , at present , sequences associated with the third mouse testes–specific MIWI protein ( MIWI2 ) , also essential for spermatogenesis and linked to transposon silencing [31] , have not yet been identified . Future identification of MIWI2-bound piRNAs—in analogy to Ago3-bound Drosophila rasiRNAs—enriched for adenosine at position 10 with partial complementary match to other piRNAs would be strongly suggestive of functional similarity between rasiRNAs and piRNAs . The discovery of large sets of piRNAs raises a number of important biological questions . In particular , what is the biochemical role and cellular function of PIWI-bound piRNAs during spermatogenesis ? Are they involved in transposon silencing , chromosome rearrangements ( as are 30-nt PIWI-bound RNAs in Tetrahymena [36 , 37] ) , or chromosome pairing ? What are the evolutionary constraints on piRNA sequences ? Answers to these questions will primarily emerge from further experiments . Here , we focused on the basic questions of how many piRNA sequences there are and how they are produced . We reasoned that a detailed computational comparison of the three major datasets , representing independent discoveries of piRNAs , provides insight into the organization of genomic clusters , the number and distribution of sequences within the clusters , and , by implication , their biogenesis .
We first compared the cluster locations in the mouse genome from datasets A–C and found extensive agreement between the datasets . The majority of clusters overlap by more than 75% of the length of the shorter cluster . All 42 genomics clusters from dataset C , the smallest of the three , matched clusters of datasets A and B ( Figure 1B ) . Given the different definitions of clusters in the three datasets , we conclude that the three sets of experiments have determined essentially the same clusters of piRNAs expressed in the pachytene stages of spermatogenesis ( Figure S1 ) . Other stages of development may yield additional and possibly distinct sets of piRNAs , such as the MILI-bound set of piRNAs ( not analyzed here ) recently identified in the pre-pachytene phase of spermatogenesis [6] . We compared the sets of individual sequences from the three groups ( A–C ) . Contrary to the agreement between clusters , we found surprisingly small overlaps between the sets of unique sequences , irrespective of the criteria used for sequence comparison ( 100% , 95% , or 90% sequence identity , Table S1 ) . For example , at a 95% sequence identity cutoff only 45% of the sequences from dataset C overlap with dataset A ( the largest fractional overlap among all pairs of datasets ) , although all the piRNA clusters from the smallest dataset C are included in dataset A . Furthermore , only 587 sequences were common to all three datasets representing 20% , 3 . 7% , and 2 . 7% of the datasets C , B , and A , respectively ( Figure 1C ) . Similarly low overlap was observed when comparing human piRNA datasets , but as the sequencing coverage is lower than in mouse , this result is not as conclusive . This small overlap between the piRNA datasets points to an apparent contradiction—how can different sets of piRNA sequences originate from a common set of genomic clusters ? The simplest explanation is that each experiment identified only a subset of sequences from a larger pool of unique piRNA sequences . To quantify this effect , we first asked whether the observed overlaps are within the expected range assuming that the complete piRNA pool is simply the union of the three datasets . To facilitate the comparison we restricted this analysis to the intersection of clusters from the three datasets , termed “intersection clusters” ( Table S2 ) . By numerical simulation and direct calculation we find that the observed sequence overlaps between all datasets is significantly lower than expected ( unpublished data ) , indicating that the total pool of piRNA sequences is indeed larger than the simple union of current datasets . Using straightforward statistical calculation , we then estimated the total number of piRNAs from the observed overlaps in the intersection clusters by considering the three studies as independent sampling experiments from a common pool of all piRNAs ( Figure S2 ) . From this estimate we conclude that the current datasets analyzed here have so far identified only 25%–30% of all potential piRNA sequences from the pachytene stage of mouse spermatogenesis . This implies that in the complete set ∼20%–25% of all “U” positions in the clusters are potential start sites for piRNA sequences when taking into account the pronounced preference for 5′ uridine . Extrapolating to saturation in all clusters reported by any of the three groups , we arrive at the overall conservative estimate of Ntotal ≈ 2 × 105 potential piRNA sequences in mouse testes ( Figure 2 ) . This does not imply that all sequences are necessarily present in any given cell . The details of piRNA biogenesis are not yet known . In particular , what is the precursor form of piRNAs ? Is it single-strand or double-strand ? What are the components of the nuclease-processing complex ? By which mechanism , in which order , and under which regulatory control do thousands of different ∼30 nt transcripts originate from a limited number of genomic regions ? The large differences in piRNA datasets and the relatively weak evolutionary conservation of piRNA sequences suggest that the processing of piRNAs from a primary precursor is not a precise step , in contrast to microRNA maturation . Instead , it appears , to a first approximation , that piRNAs are generated by a random mechanism in which any U position is a potential 5′ piRNA start . This notion is supported by the fact that sequence overlap between the datasets remains low even when we compare only the more abundant sequences ( Figure S3 ) , and that there is no evidence for repetitive spacing between consecutive sequences ( unpublished data ) . However , there appears to be some non-randomness in that some positions are preferentially processed into piRNAs ( see patterns in Figure 1A , panels 4 and 5 ) . In particular , a sizable fraction ( ∼20% ) of all piRNA sequences were cloned three or more times , and we find that many piRNA sequences from the same strand are partially overlapping ( Figure S4 ) . This suggests some , albeit weak , sequence effects within a genomic cluster , either at the level of nuclease processing or at the level of loading into a PIWI complex . We use the term “quasi-random” to reflect this weak departure from random processing . We therefore attempted to identify a distinguishing sequence signal that predicts which U bases are 5′ piRNA cleavage sites . Using a sequence classification algorithm , we identified , with 61% accuracy , the correct 5′ U piRNA sites from all other U positions using both 10-fold cross-validation on the training set and by testing on randomly withheld test set excluded from training ( see Methods ) . Although the classification accuracy is low , it is significantly better than random prediction ( classification on randomized data did not exceed 50% ) . Furthermore , the classification accuracy improved to 72% when the algorithm was trained and tested on the abundant piRNA sequences ( clone counts >2 ) . The differentiating signal is a weak preference for a G or A in the +1 position ( relative to the 5′ U ) , an A in the +4 position , and a slight under-representation of G at the −1 position ( Figure 3 ) . These results suggest that the processing of the precursor is quasi-random in that there is a weak yet significant non-random sequence preference at the 5′ cleavage site . The precursor form of piRNA primary transcript— single- or double-stranded—is currently unknown . However , the strong 5′ uridine bias and the presence of the 5′ phosphate group [4] in mature piRNAs is indicative of a dsRNA precursor that is processed by an RNase III type enzyme [3] , although no such nuclease has so far been implicated in piRNA processing , and piRNA processing is independent of Dicer [9] . In Caenorhabditis elegans , germline silencing of transposable elements by the RNAi pathway is initiated by a dsRNA structure formed by base pairing of the terminal inverted repeats of the transposon in a fold-back structure [38] . To investigate whether a similar mechanism may be involved in piRNA biogenesis , we searched for inverted repeats in or near the vicinity of piRNA clusters . Such inverted repeats may form precursors containing dsRNA that initiate enzymatic processing . Overall , we found that 63% of all clusters have inverted repeats of length 100 bases or longer ( see Methods ) and that 25% of all clusters are bracketed by inverted repeats , i . e . , the complementary segments are at the ends of the clusters ( Figure 3B ) . Surprisingly , some of the flanking inverted repeats coincide with inverted transposable elements such as SINEs , LINEs , and LTRs that are on opposite strands , one on each side of the cluster ( Figures 3B and S5 ) , suggesting a link between transposable elements and piRNA biogenesis . Recent studies propose that mammalian piRNAs may be involved in transposon silencing analogous to Drosophila rasiRNAs , although the mechanistic details remain to be determined [6 , 13] . The model of transposon silencing by rasiRNAs put forward by [13 , 20] explains the feed-forward amplification of the silencing process but not its initiation . They propose that the induction requires a pool of initiating rasiRNAs that triggers a mutual amplification loop between the Ago3-bound and the Piwi/Aub-bound rasiRNAs . The source of the initiating rasiRNAs is unknown , and they may be maternally inherited by the developing oocytes . We hypothesize that one plausible model of piRNA biogenesis involves long transcripts that contain flanking inverted transposable elements , one at each end of the cluster ( Figure 3B ) . Such precursors can arise , for example , by continuous transcription of one of the repeats past its termination site . If the transcript reaches the other end of the cluster and includes the sequence complementary to the repeat element on the opposing strand , the transcript can potentially form a dsRNA segment . piRNA biogenesis is then triggered by processing of the dsRNA segments which generate the initiating pool of piRNAs . Similar to the Drosophila model of rasiRNA generation [13 , 20] , these initial sequences may act on transcripts derived from other locations ( in trans ) containing at least one copy of the initiating repeat element and resulting in the production of a much larger pool of piRNAs . We cannot exclude the possibility that the bracketing inverted transposable elements are not part of the primary transcript but simply the result of statistical coincidence . In fact , similar numbers of such repeats are found in randomly chosen genomic regions ( unpublished data ) , as remnants of transposable elements account for over a third of the mouse genome [35] , but most of these may not be expressed . In contrast , the bracketing inverted repeat structures must be transcriptionally active , and we do find that a number of the transposable elements near piRNA clusters are indeed expressed in testes ( as indicated by ESTs recorded in genome databases ) . Alternatively to the initiating dsRNA structure , a single-strand RNA precursor may be a direct substrate of a nuclease , yet to be discovered , that generates approximately 30-residue long 5′ P products . The novel discovery of piRNAs has extended the multifaceted family of small interfering RNAs that includes microRNAs , siRNAs , and rasiRNAs . The tens of thousands of distinct mouse piRNAs observed so far map to ∼117 distinct genomics locations in the genome . The details of piRNA transcriptional control , such as promoter sites and transcription factors , remain to be determined . Our analysis has revealed low sequence overlap between the currently known pachytene-stage mouse piRNA datasets , although the sequences originate from a common set of genomic clusters . This apparent contradiction is resolved by noting lack of saturation in each individual experiment . We interpret the low sequence overlap as suggestive of quasi-random sub-saturation processing from common precursors , such that different experiments yield different and only partially overlapping sets of piRNAs . In addition , based on the observation of repeat structures bracketing some of the clusters , we propose that one plausible mechanism for initiation of piRNA biogenesis involves long transcripts with terminal inverted repeats , possibly derived from ( remnants of ) transposable elements . Such transcripts may form partial dsRNA intermediates initiating enzymatic degradation . Subsequent stages of piRNA biogenesis may then follow the ping-pong model proposed by [13 , 20] . The notion that piRNAs both direct the degradation and are the degradation products of their own precursors suggests that piRNA transcripts are under strict regulation at a crucial stage of meiosis . What is their function ? The PIWI proteins are highly expressed in the pre-pachytene and pachytene stages of meiosis when chromosome pairing is completed ( zygotene ) and synapsis is peaked . This raises the intriguing possibility that the transcripts from which the piRNAs derive , and/or the piRNAs themselves , are involved in one of the crucial processes of meiosis , correct chromosome pairing , for which the molecular mechanism remains a mystery . The connection between this and the proposed piRNA function of transposon silencing remains to be elucidated . We look forward to directed biochemical and genomic experiments that will invalidate or confirm the models proposed here and explain the function of piRNAs .
Mouse piRNA sequences were collected from the following sources: Dataset A from Lau et al . [1]; Table S4 contains 65 , 535 unique small RNA sequences . After removing known small RNA sequences , the remaining 40 , 102 were considered as piRNA sequences for this dataset . Dataset B from Girard et al . [2] ( personal communication ) includes 51 , 331 reads representing 28 , 956 unique sequences . Dataset C from Aravin et al . [3] ( Table 4 therein ) contains 5 , 444 small RNA sequences of which 3 , 482 are unique sequences annotated as piRNAs . Dataset D from Watanabe et al . [4] ( Table S7 therein ) contains 357 unique small RNA sequences . Dataset E from Grivna et al . [5] ( Table S1 therein ) contains 40 unique sequences . Duplicate and subsequences were removed from each dataset at 100% nucleotide identity ( Table 1 ) . In cases where genomic annotation was provided we removed known small RNAs ( miRNAs , tRNAs , and snoRNAs ) as well as apparent rRNA and mRNA fragments from the dataset . All sequences and clusters were mapped to mouse genome build mm7 ( August 2005 ) taking the best genomic match up to a maximum of two mismatches or gaps . Sequence matching to the genome was performed using a combination of WU-BLAST ( http://blast . wustl . edu/ ) and our in-house alignment software developed jointly with M . Zavolan . The following BLAST arguments were used for short sequence alignments: −W = 6 − X = 50 − gapX = 50 − S2 = 50 − gapS2 = 50 − hspmax = 1 , 000 − gspmax = 1 , 000 − E = 1 , 000 − filter = none . Over 90% of the sequences mapped to unique genomic locations . In the remaining cases where there was more than one match to the genome , all positions were considered as a possible origin of the piRNA . Coordinates of piRNA clusters from dataset C were translated from mm6 to mm7 , in some cases resulting in a change in cluster length due to partial mapping: The datasets were not significantly biased to specific sequences or nucleotide composition by experimental protocol . The two larger datasets ( A and B ) were produced using similar ligation adaptors and sequencing methods excluding the possibility of sequence bias due to different methodologies . Indeed , we found no differences in mononucleotide or dinucleotide frequencies between the datasets . Overlaps between genomic clusters from different datasets were determined by intersection of their genomic locations . The length of the overlaps ranged from 19% to 100% of the shorter cluster . In the majority ( 70% ) of the overlapping clusters , the extent of the overlap covered >75% of the length of the shorter cluster . Instances where two clusters from one dataset overlapped a single cluster from another dataset were counted as one overlap . Intersection clusters were defined as the genomic regions where clusters from all three datasets overlapped ( See Table S2 ) . Sequence comparison was performed as follows: All sequences ( after initial processing ) from all datasets were combined and compared all-against-all using WU-BLAST and in-house software . Sequences were grouped into similarity sets by hierarchical clustering and a defined identity measure . To explore sensitivity of the analysis to variation in parameters , we performed three clustering procedures using these identity measures: ( i ) 100% sequence identity over the entire length of the shortest sequence; ( ii ) 95% sequence identity over 95% length of the shortest sequence; and ( iii ) 90% sequence identity over 90% length of the shortest sequence . Considering all sequences in a similarity cluster to be essentially identical , the degree of overlap between two datasets is determined by counting the number of similarity clusters that contain sequences from both datasets ( Table S1 ) . Similarly , the three-way overlap between datasets A , B , and C was determined by counting the number of similarity clusters that contained sequences from all three groups ( Figure 1C ) . The comparison of the abundant piRNA sequences ( higher clone counts ) was performed in the same way using only sequences that were cloned >2 times ( Figure S3 ) . Human piRNA sequences were retrieved from Girard et al . ( dataset B ) and Aravin et al . ( dataset C ) studies . Similarly to mouse piRNAs , sequences that matched known small RNAs and mRNAs were removed resulting in 9 , 600 unique piRNA sequences from dataset B and 120 sequences from dataset C . Sequences comparison was performed as outlined above . Under 95% sequence identity measure , 29 sequences were shared between the two datasets corresponding to ∼24% of dataset C sequences . The degree of overlap between two independent datasets , say X and Y , in a genomic intersection cluster is modeled by a hypergeometric distribution with a mean where nx and ny are the number of piRNA sequences in the cluster in datasets X and Y , respectively , and N is the total number of piRNAs in the cluster , which is unknown . This corresponds to random selection of nx and ny piRNA sequences from a total pool of N unique sequences , i . e . , ignoring varying clone counts . Under the maximum likelihood assumption , the observed overlap between the two datasets is the most likely value . That is , where nx ∩ y is the size of the observed overlap between datasets X and Y . For the purpose of this approximation , the size of the overlap nx ∩ y was determined by a 95% sequence identity criterion ( see above ) . The value of total number of piRNAs N can then be computed directly as: For each intersection cluster we computed three estimates of N: NAB , NAC , NBC based on the observed overlaps between datasets AB , AC , and BC ( Figure S2 ) . The total number of piRNAs was computed as the average of the three approximations summed over all clusters: where i is an intersection cluster , c is the set of all intersection clusters , is the computed total number of piRNAs in cluster i based on the overlap between datasets A and B , and similarly for and . To approximate the total number of piRNAs in the mouse genome we extrapolated the total number in all intersection clusters , to the union of all clusters from datasets A , B , and C ( Table S3 ) , by multiplying NTotal by the ratio of the combined length of the union of all clusters to the combined length of all intersection clusters ( Figure S2 ) . Sequences assigned to genomic positions were sorted by chromosomal position . The distance between two adjacent sequences i , j mapped to the same strand is determined by: When i and j are overlapping di , j ≤ 0 . To identify a distinguishing signal for 5′ piRNA processing in cluster regions , we trained a support vector machine classifier to discriminate between 5′ piRNA and all other uridine positions . Positive set included all of the piRNAs 5′ uridine positions extended ten bases upstream and downstream; a total of 24 , 604 sequences . Similarly , the negative set was constructed by selecting random non-piRNA uridine positions in the intersection clusters and ten nucleotides upstream and downstream . Both sets were split into two , one part used for training and the other for testing . Feature vectors were constructed by converting the 21-base sequences into 84-bit vectors ( 21 nt × 4 bases ) , i . e . , each nucleotide position is converted to a 4-bit vector representing the RNA base . Support vector machine training and classification was performed using an R interface of “libsvm” ( http://cran . r-project . org/src/contrib/Descriptions/e1071 . html ) using a polynomial kernel of degree 3 . Classification accuracy in a 10-fold cross-validation on the training set and testing procedure on an independent test set was ∼61% , whereas classification using a randomized training set did not exceed 50% accuracy . Using the high frequency piRNAs ( cloned >2 times ) as the positive training set , the prediction accuracy in 10-fold cross-validation and with the test set improves to 72% . In a feature selection process we found that positions −1 , +1 , and +4 ( relative to the starting uridine position 0 ) were the largest contributors to the classification ( Figure 3D ) . Information content analysis revealed a preference for G or A in positions +1 , for an A in positions +4 , and under-representation of G at position −1 . For detection of inverted repeats in the vicinity of cluster , sequences were collected from the union clusters ( Table S3 ) and extended by 10 kb in both 5′ and 3′ directions . The sequences were aligned to their complements by “bl2seq” ( a BLAST implementation for aligning two sequences ) in gapless mode ( using –g F flag ) . Alignments longer than 100 bases with >90% identity were mapped to the mouse genome and used in subsequent analysis .
|
The discovery of a new class of mammalian small regulatory RNAs termed PIWI-interacting RNA ( piRNA ) has extended the diverse family of small regulatory RNAs . PIWI proteins are a subclass of the larger Argonaute proteins family , of which the Ago members bind microRNAs and play a critical role in gene silencing . Despite the homology between PIWI and Ago proteins , piRNAs are strikingly different from microRNAs in their length , expression pattern , and genomic organization . In contrast , piRNAs are similar to repeat-associated small interfering RNA ( rasiRNAs ) , a class of small RNAs that are responsible for transposon silencing in Drosophila germline , although it is unclear if piRNAs function in a similar way . This paper describes a computational comparison and analysis of the existing comprehensive piRNA datasets identified independently by three groups at the pachytene stage in mouse spermatogenesis . We find that the studies have identified similar genomic piRNA clusters , but differ substantially in the piRNAs that were cloned from those clusters . Based on these results we quantify the expected number of piRNAs and suggest that the processing of piRNAs from genomic transcripts is quasi-random . We find that a weak sequence signature may guide the piRNA 5′end processing that accounts for the departure from fully random processing . We further show partial evidence that piRNA biogenesis may be initiated by neighboring transposable elements .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] |
[
"mus",
"(mouse)",
"homo",
"(human)",
"genetics",
"and",
"genomics",
"computational",
"biology"
] |
2007
|
Computational Analysis of Mouse piRNA Sequence and Biogenesis
|
Influenza is a major cause of morbidity and mortality in immunosuppressed persons , and vaccination often confers insufficient protection . IL-28B , a member of the interferon ( IFN ) -λ family , has variable expression due to single nucleotide polymorphisms ( SNPs ) . While type-I IFNs are well known to modulate adaptive immunity , the impact of IL-28B on B- and T-cell vaccine responses is unclear . Here we demonstrate that the presence of the IL-28B TG/GG genotype ( rs8099917 , minor-allele ) was associated with increased seroconversion following influenza vaccination ( OR 1 . 99 p = 0 . 038 ) . Also , influenza A ( H1N1 ) -stimulated T- and B-cells from minor-allele carriers showed increased IL-4 production ( 4-fold ) and HLA-DR expression , respectively . In vitro , recombinant IL-28B increased Th1-cytokines ( e . g . IFN-γ ) , and suppressed Th2-cytokines ( e . g . IL-4 , IL-5 , and IL-13 ) , H1N1-stimulated B-cell proliferation ( reduced 70% ) , and IgG-production ( reduced>70% ) . Since IL-28B inhibited B-cell responses , we designed antagonistic peptides to block the IL-28 receptor α-subunit ( IL28RA ) . In vitro , these peptides significantly suppressed binding of IFN-λs to IL28RA , increased H1N1-stimulated B-cell activation and IgG-production in samples from healthy volunteers ( 2-fold ) and from transplant patients previously unresponsive to vaccination ( 1 . 4-fold ) . Together , these findings identify IL-28B as a key regulator of the Th1/Th2 balance during influenza vaccination . Blockade of IL28RA offers a novel strategy to augment vaccine responses .
Generation of a protective and durable immune response is the major challenge of effective vaccinations against influenza . On a global scale , infection with influenza viruses is associated with increased morbidity and mortality in elderly persons , pregnant women and immunosuppressed individuals [1] . The primary means to limit this disease is through annual influenza vaccination as recommended [2] . However , annual influenza vaccines are poorly effective in the elderly , and immunocompromised populations [3]–[5] . For example , after organ transplantation , post-vaccine seroconversion rates only approach 30 to 50% [4] , [6] , [7] . Although this may be a function of diminished adaptive immune responses , there are increasing data that interferons ( IFN ) may modulate vaccine responses [8]–[12] . Understanding the factors involved in a successful vaccine response and seroconversion will allow optimization of vaccine strategies [13] . The IFN-λ family ( Interleukin-28A , -28B , -29 , and IFN-λ4 ) is a recently described class of IFNs with antiviral properties similar to IFN-α and -β [14]–[17] . IFN-λ is known to induce phosphorylation of STAT-1 and -2 via binding to its receptor , which is a heterodimer consisting of the IL-28 receptor alpha subunit ( IL28RA ) and IL-10 receptor beta subunit ( IL10RB ) [16] . In addition to their anti-viral effects , one of the IFN-λ family members ( IL-29 ) has been shown to increase Th1 and suppress Th2 cytokine producing T-cells [18]–[21] . Furthermore , IFN-λs induced the development of T-regulatory cells in vitro [22] , [23] . These findings indicate the substantial role of IFN-λs in immune responses , however , this has not been explored in the context of vaccine responses or influenza infection . Single nucleotide polymorphisms ( SNPs ) in IL-28B are divided according to their frequencies in a population . At rs8099917 , TT is the major-allele and TG or GG are minor-allele genotypes; at rs12979860 , CC is the major-allele and CT or TT are minor-allele genotypes [24]–[27] . We selected these two SNPs as they are commonly described in the literature to affect IL-28B functions . Since IFN expression is involved in multiple aspects of the immune response , we hypothesized that the effectiveness of vaccination may be modulated by variation in IL-28B expression as a consequence of SNPs . We further explored the possibility that altered expression of IL-28B might be associated with changes in B- and T-cell responses . In this study , we chose to use clinical samples obtained from organ transplant patients . These patients receive lifelong immunosuppression and have impaired adaptive immune responses to vaccination . Therefore , any impairment of the innate immune response which alters stimulation of adaptive immunity , is likely to take on greater importance . This population also stands to have the greatest gain from strategies to augment vaccine responses . Here we show that transplant patients that carry minor-alleles in the IL-28B ( rs8099917 , TG or GG ) gene have significantly higher rates of seroconversion following influenza vaccination . PBMCs from transplant patients with the minor-allele expressed less Th1 cytokines , had more IL-4 producing H1N1-specific T-cells , and higher HLA-DR activation marker expression on naive B-cells than those from major-allele carriers . Consistent with these findings , in vitro addition of IL-28B to PBMCs increased Th1 cytokine expression , decreased Th2 cytokines , and decreased H1N1 stimulated B-cell proliferation and IgG production . We also show healthy volunteer PBMCs from minor-allele carriers stimulated with H1N1 expressed less IL-28B . Antagonistic peptides designed to block the interaction between IL-28B and its receptor , reversed these effects and could potentially be used as a novel class of vaccine adjuvants .
We have determined in an immunocompromised transplant population that the presence of the rs8099917 single nucleotide polymorphism ( SNP; TG or GG ) in the IL-28B gene significantly increases the likelihood of seroconversion to an influenza vaccine especially in those people on high doses of immunosuppression . We also showed that IL-28B affects Th2 and B-cell responses in the context of influenza stimulation . Other important factors associated with B-cell functions such as T-follicular helper cells and IL-21 [40] were not studied . IL-28B ( IFN-λ3 ) belongs to the family of IFN-lambdas and shares antiviral properties similar to IFN-α via induction of interferon stimulated genes ( ISGs ) such as MX1 or OAS1 [41] . In addition , IL-28B has been shown to induce IL-12 production in monocytes and macrophages . IL-12 is a key cytokine for the induction of Th1 cells and cytotoxic lymphocytes [42] , [43] . SNPs in the IL-28B gene ( minor-allele genotypes ) have been associated with reduced IL-28B expression [25] , [29]–[33] , which could impact adaptive immune functions during vaccination . One limitation of our study is that we did not measure serum levels of IL-28B . To the best of our knowledge , no reliable ELISA assay is currently available , which can differentiate between the high sequence homology of IL-28A and IL-28B [44] . In addition , since our study only evaluates patients 30 days post-vaccination , we may not capture the peak of IL-28B secretion . Approximately 40% of Caucasians and 10% of Asian populations carry IL-28B minor-allele genotypes [27] . SNPs in IL-28B have been best studied in the context of response to Hepatitis C therapy . Minor-allele genotypes in IFN-λ signalling have been associated with reduced sustained virologic response of hepatitis C virus ( HCV ) following IFN-α treatment [24]–[27] . In contrast patients with a IL-28B minor-allele genotype and also at high-risk for primary infection with Cytomegalovirus , had lower frequencies and shorter episodes of primary CMV replication [29] , [45] . Patients with minor-allele genotypes of IL-28B showed lower expression of IFN-λ during HCV infection in liver biopsies [25] , [30] and during stimulation of PBMCs with CMV [29] . Previous studies have shown that vaccine responses may be influenced by SNPs in interleukin genes . For example , hepatitis B vaccine responses may be influenced by SNPs in the IL-4 gene [46] . SNPs in interleukin genes may also affect humoral and cellular responses to the measles vaccine [47] . In a large cohort of children vaccinated against measles , the rs10853727 SNP in the IL-28B promoter was strongly associated with post-vaccine titers . The major-allele genotype ( AA ) showed significantly lower measles antibody titers compared to the minor-allele genotypes ( AG and GG; median 807 vs . 1004 , and 1727 mIU/mL , respectively; p = 0 . 021 ) [47] . The effects of SNPs on vaccine responses in the general population may be demonstrated through large-scale genome-wide association studies . However , an immunosuppressed cohort with poor adaptive immunity can be ideal to demonstrate immunogenetic differences in vaccine responses . We found that the association of the IL-28B SNP with influenza vaccine seroconversion and seroprotection ( to at least two vaccine antigens ) was even more significant in transplant patients on high doses of mycophenolate mofetil ( MMF ) . MMF is well known to significantly suppress influenza vaccine responses [28] , [48] by having an effect on virus-specific Th2 cytokines [49] and on B-cell activation markers , and seroconversion rates [28] . The minor-allele genotype in patients treated with more than two grams MMF per day demonstrated significantly higher seroconversion rates – essentially acting similar to a “rescue” mutation . We speculate that major-allele carrier status with high IL-28B expression in addition to receiving high-dose MMF therapy leads to a “double hit phenomenon” on Th2 responses . As a limitation of our work , we recruited healthy volunteers over multiple months , therefore the numbers within various experiments are variable and some intra-individual variation may be present . We also determined that the IL-28B rs8099917 SNP affected not only humoral responses to the influenza vaccine but also had a potent effect on cellular responses by modulating the Th1/Th2 cytokine balance . We show that the IL-28B minor-allele genotype is characterized by a predominant Th2-response upon stimulation with H1N1-influenza virus , and is associated with increased B-cell activation ( HLA-DR , CD86 ) and function ( IgG production ) . Although we did not measure H1N1-specific IgG in transplant patients , we do show in healthy volunteers that virus-specific IgG decreases upon pre-treatment with IL-28B in vitro at similar inhibition levels . Furthermore , exogenous treatment with IL-28B simulated a major-allele phenotype with significantly reduced Th2 cytokine expression . In PBMCs from healthy volunteers , this phenomenon was independent of MMF treatment . Our findings confirm the previous observation that IL-29 may skew the balance of Th1 and Th2 cytokine towards Th1 and a pronounced cytotoxic T-cell response [18] , [19] , [21] , [38] . Secretion of Th1-cytokines acts as an important suppressor of Th2-cell differentiation [50] , [51] via IFN response factors [51] and is associated with lower antibody titers after influenza vaccination [52] . Interestingly , the effect of IL-28B treatment was stronger in minor allele genotypes . The reduced effects in major allele genotypes could be due to a higher baseline expression of IL-28B and saturation of the signalling cascade . This is supported by a study in hepatocytes , where the minor allele genotype was associated with a higher baseline IL28RA expression and increased susceptibility and responses to IFN-λs [53] . A similar mechanism could be present in antigen presenting cells , which in turn has then affects the down-stream effects on T-cells and antibody production . We further used peptides to inhibit IL-28R signalling . These peptides have previously been described [29] . Inhibition of the IL-28B signalling during vaccination offers the potential to enhance Th2 cytokine release and thereby boost pathogen-specific IgG . It has been previously shown that signalling of the IFN pathway suppresses IgG secretion via increasing Th1 cytokines and a more cytotoxic immune response [52] . In particular antagonistic peptides 1 , 6 , 16 and 17 are promising candidates with high binding affinities to IL28RA , a strong potential to inhibit binding of IFN-λs and the ability to significantly increase in vitro H1N1-induced IgG production . These antagonistic peptides may enable immunomodulation towards Th2 cytokines and have the potential to become a new class of adjuvants by modulating IFN . An important strength of our study is the use of a clinical cohort including immunosuppressed transplant recipients to confirm our findings in the clinical setting . We then sought to define additional observations to support our clinical findings making our study unique . One limitation of our study is the heterogeneity of the transplant cohort due to a variety of underlying conditions leading to organ failure . However , the immunosuppressive treatment was not significantly different between genotype groups . In addition , we have shown that mycophenolate mofetil ( MMF ) and the IL-28B major-allele genotype are independent factors for IgG production and that IFN-λ mRNA expression is not influenced by MMF . Another limitation of our study was that at the time of peptide design , only the crystal structure of IL-29 was available and the peptides are therefore based on IL-29 and not IL-28B . However , as IL-29 has a significantly greater binding affinity towards IL28RA compared to IL-28B , this could be advantageous as we are potentially blocking all IFN-λs with greater efficiency . The role of the IL10RB co-recruitment also needs to be further defined . In summary , SNPs in IL-28B play a key role in vaccine responses especially for influenza vaccine response in immunosuppressed patients . Peptides used to inhibit IFN lambda receptor signalling may play a role in augmenting vaccine responses and as such , represents a novel avenue for developing new adjuvants . Further studies in other populations such as other immunosuppressed populations , elderly persons and healthy individuals would also lead to improved vaccine strategies .
A previously described cohort of immunosuppressed adult solid organ transplant recipients was used for this study [28] . Healthy non-immunosuppressed non-vaccinated volunteers ( HV ) were recruited as controls . Peripheral blood mononuclear cells ( PBMCs ) from 47 transplant recipients were available . The study protocols were approved through the University of Alberta research ethics board and written informed consent was obtained from all participants ( patients and healthy volunteers ) . HAI titers were determined as previously published [28] . Definitions of vaccine immunogenicity were based on recommendations for annual licensure of influenza vaccine ( European Medicines Agency , document: CHMP/VWP/164653/2005 ) . Seroconversion was defined as a ≥4-fold rise in titer from pre-vaccination . Vaccine response was defined as seroconversion to at least one of the three vaccine antigens: influenza A/California/7/2009 ( H1N1-like ) , A/Victoria/210/2009 ( H3N2-like ) and B/Brisbane/60/2008 [28] . SNP genotypes were determined as previously published [29] , [54] , [55] . Briefly , the probe set to discriminate the rs12979860 discriminates the C and T alleles , where C is the major , and T is the minor-allele [55] . For the rs8099917 SNP , the probe set discriminates the T and G alleles , where T predicts the major , and G is the minor-allele . SNP detection was performed on 6 ng of genomic DNA . S9 Table shows all primer sequences . In each allelic discrimination assay 50 bp synthetic positive control oligonucleotides were included . SNP genotype was determined using the automatic call algorithm in conjunction with the allelic discrimination plot . For immune stimulation we used formalin inactivated , partially purified A/California/7/2009 ( H1N1 ) ( NIBSC , NXMC-X179A , UK ) . The H1N1 stock contained 50 µg/mL of hemagglutinin protein and was re-constituted in water . For all experiments a final concentration of 0 . 3 µg/mL was used . IL-28B primers and probe were designed based on Homo sapiens IL-28B mRNA-sequence ( NM_172139 . 2 ) using Primer3 Input ( version 0 . 4 . 0 ) ( http://frodo . wi . mit . edu/ ) . The forward primer ( IL-28BF1: CAAAGATGCCTTAGAAGAGTCG ) spans the exon/exon junction of exons 1 and 2 of IL-28B . The IL-28B-specific probe ( IL-28B probe: GCTGAAGGACTGCAAGTGCCG ) is located in the second exon and the reverse primer ( IL-28BR1: TCCAGAACCTTCAGCGTCAG ) is in the third exon of the IL-28B gene . For IL-28A , the Homo sapiens IL-28A mRNA-sequence ( NM_172138 . 1 ) was utilized . The forward primer ( IL-28AF1: CAAAGATGCCTTAGAAGAGTCG ) spans the exon/exon junction of exons 2 and 3 of IL-28A . The IL-28A-specific probe ( IL-28A probe: GCTGAAGGACTGCAGGTGCCA ) is in exon 3 and the reverse primer ( IL-28AR1: TCCAGAACCTTCAGCGTCAG ) is found in the fourth exon . Forward and reverse primers were identical for both genes due to high percentage sequence homology . Assay specificity was conferred by a two-nucleotide difference in the probe sequence ( underlined ) . Both assays yield 150 nt products . Primers and a probe specific for IL-29 were designed based on the Homo sapiens IL-29 mRNA-sequence ( NM_172140 . 1 ) . The forward primer ( IL-29F1: GGACGCCTTGGAAGAGTCA ) spans the exon/exon junction of exons 1 and 2 of IL-29 . The IL-29-specific probe ( IL-29 probe: CTCAAGCTGAAAAACTGGAGTTGCAGC ) is in the second exon of the IL-29 gene and the IL-29 reverse primer ( IL-29R1: CCAGGACCTTCAGCGTCA ) is in the third exon . The IL-29 assay yields a product of 146 nucleotides . Primers and probes were manufactured by IDT ( Integrated DNA technologies , Iowa , USA ) . The specificity of the three sets of qRT PCR assays was tested against Invivogen expression plasmids containing complete IL-28A ( puno1-hIL-28A ) , IL-28B ( puno1-hIL-28B ) and IL-29 ( punoIL-29 ) sequences [16] . The specificity of all qRT PCR assays has been previously assessed [29] . As a house keeping gene HPRT was used [29] . Cell-free supernatants were collected from PBMC cultures at indicated time points and stored at −80°C until analysis . An in-house human IgG ELISA assay was developed using antibodies and human IgG standard . In brief , 96 well EIA/RIA plates ( Costar ) were coated overnight with donkey anti-human IgG antibody at 5 µg/ml . Plates were washed with PBS/0 . 05% Tween and supernatant samples ( diluted 1∶5 ) or ChromPure Human IgG standard ( Jackson Immunoresearch ) were added in duplicate for 2 hrs at room temperature . After washing extensively , detection antibody ( goat anti-human IgG alkaline phosphatase , 1∶15 , 000 ) was added for 1 hr at room temperature . After washing , PNPP substrate was added and the plate was read every 5 min at 405 nm with correction at 570 nm . Virus-specific IgG production was assessed by coating the previously mentioned plates either with pH1N1 antigen ( contained in the vaccine , and used also for T-cell stimulation assays: NIBSC , NXMC-X179A ) or purified pH1N1 hemagglutinin ( Influenza reagent resource ( IRR ) , FR-559 ) . Supernatants from stimulated PBMC cultures ( day 7; diluted 1∶2 ) were added and the amount of bound antibody was determined as above except supernatants were incubated overnight to increase sensitivity . To confirm specificity , supernatants were added to plates coated with hepatitis B virus surface antigen ( Creative Biomart ) or HCV E2 antigen ( Immunodiagnostics , Inc . ) and no signal was detected above background ( Median ODs for HA coated wells: 0 . 607; HBs Ag: 0 . 00625; HCV E2: 0 . 00425; and unstimulated sample supernatant with HA coated wells: 0 . 01175 ) . Results are expressed as absorbance values ( 405 nm–570 nm ) with the plate blank subtracted . Antagonistic peptides were designed as previously published [29] . Briefly , based on previous publications of the crystal structures of IL-29 and the receptor IL28RA ( PDB: 3OG4 , 3OG6 ) , we determined the amino acid residues , which are in close proximity to mediate interaction between the two molecules . The selected amino acids were compared with the crucial amino acids described in the literature [39] , [56] . In order to preserve the interaction domain structure ( helix or loop ) we selected nearby amino acids that may stabilize the binding domain for inclusion in peptides . Based on the oligomeric state structure suggested , we designed peptides , which have the potential to bind both IFN-λ and IL28RA . We used the crystal structure of IL-28B oligomer ( PDB: 3HHC [56] ) focusing on amino acids , whose residues may be involved in the interactions responsible for the formation of the oligomeric state . We then designed peptides to mimic these interaction domains in order to prevent the formation of oligomers . All peptides ( and all other reagents ) were tested for endotoxin and had <0 . 25 endotoxin units ( EU ) /ml . Recombinant IL28RA was coated on an ELISA plate and pre-treated with increasing concentrations of peptides . Next , recombinant , his-tagged IL-28B at a fixed concentration of 100 ng/mL was added . Anti-his secondary antibody was used to determine the relative amount of bound IFN-λ to the IL28RA . These dose-response curves allowed us to determine the effectiveness of binding inhibition of each peptide . Antagonistic peptides were added in a range from 10 nM to 100 µM . THP1-derived macrophages were generated as previously described [57] . Briefly , THP1 cells were seeded in presence of PMA ( 100 nM ) and incubated at standard conditions in RPMI 10% heat inactivated FCS for 3 days . Then media and non-adherent cells were removed and fresh media without PMA added for another 5 days incubation . These cells ( THP1-derived macrophages ) where used for the peptide screening assays . Prior to surface staining , LIVE/DEAD staining was performed ( near-IR; Invitrogen ) . Markers for identifying T-cell subsets were CD3 , and CD4 . Intracellular cytokine staining was performed according to previously published protocols after overnight stimulation [58] . IL-4 was used as a key representative for Th2 cytokine production . Background ( unstimulated samples ) were subtracted from stimulated results . All reagents including perm and fixation buffers and antibodies were from eBioscience . Isotype controls have previously been used to establish the assays . Markers for identification of B-cell subsets were CD20 and CD27 , where naïve B-cells are CD20+CD27− and memory B-cells are CD20+CD27+ . MHC-II , CD86 and CD69 served as activation markers ( Biolegend or eBioscience; see S4 Figure ) . For B-cell expansion experiments , PBMCs were labeled with Cell Trace Violet proliferation dye ( Invitrogen ) . Labeled PBMCs were washed and resuspended in RPMI with 10% FBS and plated in a 96-well format . Stimulation was according to the respective experimental condition in 5% CO2 at 37°C . 2 days after initial stimulation , 50 µL of fresh RPMI was added . THP1-derived macrophages were stained using STAT1-phosphorylation antibodies ( BD Bioscience , AF647 Mouse anti-stat-1 pY701 ) and respective isotype controls . Macrophages were pretreated with blocking peptides and challenged with IL-28B ( 100 ng/mL ) for 15 min . Then cells were fixed and permeabilized as previously described . Two luminex-based cytokine profiling kits were used ( Eve Technologies , Calgary , Canada ) . ( i ) 17-plex: Fractalkine , IFN-α , IFN-γ , GRO , MCP-3 , IL-13 , sCD40-L , IL-9 , IL-1β , IL-2 , IL-4 , IL-5 , IL-6 , IP-10 , MCP-1 , MIP-1α , and TNF-α . ( ii ) 41-plex: EGF , Eotaxin , FGF-2 , FLT3 , Fractalkine , G-CSF , GM-CSF , GRO . IFN-α2 , IFN-γ , IL-1α , IL-1β , IL-1ra , IL-2 , IL-3 , IL-4 , IL-5 , IL-6 , IL-7 , IL-8 , IL-9 , IL-10 , IL-12p40 , IL-12p70 , IL-13 , IL-15 , IL-17 , IP-10 , MCP-1 , MCP-3 , MDC , MIP-1α , MIP-1β , PDGF AA , PDGF AB/BB , RANTES , sCD40L , sIL2ra , TGF-α , TNF-α , TNF-β , and VEGF . Our independent experiment for examining a time course of cytokine induction was a custom-plex based on the 17-plex and still run and analyzed by Eve Technologies . GeneSpring GX version 12 ( Agilent Technologies , Canada ) was used for cluster and principal component analysis ( PCA ) of the cytokine data measured in H1N1-stimulated PBMCs . Non-stimulated background samples were subtracted prior . Percentile shift was used as normalization algorithm and baseline transformation was performed to median of all samples . Hierarchical clustering of both conditions and cytokines was done using Euclidean as similarity measure and Centroid as linkage rule . PCA was used to detect major trends in the experimental conditions . 2D PCA Scores are shown for the first and second PCA components . They capture about 90% of the variation and visualize the separation of the conditions . The PCA loading plot indicates the separation in subsets of cytokines ( x-axis ) and denotes their relative contribution to the principal components on the y-axis . All pre- and post-vaccination samples of 47 transplant recipients were included . The conditions considered for analysis were: pre- vs . post-vaccination and minor- vs . major-allele IL-28B genotype . Statistical analyses were performed using PASW Statistics ( version 20 . 0 , Chicago , Ill . ) and GraphPad Prism ( version 4 . 0 , La Jolla , CA ) . Data are shown with median and inter-quartile ranges unless otherwise indicated . Categorical variables were analyzed using a Chi-Square ( Chi2 ) . Continuous non-normal distributed data ( Shapiro Wilk test ) were analyzed using a Mann-Whitney U test ( MWU ) or Kruskal-Wallis test ( KW ) . Paired data were analyzed using Wilcoxon matched pairs rank test ( WCR ) . All tests were two-tailed .
|
Infection with influenza viruses is associated with high morbidity and mortality . Therefore , vaccination is recommended in immunosuppressed patients , however often the post-vaccine induced protection is insufficient . Factors associated with reduced vaccine responses may guide preventive strategies and could offer novel targets for adjuvants . Here , we explore the impact of IL-28B on B- and T-cell responses during vaccination . We found that a single nucleotide polymorphism ( minor allele genotype ) in the IL-28B gene was associated with a significant increase in the antibody seroconversion rate following influenza vaccination . Interestingly , this SNP reduces the expression of IL-28B . In addition , in vitro stimulation of peripheral blood mononuclear cells from patients with the SNPs had increased IL-4 production in CD4 T-cells . As a potential mechanism , we show that recombinant IL-28B inhibits influenza stimulated Th2 cytokine release , B-cell activation/proliferation and H1N1-induced IgG secretion . Next , we developed antagonistic peptides to block the IFN-λ receptor . Pre-treatment with the antagonistic peptides increased in vitro B-cell activation and antibody production in healthy individuals and transplant recipients . Together , these findings identify IL-28B as a key regulator of Th1/Th2 balance during influenza vaccination . Blockade of the IFN-λ receptor with antagonistic peptides may offer a novel strategy to augment vaccine responses .
|
[
"Abstract",
"Introduction",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"viral",
"vaccines",
"clinical",
"immunology",
"virology",
"immune",
"deficiency",
"biology",
"and",
"life",
"sciences",
"immunology",
"vaccination",
"and",
"immunization",
"microbiology",
"immune",
"response",
"immunocompetence"
] |
2014
|
IL-28B is a Key Regulator of B- and T-Cell Vaccine Responses against Influenza
|
Learning arises through the activity of large ensembles of cells , yet most of the data neuroscientists accumulate is at the level of individual neurons; we need models that can bridge this gap . We have taken spatial learning as our starting point , computationally modeling the activity of place cells using methods derived from algebraic topology , especially persistent homology . We previously showed that ensembles of hundreds of place cells could accurately encode topological information about different environments ( “learn” the space ) within certain values of place cell firing rate , place field size , and cell population; we called this parameter space the learning region . Here we advance the model both technically and conceptually . To make the model more physiological , we explored the effects of theta precession on spatial learning in our virtual ensembles . Theta precession , which is believed to influence learning and memory , did in fact enhance learning in our model , increasing both speed and the size of the learning region . Interestingly , theta precession also increased the number of spurious loops during simplicial complex formation . We next explored how downstream readout neurons might define co-firing by grouping together cells within different windows of time and thereby capturing different degrees of temporal overlap between spike trains . Our model's optimum coactivity window correlates well with experimental data , ranging from ∼150–200 msec . We further studied the relationship between learning time , window width , and theta precession . Our results validate our topological model for spatial learning and open new avenues for connecting data at the level of individual neurons to behavioral outcomes at the neuronal ensemble level . Finally , we analyzed the dynamics of simplicial complex formation and loop transience to propose that the simplicial complex provides a useful working description of the spatial learning process .
Considerable effort has been devoted over the years to understanding how the hippocampus is able to form an internal representation of the environment that enables an animal to efficiently navigate and remember the space [1] . This internal map is made possible , in part , by the activity of pyramidal neurons in the hippocampus known as place cells [2] , [3] . As an animal explores a given environment , different place cells will fire in different , discrete regions of the space that are then referred to as that cell's “place field” [2] , [3] . Despite decades of research , however , the features of the environment that are encoded , the identity of the downstream neurons that decode the information , and how the spiking activity of hundreds of cells is actually used to form the map all remain unclear . We recently developed a computational model for spatial learning , focusing on what information is available to the still-unidentified downstream neurons [4] . We reasoned that the information they decode must be encapsulated in the temporal patterns of the place cell spike trains , specifically place cell co-firing [4] , [5] . Because place cell co-firing implies that the respective place fields overlap , the resulting map should derive from a sequence of overlaps between parts of the environment . The information encoded by the hippocampus would therefore emphasize connectivity between places in the environment , which is a topological rather than a geometric quality of space [4] . One advantage of this line of reasoning is that a topological problem should be amenable to topological analysis , so we developed our model using conceptual tools from the field of algebraic topology and , in particular , persistent homology theory [6] , [7] . We simulated a rat exploring several topologically distinct environments and found that the information encoded by place cell co-firing can , in fact , reproduce the topological features of a given spatial environment . We also found that , in order to form an accurate spatial map within a biologically reasonable length of time , our simplified model hippocampus had to function within a certain range of values that turned out to closely parallel those obtained from actual experiments with healthy rodents . We called this sweet spot for spatial learning the learning region , L [4] . As long as the values of the three parameters ( firing rates , place field sizes , and number of active neurons ) remain within the learning region , spatial learning is reliable and reproducible . Beyond the perimeters of L , however , spatial learning fails . Several features of this model are intuitively appealing . First , the size and shape of L vary with the difficulty of the task: the greater the complexity of the space to be learned , the narrower the range of values that can sustain learning and thus the more compact the learning region . Second , there is a certain tolerance for variation among the three parameters within L: if one parameter begins to fall outside the sweet spot , spatial learning can still occur if there is sufficient compensation in the other two parameters . Our model suggests that certain diseases ( e . g . , Alzheimer's ) or environmental toxins ( e . g . , ethanol , cannabinoids ) disrupt spatial learning over time by gradually shifting mean neuronal function ( place cell firing , neuronal number , or place field size ) beyond the perimeter of the learning region . This notion receives support from studies of mouse models that show a correlation between impairment in spatial cognition and larger , more diffuse place fields , lower place cell firing rates , and smaller numbers of active cells [8] , [9] . All this corresponds well with our subjective experiences of learning: the complexity of the task influences learning time; when the task is difficult we can feel we are at or just beyond the limits of our capacity; disease or intoxication can reveal limits in our spatial cognition that would normally be compensated for . In this paper we focus on analyzing the structure of the learning region itself . We begin by making the computational model more physiologically accurate . There is a θ ( theta ) component of subcortical LFP oscillations that occurs in the frequency range of 6-12 Hz and regulates spiking activity [10] . The timing of place cell spiking in the hippocampus is coupled with the phase of θ-oscillations so that , as a rat progresses through a particular place field , the corresponding place cell discharges at a progressively earlier phase of each new θ-cycle [11] . This phenomenon , called theta phase precession , reproduces short sub-sequences of an animal's current trajectory during each θ-cycle [11] . This has been construed to suggest that θ-phase precession helps the hippocampus remember the temporal sequence of the rat's positions in a space ( i . e . , its trajectory ) [12] , [13] , thereby enhancing spatial learning and memory . If this is the case , θ phase precession should enhance learning in our computational model . Indeed , we find that it significantly improves and stabilizes spatial learning . We also find that different temporal windows to define co-firing exert a pronounced influence on learning time , and the most efficacious window widths correspond with experimental predictions . Finally , we analyze simplicial complex formation within the learning region , examining both the structure of the complexes and the dynamics of loop formation , and find an explanation for the poor efficiency of ensembles at the boundary of the learning region compared to peak-performing ensembles .
We will first briefly describe the fundamental concepts on which our model is based ( this section is an abbreviated version of the approach described in [4] ) . Central to this work is the concept of a nerve simplicial complex , in which a space X is covered by a number of smaller , discrete regions [14] . If two regions overlap , the corresponding vertices , vi and vj , are considered connected by a 1D link vij ( Figure 1 ) . If three regions overlap , then vij , vjk , and vki support a 2D triangular facet or simplex σijk , and so on as the number of overlaps and links increase . The structure of the simplicial complex approximates the structure of the environment: the complex N ( X ) obtained from a sufficiently dense cover of the space X will reproduce the correct topological indices of X ( see [4] for details ) . For our model we developed a temporal analogue to the simplicial complex , i . e . , a simplicial complex that builds over time: when the animal is first introduced to the environment , there will be only a few data points from place cell firing , but as the animal explores the space the place cell firing data accumulate . ( Rodent experiments indicate that place fields take about four minutes to develop [15] . ) As the animal explores its environment and more place cells fire ( and co-fire ) , the simplicial complex T grows with T ( time ) ( T = T ( T ) ) . Eventually , after a certain minimal time Tmin , the space's topological characteristics will stabilize and produce the correct topological indices , at which point the topological information is complete . Tmin is thus the minimal learning time , the time at which a topologically correct map is first formed . The correct topological indices are indicated by Betti numbers , which in turn are manifested in persistent cycles ( see [4] , [7] , [16] ) . As the rat begins to explore an environment , the simplicial complex T ( T ) will consist mostly of 0-cycles that correspond to small groups of cofiring cells that mark contractible spatial domains . As the rat continues to explore the environment , the co-firing cells will produce links between the vertices of T ( T ) , and higher-dimensional cycles will appear . As T increases , most cycles in each dimension will disappear as so much “topological noise , ” leaving only a few persisting cycles that express stable topological information ( Figure 1C ) . The persistent homology method [6] ( see [4] Methods ) enables us to distinguish between cycles that persist across time ( reflecting real topological characteristics ) and transient cycles produced by the rat's behavior ( e . g . , circling in a particular spot during one trial or simply not venturing into one part of the space during early explorations ) . The pattern of cycles is referred to as a barcode [16] that can be easily read to give topological information about a given environment ( Figure 1C ) [6] , [7] . If theta precession serves to enhance learning , as has been predicted [17]–[19] , then it should enhance spatial learning in our model . This could occur by any of several means . First , theta precession might enlarge the number of ensembles capable of the task by expanding the scope of the parameters ( including firing rates or place field sizes normally out of the bounds of L ) . Second , it might make the ensembles that are in L converge on the correct topological information more rapidly . Third , it might make the same ensembles perform more reliably ( e . g . , succeeding in map formation a greater percentage of times in our simulations ) . To test the effect of theta precession in our model , we compared the rates of map formation for those formed with and without θ-precession . We tested 1710 different place cell ensembles by independently varying the number of place cells ( N; 19 independent values , from 50 to 500 ) , the ensemble mean firing rate ( f; 10 independent values , from 4 to 40 ) , and the ensemble mean place field sizes ( s; 9 independent values , from 5 to 30 ) [Methods; see [4] and Methods therein for further details] . For statistical analysis , we simulated each map 10 times so that we could compute the mean learning time and its relative variability , ξ = Δ Tmin/Tmin , for each set of ( s , f , N ) values . In the following we will suppress the bar in the notation for the mean f , s , N , and Tmin . Figure 2 shows the results of these simulations in a 1×1 m space with one hole . ( The size of the environment in this study is smaller than the ones used in [4] , for two reasons: to avoid the potential problem of place cells with more than one field , and to reduce computational cost; see Methods . ) The learning region L is small and sparse in the θ-off case , but notably larger and denser in the θ-on case ( Figure 2A ) . Values that would be just beyond the learning region—N that may be too small , or place fields that are too large or too small , or firing rates too high or too low [4]—thus become functional with the addition of θ-precession . Two criteria reveal the quality of the map-forming ensembles: speed and consistency in converging toward the correct topological signature . The fastest map formation times ( under 4 minutes ) are represented by blue dots; as the color shifts toward red , map formation times become longer and the error rate ( failure to converge ) increases . The size of the dot represents the success rate: small dots represent ensembles that only occasionally converge on the correct information , large dots represent ensembles that converge most or all of the time . θ-precession increases the probability of convergence across all ensembles that can form accurate maps at all ( Supplemental Figure S1 ) . Since we were interested in understanding the dynamics of efficient learning , however , we created a more stringent definition of the learning region to focus on the core of L where map-formation is most rapid and reliable , as well as to make the results more legible ( L can be quite dense , as in Figure 2A and Supplemental Figure S1 ) ; if θ-precession truly enhances learning , its effect should be apparent even in the most successful ensembles , and indeed this was the case . The point clouds in Figure 2B depict those ensembles that formed maps with a convergence rate of ρ≥0 . 7 ( i . e . , those that produced correct topological information at least 70% of the time ) and simultaneously had low relative variability of the Tmin values , ξ<0 . 3 . Even within this more efficient core of L , the effect of θ-precession was pronounced . The histograms of the computed mean learning times are closely fit by the Generalized Extreme Value ( GEV ) probability distribution ( Figure 2B ) . The distributions show that θ-precession reduced the mean learning times Tmin: the mode of all the θ-on GEV distributions decreased by ∼50% compared with the θ-off case for the learning region as a whole ( Figure 2C ) and by ∼15% for the efficient ensembles at the core of L ( Figure 2D ) . Moreover , the effects of adding θ-precession—reducing map formation time and decreasing the relative variability of the Tmin values—were manifested in all maps , not just those with high ( ρ ≥0 . 7 ) convergence rates ( Figure 2C , D ) . The histograms for all maps ( all ρ -values ) fit by the GEV distribution reveal that the typical variability ( the mode of the distributions ) in the θ-on cases is about half the size of the θ-off case ( Figure 2E ) . In our model , therefore , θ-precession strongly enhances spatial learning . Since we do not know what features of θ-oscillations might be important [20] , we studied four different θ-oscillations , two simulated and two derived from electrophysiological experiments in wild-type rodents . Specifically , we modeled the effect of theta precession on the topological map by coupling the place cells' Poisson firing rates , λc , with the phase of the following four θ-oscillations: 1 ) θ1 – a single 8 Hz sinusoidal wave , 2 ) θ4 – a combination of four sinusoids , 3 ) θM – a subcortical EEG signal recorded in wild-type mouse , and 4 ) θR – a subcortical EEG signal recorded in a rat ( Supplemental Figure S2; see Methods ) . The last three signals were filtered in the θ-domain of frequencies ( 6–12 Hz ) . The distribution of the learning times , the histograms of the mean learning times , and the histograms of the relative variability , ξ , for all four different theta cases are shown in Supplemental Figures S3 and S4 . To compare the θ-off and θ-on cases , we performed two-sample Kolmogorov-Smirnov ( KS ) tests for all pairwise combinations of the studied sample sets [21] . This produced a 5×5 matrix of the p-values , pij , where i , j = 0 ( no theta ) , θ1 , θ4 , θM , and θR . Black squares signify a statistically significant difference between cases i and j ( p<0 . 05 ) ; gray squares signify no statistically significant difference . The statistical difference diagrams for the sets of Tmin values ( Supplemental Figure S3 ) and for the learning time variability ( Supplemental Figure S4 ) indicate that the distributions of learning times in the various θ-on cases were very similar , but the difference between all of these and the θ-off case was statistically significant . So far we have described the outcome of place cell ensemble activity in terms of the time at which the correct number of loops in the simplicial complex T emerges . But the learning process can also be described by how spurious loops are handled in the system . These loops are a fair representation of the subjective experience of learning . It takes time to build a framework into which new information can be properly slotted: until that framework is in place—whether it's a grasp of the layout of a neighborhood or the basic principles of a new field of study—we have incomplete hunches and many incorrect notions before experience ( more learning ) fills in our understanding . Translating this into topological terms , as the knowledge gaps close , the spurious loops contract . We therefore wanted to study the effects of theta precession on the dynamics of loop formation . Does a “smarter” ensemble form more spurious loops or fewer ? Does it resolve those loops more quickly ? We concentrated on the 1D cycles , which represent path connectivity within the simplicial complex , because they are more numerous and thus produce more robust statistics than the 0D cycles . Figure 3 shows that θ-precession shortened the duration of the spurious loops . The KS test reveals a statistically significant difference between the lifetimes of spurious loops in the θ-off case and those in all the θ-on cases ( Figure 3B ) . To simplify the presentation of the results produced by the statistically similar θ-on cases , we combined the data on spurious loop duration from all four θ-driven maps into a single histogram . It is interesting to note that the probability distributions for loop dynamics are typically better fit by the gamma distribution ( Figure 3B–D ) . In the θ-driven maps , a typical spurious loop persisted for 50% less time than it would without θ-precession ( Figure 3B ) . It is worth noting that the spurious loops persisted longer at the lower boundary of the learning region , where the mean firing rates and place field sizes are smallest . This makes sense , insofar as whatever information appears will take longer to be corrected . Statistical analysis of the largest number of loops observed at any given point over the course of the map formation period also differentiated θ-driven from θ-off maps . Curiously , θ-driven cases tended to produce a significantly higher mean number of spurious loops than the θ-off case ( Figure 3C ) , but with a lower peak number of loops ( Figure 3D ) . This implies that θ-precession enhances the speed of spatial learning overall at the price of creating more ( transient ) errors; lots of spurious loops are formed early on , but they disappear faster . The KS test shows that the distributions of the mean number of loops in all θ-on cases differ from one another; only the maps driven by the two simulated θ-signals gave statistically similar results . In our model , spatial learning can be quantified by the time required for the emergence of correct topological information , but it can also be quantified by studying the simplicial complex itself . We noted earlier that the structure of the simplicial complex approximates the structure of the environment . Similarly , it is possible to conceive of a simplex as a mathematical analogue to a cell assembly ( a group of at least two cells that repeatedly co-fire and form a synapse onto a readout neuron ) , and to view the simplicial complex as analogous to the realm of possible connections within the hippocampus . We were therefore curious: since it is in the interest of neural function to be efficient , how many cell assemblies ( simplices ) does it take to encode a given amount of information ? We would predict that the fewer the connections , the better , for the sake of efficiency . One of the major characteristics of a simplicial complex T is the number of n-dimensional simplices it contains , traditionally denoted as fn . The list of all fn –values , ( f1 , f2 , … , fn ) , is referred to as the f-vector [22] . Since the D-dimensional simplices in T correspond to ( D+1 ) -ary connections , the number of which depends on the number of vertices , N , we considered the fn values normalized by the corresponding binomial coefficients , which characterize the number of simplices connecting vertices in the complex T . We can consider η an index of the connectivity of the simplicial complex . Since we model 2D spatial navigation , we analyzed the connections between two and three vertices , i . e . , the 1D and 2D simplices , of T ( the number of 0D simplexes normalized by the number of vertices in T is η0≡1 ) . Figure 4 shows the distribution of the normalized number of simplices at the time the correct signature is achieved ( η1 and η2 , for 1D and 2D , respectively ) . As expected , the number of simplices was smaller at the lower boundary of the learning region L ( the base of the point cloud ) and increased towards the top of L where place fields are larger and the firing rates are higher , each of which would produce more place cell co-firing events . Remarkably , the number of simplices depended primarily on the mean place field size and on the mean firing rate of the ensemble , and not on the number of cells within the ensemble . This suggests a certain universality in the behavior of place cell ensembles that is independent of their population size . In the ensembles with smaller place fields and lower firing rates , about 1 . 5% of place cell pairs and 1 . 7% of the triplets were connected , and this was enough to encode the correct topological information , whereas in the ensembles with low spatial selectivity and higher firing rates , 25% of pairs and 8% of triplets were connected . These ensembles , in which the place fields and spike trains will by definition have a lot of overlap , are forced to form many more 1D and 2D simplices in order to encode the same amount of information and are thus less efficient ( Figure 4 , third column ) . According to our model , such ensembles and the hippocampal networks whose activity they represent are inefficient on two counts . First , these larger , more complex temporal simplicial complexes ( analogous to a larger number of coactive cell groups ) will take longer to form correct topological information , if they can manage it at all . Second , a larger number of coactive place cells would hamper the training of downstream readout neurons , thereby impeding reliable encoding of spatial information . This is consistent with studies showing that the number of cells participating in a particular task decreases until it reaches an optimal number that fire at a slightly higher rate than their no-longer-participating neighbors [23] . Our model depends on patterns of place cell co-firing , but we had not previously explored what the optimal temporal window for defining co-activity might be . Experimental work supports the widely held assumption that the temporal unit for defining coactivity ranges between tens [24] and hundreds of milliseconds [25]–[27] . Our model , however , enables us to approach the question of optimal width for the coactivity window theoretically . Clearly , if the time window w is too small , then the spike trains produced by the presynaptic place cells will often “miss” one another , and the map will either require a long time to emerge or it may not be established at all . One would thus expect large values Tmin ( w ) for a small w . On the other hand , if w is too large , it will allow cells whose place fields are actually distant from one another to be grouped together , yielding incorrect topological information . Theta rhythm itself will have a tendency to group sequential spike trains together , but clearly there must be limits to this , or else some place cells would be read downstream as co-firing when they actually are not . Therefore , there should exist an optimal value of w that reliably produces a finite , biologically relevant learning time Tmin at which the learning region L is robust and stable . We assume that the capability of a read-out neuron to detect place cell co-activity is specified by a single parameter , the width of the integration time window w , over which the co-appearance of the spike trains is detected . ( We considered the possible effect of time bin position on co-activity , but found this did not affect outcome; see Methods . ) We defined cell coactivity as follows: if presynaptic neurons c1 and c2 send a signal to a read-out neuron within a certain time window w , their activity will be interpreted as contemporaneous . The width of this time window may be positive ( c2 becomes active w seconds after c1 ) or negative ( c2 becomes active while c1 is still firing ) . We studied window widths w for which the place cell spike trains would eventually be able to produce the correct topological signature ( the Betti numbers , see Methods in [4] ) . In order to describe the dependence of learning times on the window width , Tmin ( w ) , we scanned an array of 24 values of w ( ranging from 0 . 1 to 5 θ-periods ) for each combination of the parameters ( mean s , f , and N ) and noted the width of the value wo , at which the map began to produce the correct topological signature . We call this initial correct window width the “opening” value . A typical result is provided by an ensemble with f = 28 Hz , s = 23 cm and N = 350 , in which an accurate topological map emerges at a fairly small window width , wo∼25 msec ( Figure 5 ) . The distribution of the opening window widths shows that wo may exceed 1 . 5 θ-periods ( ∼25 msec ) , which matches the slow γ-period [24] , [28] ( Supplemental Figure S5 ) . Since at this stage γ-oscillations have not been explicitly built into the model , this correspondence is coincidental , if suggestive . As expected , the values of learning times at wo were rather large: Tmin ( wo ) ∼20 minutes in θ-off case and Tmin ( wo ) ∼30 minutes in the θ-on case , and in some cases exceeding one hour ( mostly for the ensembles with low firing rates ) . For small window widths , the value of the learning time Tmin ( w ) was very sensitive to variations in w ( Supplemental Figure S6 ) . As w increased , however , the learning time reached a plateau around some larger value ws . This implies that in order to produce stable values for Tmin that are biologically plausible , the values of the window widths should start around ws . The distribution of the ws values demonstrates that in the θ-off case the stabilization is typically achieved at approximately one θ-period , and in the θ-on case at about ∼1 . 2–1 . 5 θ-cycles ( Figure 6 ) , which justifies our choice of a two θ-period window width for the computations and corresponds well with the predicted limit of 150 msec for θ-cycle cofiring in sequence coding [29] . Further increasing the integration time window w did not significantly alter the learning time Tmin in L; instead , the rate of map convergence decreased until the maps completely fail to encode the correct topological information at w ζ 4 . 5 θ-periods . From the perspective of our current model , the range of optimal window widths w is between 20–25 msec and 0 . 5 secs . Finally , we sought to uncover a relationship between learning time and window width . Our analysis suggests that Tmin is inversely proportional to a power of the window width ( Figure 5B , Supplemental Figure S7 ) .
Numerous experiments have demonstrated that θ precession is important for spatial learning . θ-power increases with memory load during both spatial and non-spatial tasks in humans [30] , [31] and in rodents [32] , [33]; spatial deficits correlate with a decrease in the power of theta oscillations in Alzheimer's disease [34] and in epilepsy [35] , [36] . If θ-signal is blocked by lesioning the medial septum ( which does not affect hippocampal place cell representations ) , it severely impairs memory [37] and the acquisition of new spatial information [38] . Recent experiments demonstrate more directly that destroying θ precession by administering cannabinoids to rats correlates with behavioral and spatial learning deficits [17] , [39] . But at what level , and through what mechanisms , does θ precession exert its influence ? The effect of θ-precession on the structure of the spike trains is rather complex [40] . On the one hand , it groups cell spikes closer together in time and enforces specific sequences of cell firing , which is typically interpreted as increasing the temporal coherence of place cell activity [41]–[43] . One might predict that grouping spikes together would ( somehow ) speed up learning . On the other hand , θ-precession imposes extra conditions on the spike timing that depend on θ-phase and on the rat's location with respect to the center of the place field through which it is presently moving . Since every neuron precesses independently , one could just as well predict that θ modulation would either restrict or enlarge the pool of coactivity events , which in turn would slow down learning at the level of the downstream networks , and that the beneficial effect of the θ rhythm is a higher-order phenomenon that occurs elsewhere in the brain . Our results suggest that θ precession may not just correlate with , but actually be a mechanism for , enhancing spatial learning and memory . The interplay of θ precession and window width , especially the extremely long learning times at the opening window width wo , is particularly illuminating here . As noted , theta precession acts at both the ensemble and the individual neuron level: it groups spikes together , but each neuron precesses independently . When the time window is sufficiently wide , the coactivity events are reliably captured , the first effect dominates , and the main outcome of theta precession is to supply grouped spikes to downstream neurons . For very small time windows , however , the system struggles to capture events as coactive , and the extra condition imposed by phase precession acts as an impediment: detected coactivities are rarer , and learning slows down . Put more simply , imagine in Suppl . Figure S8 that the window is only one spike wide: in a train of 10 spikes that overlaps by one spike with another train , it will take 10 windows before the overlap is detected . It is noteworthy that the presence of theta precession was clearly more important than the details of the oscillation . Although theta precession enhanced learning in our simulations , learning times were relatively insensitive to the details of the theta precession chosen . One might expect differences in spike train structure induced by the four different θ-signals studied to alter the dynamics of the persistent loops and thus learning efficiency . Our results show , however , that differences that would matter at the level of individual cells are averaged out at the level of a large ensemble of cells . Here again the model shows its particular strength: it allows us to correlate parameters of activity at the level of individual neurons with the outcome at the level of an ensemble of hundreds of cells , providing a framework for understanding how micro-level changes play out at the behavioral level . Interestingly , we also saw a difference between the micro and macro levels when we considered whether the placement of a temporal window affected what would be considered co-activity ( see Methods and Supplementary Figure S8 ) . In theory it should , but the effect at the macro level washes out and we found that only the temporal width of the window matters for learning time . Beyond validating the model as a reliable way to study physiological aspects of spatial learning , we have gone further in this work to analyze the simplicial complex itself as a way of describing learning . As a rat starts to explore an environment , some cells begin to form place fields . Then , the co-firing of two or more place cells will define the respective places as connected in space and temporal experience and will create corresponding simplices in the simplicial complex T . With time , these simplices form a chain corresponding to the animal's route through the space . If the environment is bounded , the rat will discover new connections between the same places ( arriving to the same location via different routes ) . As a result , the chains of connected simplices grow together to form loops . Existing loops become thicker and may eventually “close up” and disappear , yielding surfaces . The appearance of such surfaces is significant: the closing up of a D-dimensional surface corresponds to the contraction , or disappearance , of one of its boundaries , which itself is a D1-dimensional loop . Eventually , the structure of the simplicial complex saturates such that no new simplices ( connections between places ) are produced and no more loops contract because all that could close have already closed . At this point , the saturated simplicial complex T encodes not only the possible locations of the rat , but also connections between the locations , along with the information about how these connections can be deformed , e . g . , whether they are contractible or whether they encircle an obstacle and cannot be contracted into a point . Thus , the saturated temporal simplicial complex T is a framework that unifies information about places and spatial-temporal relationships between them . This framework might correspond fairly well to the structure formed by synaptic connectivity in the rat's hippocampus , which allows the rat to explore and retrieve information by “pinging” the network , without physical navigation [44] . In addition to the practical benefits of a model that consistently produces biologically relevant results , there is a special appeal in the ability of algebraic topology to provide insight into the mechanisms of learning . It is fitting that a method developed to simplify the analysis of high-dimensional data with many coordinates might itself represent how the brain approaches a similar challenge in the real world .
An ensemble of N cells is described by N peak rates , f1 , f2 , … , fN , and N place field widths , s1 , s2 , … , sN . As in [4] , we assume that the values fi , and si , are log-normally distributed around a certain ensemble-mean firing rate and the ensemble-mean place field size , with the variances σf and σs respectively . To simplify the analysis , we assume that the variance of the firing rate , σf , and of the place field size , σs , are proportional to their means , σf = af and σs = bs , so that the distribution of the firing rates and of the place field sizes are defined by a single parameter , f and s respectively . The protocol of the simulations was similar to [4]: the trajectory was fixed , but the place field centers , rc , are randomly scattered in the environment for each simulation . The first simulated θ-signal ( θ1 ) contained a single sinusoidal oscillation with the frequency f = 8 . 0 Hz . The second simulated θ-signal ( θ4 ) was obtained by combining four sinusoids , with frequencies f1 = 6 . 5 Hz , f2 , = 8 . 65 Hz , f3 = 10 . 0 Hz , and f4 = 11 . 5 Hz , filtered between 6 and 12 Hz . The third and the fourth θ-signals ( θM and θR ) were obtained by filtering subcortical EEG signals recorded in mice and in rat , filtered between 6 and 12 Hz . As the rat just enters the place field of a cell ci , that cell prefers spiking at a high θ-phase , φ*∼2π , and as the animal moves over a distance l towards the center , the preferred phase decreases , reaching φ*∼0 as the rat exits the place field [10] , [45] . We approximate this dependence by a linear function , where Li is the size of the place field ci , Li∼3si , To simulate the coupling between the firing rate and the θ-phase , we modulate the original Gaussian firing rate by a phase-dependent factor Λ ( φ ) , which peaks around 0 , where the width ε we defined as the ratio of the mean distance that rat travels during one θ-cycle to the size of the place field , ε = v/Lf . Typical examples of the resulting theta precession of spiking are shown in Supplemental Figure S2 . For persistence methods see [4] . We conducted our simulations in small environments ( 1×1 m ) . In the “Reducing computational cost by subdividing maps” section ( see also Supplementary Figure S9 ) we show that direct computations based on the Mayer-Vietoris theorem [14] demonstrate that the spatial map corresponding to a larger environment can be split in several smaller pieces , and that the connectivity of the entire spatial map can be computed piece-by-piece , so that the total learning time is approximately equal to the times required to “learn” its parts . This observation helps to reduce the computational cost of the algorithm . In addition , simulating the maps in smaller environments allows us to avoid multiply connected , “patchy” place fields ( in smaller environments there is a lesser chance of observing more than one component of a place field , as occasionally happens when a place cell fires in more than one place ) and helps to bring the density of place cells closer to experimental values . Finally , we notice that , given a particular integration time window width w , the co-activity between two specific spike trains may or may not be detected depending on the position of the time bins ( Suppl . Figure S5 ) . We studied this effect by shifting the time windows 10 times over 10% of the fill window width w and recomputing Tmin , and we saw no difference in the outcome . This implies that Tmin does not depend on the shift of the time window: over the typical learning time scale , the effect produced by the bin shifts averages out . Second , it is clear that a fixed time width is not physiological , because a realistic window size will “jitter” from cell to cell and from moment to moment . However , a direct simulation conducted for w equal to two θ-periods shows that adding up to 50% “jitter” to the window width does not affect the learning times: Tmin remains virtually invariant with respect to the amplitude of the bin size noise . This allows us to simplify the computations by using a single parameter w to characterize coactivity . One of the major computational difficulties in simulating hippocampal spatial maps is the time it takes to analyze a large number of simplices: in our original paper [4] , simulating the map formation time for each set of parameters ( each variation tested 10 , 000 times ) took several months for a 2×2 m virtual space . Considered in mathematical terms: given N vertexes in the simplicial complex T , the number of 1D links scales as ∼N2 , and the number of 2D facets scales as ∼N3 . The topological relationships between them are then defined by a∼N2× N3 incidence matrix [46] . Due to restrictions in computational power , we can investigate ensembles that include up to 400 cells , but in the actual hippocampus there are on the order of 4000 cells active in a given experimental environment [47] , [48] . To reduce the computational load , we took advantage of the Meyer-Vietoris theorem , which allows us to simulate the map by breaking it into its constituent parts . Using Meyer-Vietoris , it is possible to compute the homological characteristics of the entire space X from the homological characteristics of its constituent parts [46] . Specifically , if a space X is split into pieces A and B , X = A∪B , then the homologies of X , Hq ( X ) , are related to the homologies of its parts , Hq ( A ) and Hq ( B ) , via the so-called long exact sequence:If the overlap , A∩B , is topologically trivial , Hq ( A∩B ) = 0 , then the sequence reduces to justin which case the exactness of homomorphisms implies thati . e . , the homologies of the whole space are equal to the direct sum of the homologies of its parts . As a consequence , the Betti numbers from both A and B can be combined to accurately provide topological information about the whole space X . This observation can be used to divide the map formation times of the temporal simplicial complex T ( T ) . If our virtual rat spends time TA and TB in part A and in part B , respectively , then assuming that A and B meet but do not overlap , the total map formation time for the whole environment , TX , can be estimated as A more complete discussion of the mathematical aspects of this dividing approach will be given elsewhere; for example , there is a scale at which space can be no further atomized , and this will require considerable effort to define the size of these ‘atoms’ and account for these size limits . ( At present , it appears that an atom of space is approximately the size of two to three place fields . ) Here we present some numerical results justifying the piecewise computations ( Supplemental Figure S9 ) . We simulated the rat's movement in a large 2×2 m environment with two holes , which we formally divided into 2 , 3 or 4 parts . After the fragments have been chosen , we counted the time spent in each region , and , by using only the spikes fired within a given region computed each region's own learning times . Supplemental Figure S9 shows that the sum of these times is similar to total time spent by rat in the entire arena ( the differences are not statistically significant ) . The second scenario is illustrated in the figure below . Using the adaptive filtering method in [49] it was estimated that in novel environments , place fields form in about four minutes , starting at the background stochastic spiking level of 0 . 1 Hz [15] , [50] , [51] and gain spatial specificity in about the same amount of time . After that , place cells begin to ( co ) fire in a place-specific manner , encoding spatial locations [52] . To include the effect of place field formation into the model , we simulated place cell ensembles with time-dependent firing rate amplitudes , fi , and time-dependent place field widths , si , defined as and in which τi defines the slope of the sigmoid . We chose the typical τi-value , τmean , equal to 240 seconds ( see Supplemental Figure S10 ) and the starting points start at 0 . The results shown in Supplemental Figure S10 demonstrate that the place field formation produces only an additive effect on the overall spatial map formation time: the average map formation time increases by 120% of the τmean with respect to the “base” value obtained for the stable place cells .
|
One of the challenges in contemporary neuroscience is that we have few ways to connect data about the features of individual neurons with effects ( such as learning ) that emerge only at the scale of large cell ensembles . We are tackling this problem using spatial learning as a starting point . In previous work we created a computational model of spatial learning using concepts from the field of algebraic topology , proposing that the hippocampal map encodes topological features of an environment ( connectivity ) rather than precise metrics ( distances and angles between locations ) —more akin to a subway map than a street map . Our model simulates the activity of place cells as a rat navigates the experimental space so that we can estimate the effect produced by specific electrophysiological components —cell firing rate , population size , etc . —on the net outcome . In this work , we show that θ phase precession significantly enhanced spatial learning , and that the way downstream neurons group cells together into coactivity windows exerts interesting effects on learning time . These findings strongly support the notion that theta phase precession enhances spatial learning . Finally , we propose that ideas from topological theory provide a conceptually elegant description of the actual learning process .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"computational",
"neuroscience",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"neuroscience",
"learning",
"and",
"memory"
] |
2014
|
The Effects of Theta Precession on Spatial Learning and Simplicial Complex Dynamics in a Topological Model of the Hippocampal Spatial Map
|
To communicate effectively animals need to detect temporal vocalization cues that vary over several orders of magnitude in their amplitude and frequency content . This large range of temporal cues is evident in the power-law scale-invariant relationship between the power of temporal fluctuations in sounds and the sound modulation frequency ( f ) . Though various forms of scale invariance have been described for natural sounds , the origins and implications of scale invariant phenomenon remain unknown . Using animal vocalization sequences , including continuous human speech , and a stochastic model of temporal amplitude fluctuations we demonstrate that temporal acoustic edges are the primary acoustic cue accounting for the scale invariant phenomenon . The modulation spectrum of vocalization sequences and the model both exhibit a dual regime lowpass structure with a flat region at low modulation frequencies and scale invariant 1/f2 trend for high modulation frequencies . Moreover , we find a time-frequency tradeoff between the average vocalization duration of each vocalization sequence and the cutoff frequency beyond which scale invariant behavior is observed . These results indicate that temporal edges are universal features responsible for scale invariance in vocalized sounds . This is significant since temporal acoustic edges are salient perceptually and the auditory system could exploit such statistical regularities to minimize redundancies and generate compact neural representations of vocalized sounds .
Efficient coding strategies for representing natural sensory signals aim to generate compact neural representations of the external world . Barlow originally proposed the efficient coding hypothesis as a theoretical model of neural coding that aims to maximize information transfer between the external world and the brain while reducing metabolic and computational cost to an organism [1] . According to this model , neural computations performed by the brain should be optimized for extracting information from natural sensory signals and thus should be adapted for statistical regularities prevalent in natural environments . One such statistical regularity is the widely-observed scale invariant relationship between the signal power and frequency of a sensory signal in which the power can be described as a power-law function of general form Sxx ( f ) ∝ f−α , where f is the signal frequency and α is the scaling exponent . Natural visual scenes , for instance , exhibit this generalized form of scaling [2 , 3] and it has been demonstrated that the spatial arrangement of object boundaries which contain edges are necessary to account for the empirically observed scaling exponent of α ≈ 2 [4–6] . Neurons in the central visual system are optimized to encode a wide range of edge orientations [3 , 7] , supporting the hypothesis that the brain is specialized for such statistical regularities in natural environments . As for visual scenes , natural sounds also exhibit various forms of scale invariance , although the acoustic features that contribute to such phenomenon have remained elusive . Long-term fluctuations in the intensity profile of speech and music where first reported to exhibit scale invariance for frequencies < 1 Hz and with a scaling exponent of α ≈ 1 [8] . Subsequent studies further demonstrated that amplitude modulations of natural sounds including speech , animal vocalizations , environmental sounds also exhibit scale invariance [9–12] . Upon representing a natural sound by the analytic signal SA ( t ) = x ( t ) eiθ ( t ) , where i=−1 , θ ( t ) is the carrier phase , and x ( t ) is the modulation signal or equivalently the temporal envelope [13] , the amplitude modulation power spectrum ( AMPS ) is defined as the Fourier transform magnitude of the envelope signal , x ( t ) . For natural sounds the AMPS is well described by a generalized power-law function of the form Sxx ( f ) ∝ f−α , such that the power in the envelope signal drops off with increasing modulation frequency ( f ) with exponent between α ≈ 1 − 2 within the approximate frequency range 1 to 100 Hz [10–12] . With the exception of water sounds , where scale invariance is accounted for by the distribution of self-similar acoustic “droplets” [9] , it remains a mystery as to whether there are universal acoustic features that contribute to scale invariance for broader categories of natural and man-made sounds . Answering this question has important implications as neurons in the mammalian auditory system efficiently encode scale invariant structure in the sound envelope [12 , 14] suggesting it is a critical driver of brain pathway function and perception abilities . Physically , vocalization production in many species entails a source generator ( e . g . , vocal folds ) that produces quasi-periodic envelope signal and articulatory gestures , for instance the opening and closing of the mouth and postural adjustments of the lips and tongue , that dynamically shape the sound envelope during speech production . Envelope fluctuations created by vocal fold vibration lie outside the modulation frequency range where scaling is observed [12] ( i . e . , >100 Hz ) and thus should not contribute to scaling directly . In contrast , transient temporal onset and offset that mark the boundaries between isolated vocalizations are evident across many species and produce transient envelope fluctuations that may contribute to scaling behavior . In human speech , for instance , these salient features are generated by the time-dependent opening and closing of the oral cavity and related articulatory gestures . Drawing analogies from the statistics of natural visual scenes and the prevailing role of object boundaries [4 , 5] , we test the hypothesis that transient temporal edges account for the scaling phenomenon observed in natural vocalized sounds . We demonstrate that temporal edge boundaries in vocalizations are responsible for producing a amplitude modulation spectrum with dual-regime lowpass structure consisting of a flat region for low modulation frequencies and f−2 scale invariant trend at high modulation frequencies .
Sequences of vocalized sounds were obtained from a variety of digital sound sources . Vocalization sequences for a rat pup ( HsdCpb/Wistar ) [15] , a mouse pup ( C57BL/6 mice ) [16] , and a crying infant [17] all consisted of a single long-duration vocalization sequence ( 5–7 min duration; Table 1 ) . Excerpts of continuously spoken speech totaling five minutes were obtained from a BBC broadcast reproduction of Hamlet [18] . Vocalization sequences were also obtained from various bird species [19] ( Track 4: superb lyrebird , 35: winter wren , 41: common loon , and 46: gray-necked wood rail ) and several species of new-world monkeys [20] ( Track 9: Black Mantle Tamarin , Track 18: Golden Lion Tamarin , Track 32: White-Throated Capuchin Monkey , Track 43: Black Howler Monkeys , Track 48: Yellow Tail Wooly Monkey , Track 49: Common Wooly Monkey ) . Single long-duration sequences were not available for either bird or monkey vocalization categories and for this reason shorter sequences ( monkey sequence range = 20–120 sec duration , average duration = 48 sec; bird sequence range = 26–135 sec , average duration = 60 sec ) from different species were used to measure the envelope group statistics and AMPS for these groups . All of the vocalization sequence segments were selected because they contained well-isolated vocalization with minimal background noise . Vocalization sequences were sampled at a sampling rate ( Fs ) to preserve the frequency content of each species ( Table 1 ) . For each vocalization sequence , we computed the vocalization sequence envelopes and computed the amplitude modulation power spectrum ( AMPS ) by extracting the temporal envelope of each sound sequence and subsequently computing the Fourier transform magnitude . Sounds were first bandpass filtered between frequencies flow and fhigh so as to encompass the frequency range of each vocalization sequence sband ( t ) =s ( t ) *hband ( t ) where hband ( t ) is a Kaiser bandpass filter impulse response ( β = 5 . 6 , filter order = 640 , sidelobe error 60 dB ) and * is the convolution operator . Since the vocalizations for each species has dominant energy over a unique frequency range , the frequencies flow and fhigh were individually selected based on visual inspection of the sound spectrum ( Table 1 ) . For the rat and mouse vocalizations the bandpass filter was selected to overlap the ultrasonic range ( flow = 30 kHz and fhigh = 100 kHz ) where the vocalizations had dominant energy . For the remaining vocalizations , flow = 500 Hz and fhigh = 20 kHz . flow was chosen as 500 Hz because we measured the AMPS up to 250 Hz modulation frequency , which requires a carrier frequency of at least 500 Hz . The upper filter cutoff was selected as 20kHz which encompasses the bandwidth of the anti-aliasing filters for each recording . For each of the bandpass filtered signals , we next extracted the envelope . This was done by first computing the analytic signal: sA ( t ) =sband ( t ) +iH{sband ( t ) }=x ( t ) eiθ ( t ) where H{∙} is the Hilbert transform [13] . The temporal envelope is then obtained by taking the analytic signal magnitude x ( t ) =|sA ( t ) | . The envelopes were next passed through an antialiasing lowpass filter ( 250 Hz cutoff ) to limit the modulation content to the range of interest ( 0–250 Hz ) , down sampled by a factor DF ( see Table 1 ) , and scaled for unit standard deviation . An example speech waveform excerpt and the corresponding filtered envelope obtained with the above procedure are shown in Fig 1 ( Fig 1A , black = original sound waveform; Fig 1B and 1C , red = 250 Hz filtered envelope ) . Finally , we computed the AMPS of each animal group . The power spectral density of the envelope , x ( t ) , was estimated using a multi-taper spectral estimator ( pmtm . m MATLAB function , NFFT = 16384 , NW = 7/2 ) . This procedure generates a power spectral density estimate with nominal frequency resolution of ~0 . 1 Hz . An NFFT value of 16384 was used to analyze all of the data with the exception the periodic simulated envelope of Fig 2 ( magenta curves in Fig 2B and 2F ) . In order to achieve sufficiently high frequency resolution to resolve all of the envelope harmonics a value of NFFT = 262144 was used for this example . To test whether the envelope of isolated vocalizations contribute to the scale-invariant structure observed in vocalization sequences , we developed a stochastic vocalization sequence model consisting of a sequence of nonoverlapping rectangular pulses , pn ( t ) . Each pulse marks the beginning and end of isolated vocalizations . The vocalization envelope can be approximated as x ( t ) =∑n=1Npn ( t ) =∑n=1NAn∙rect ( t−tnDn ) ( 1 ) where n is the pulse number and rect ( ∙ ) is a unit amplitude rectangular pulse with start time zero and 1 s duration . The number of isolated vocalizations within the T second interval is N ≅ λT where λ is the average vocalization rate in units of vocalizations/s . To account for the vocalization-to-vocalization variability in the sequence , pulse amplitudes ( An ) , onset times ( tn ) and durations ( Dn ) are modeled as random variables . The envelopes from each vocalization sequence were fitted to the model of Eq 1 to assess how temporal sequence parameters ( vocalization peak amplitudes , durations and onset times ) contribute to 1/f structure . The fitting procedure consisted of two separate steps outlined in the following sections . This includes 1 ) a segmentation phase in which we detected and segmented the sequence into isolated vocalizations that stand out above the background noise level followed by 2 ) fitting the envelope from the segmented vocalization to rectangular pulses . In order to fit the vocalization sequence data for each animal group to the model of Eq 1 , we first segmented the data into segments that contain single isolated vocalizations . Since isolated vocalizations occur at relatively low rates [21 , 22] the envelopes of each vocalization sequence , x ( t ) , were initially filtered to a maximum frequency fm = 30 Hz with a 5-th order B-spline lowpass filter with continuously differentiable impulse response ( differentiable to 5th order ) as shown for a speech segment ( Fig 1A , black = original sound waveform; Fig 1B , blue = 30 Hz envelope ) . This 30 Hz lowpass filter is only applied during the vocalization segmentation phase and is used to identify sequence segments that contain isolated vocalizations ( consisting of an onset and an offset component ) . Envelope segments that contained both an onset and offset were identified if the envelope exceeded a designated threshold level ( Tx , Table 1 ) above the envelope of sequence segments containing background noise . A short 2 . 7 sec long noise segment from each vocalization was identified audio-visually and used to measure the noise variance for each recording . The threshold level was set to 30 standard deviations ( SD ) above the noise floor for all vocalizations except for the rat and speech sequence ( Tx = 10 SD ) which required a lower threshold to minimize false negatives ( i . e . , vocalizations not detected by the algorithm as identified audio-visually ) . Using this approach total 2957 vocalization segments were identified ( rat = 571 , mouse = 492 , bird = 518 , monkey = 590 , infant = 389 , speech = 801 ) . The model fitting procedure was performed on each isolated vocalization following the segmentation . During the fitting procedure , we used the signal envelopes that were lowpass filtered with a cutoff of 250 Hz ( Fig 1B , red , shown for the speech segment in Fig 1A ) . Although it is theoretically possible to fit the vocalization model sequence of Eq 1 directly to the vocalization sequence envelope , the large number of parameters that would be required in the optimization are prohibitive . For instance , for the crying infant vocalization sequence there are N = 389 detected vocalizations , which would require that the algorithm optimize for a total of 1167 ( 389x3 ) model parameters . We tested such a global fitting procedure using least-squares and were unable to achieve convergence because of the high parameter dimensionality . Instead , we optimized for each of the detected vocalization sequence segments , which individually requires only three parameters . That is , for each detected vocalization segment , xn ( t ) , we fitted a rectangular pulse , pn ( t ) , of variable start time ( tn ) , duration ( Dn ) , and peak amplitude ( An ) using least-squares optimization . The optimization was carried out for all of the detected segments in each sequence . The results of this fitting procedure are illustrated for brief speech segment ( Fig 1B and 1C , green ) . The model envelope accounts for the transient onsets and offsets that mark the beginning and end of vocalizations . It is not intended to model fast modulations that are also evident in the envelopes , such as those arising from periodic vocal fold vibration and which can be seen as a superimposed components ( red ) that ride on top of the slower vocalization envelope ( blue ) ( zoomed in view in Fig 1C ) . The optimal parameters for each segment were then combined into three time-series ( tn , Dn and An ) that were used to implement the full vocalization sequence model ( Eq 1 ) . For each of the vocalization AMPS , we empirically estimated the cutoff frequency , fc , where the AMPS transitions from a predominantly flat curve at low frequencies to a f−2 trend at higher frequencies . This was done by fitting the AMPS of each vocalization to a first-order lowpass spectrum model of the form Sxx ( f ) =C/ ( f2+fc2 ) where C and fc are free parameters to be determined . The estimated cutoff frequency was derived from the best fit solution of the first order model obtained numerically using least squares . In this section we derive a closed form solution for the AMPS of the stochastic vocalization sequence model of Eq 1 . The modulation power spectrum of the vocalization model is obtained by taking the long-term expectation of the Fourier Transform Magnitude: Sxx ( f ) =limT→∞1TE[X ( f ) X ( f ) *] where E[∙] is the expectation operator taken across the three random variables ( onset time , duration and amplitude ) , X ( f ) =I{x ( t ) }=I{∑n=1NAn∙rect ( t−tnDn ) }=∑n=1NAn∙sin ( πDnf ) πfe−j2πftn is the envelope Fourier transform ( I{∙} ) , and * represents the complex conjugate . The model AMPS is then obtained as Sxx ( f ) =limT→∞1TE[X ( f ) X ( f ) *]=limT→∞1TE[ ( ∑n=1NAn∙sin ( πDnf ) πfe−j2πftn ) ( ∑k=1NAk∙sin ( πDkf ) πfe+j2πftk ) ]=limT→∞1TE[∑n=1NAn2∙sin2 ( πDnf ) π2f2+∑n=1N∑k≠nAk∙An∙sin ( πDnf ) πfsin ( πDkf ) πfe+j2πf ( tk−tn ) ] . As will be illustrated subsequently the measured model parameters are largely independent and onset times are serially uncorrelated for the experimental data . This allows us to assume independence of the model parameters so that the second term inside the expectation approaches zero so that Sxx ( f ) =limT→∞1T∑n=1NE[An2∙sin2 ( πDnf ) π2f2]=limT→∞NTE[An2∙sin2 ( πDnf ) π2f2] . Since in the limiting case λ ≃ N/T and the random variables are approximately independent the AMPS simplifies as follows Sxx ( f ) =λ∙E[An2]∙E[sin2 ( πDnf ) π2f2]=λ∙ ( μA2+σA2 ) ∙E[sin2 ( πDnf ) π2f2] . Finally , under the assumption that the vocalization durations are uniformly distributed within the interval [T1 , T2] E[sin2 ( πDnf ) ]=∫p ( γ ) sin2 ( πγf ) dγ=1T2−T1∫T2T1sin2 ( πγf ) dγ=1 ( T2−T1 ) ∙2∫T2T11−cos ( 2πγf ) dγ=12∙1T2−T1∙[T2−T1−sin ( 2πT2f ) −sin ( 2πT1f ) 2πf] so that the AMPS is Sxx ( f ) =λ∙ ( σA2+μA2 ) 2∙π2f2∙[1−sin ( 2πT2f ) −sin ( 2πT1f ) ( T2−T1 ) ∙2πf]=λ∙ ( σA2+μA2 ) 2∙π2f2∙[1−T2T2−T1∙sinc ( 2πT2f ) +T1T2−T1∙sinc ( 2πT1f ) ] . Given that the experimental and model AMPS both have lowpass structure , we derived in closed form the vocalization model AMPS cutoff frequency in order to relate this AMPS parameter to the vocalization model parameters ( amplitude , duration and onset times ) . The vocalization model AMPS cutoff frequency ( fc ) is defined as the frequency where AMPS achieves half power ( - 3dB ) relative to the AMPS at zero frequency Sxx ( fc ) =12∙Sxx ( 0 ) , which for the model requires that the following equation be satisfied λ∙ ( σA2+μA2 ) 2∙π2fc2∙[1−sin ( 2πT2fc ) −sin ( 2πT1fc ) ( T2−T1 ) ∙2πfc]=λ∙ ( σA2+μA2 ) 6 ( T2−T1 ) [T23−T13] . An approximate solution is obtained by noting that for large fc > 1/2π ( T2 − T1 ) sin ( 2πT2fc ) −sin ( 2πT1fc ) ( T2−T1 ) ∙2πfc<1 ( T2−T1 ) ∙2πfc<1 . Considering this upper bound , the above equation is approximated as 12∙π2fc2≈16 ( T2−T1 ) [T23−T13] and solving for the cutoff frequency yields fc≈1π3∙ ( T2−T1 ) [T23−T13] . Finally , since μD = ( T1 + T2 ) /2 and σD2= ( T2−T1 ) 2/12 for a uniform distribution the cutoff can be expressed as fc≈1π1μD2+σD2=1π1E[Dn2] .
We explore which temporal cues contribute to scaling phenomena in vocalization sequences . We consider a stochastic model of vocalization envelope sequence , x ( t ) , containing three distinct forms of temporal variability ( Eq 1 , Materials and methods ) . The envelope of each vocalization sequence is approximated as a superposition of rectangular pulses each with a distinct onset time ( tn ) , pulse amplitude ( An ) , and duration ( Dn ) . Each parameter is modeled as a random variable to account for vocalization-to-vocalization variability in the sequence . Fig 2 illustrates how each of the model acoustic features contributes to the AMPS of natural vocalization sequences from an infant ( a-d ) and a rat pup ( e-h ) , respectively . Vocalization amplitudes , onset times , and duration parameters are obtained for each vocalization in the sequence by fitting the model ( a and e; red curve ) to the original sound envelope ( a and e; black curve ) and the AMPS of the model envelope is computed ( Fig 2B and 2F; see Materials and methods ) . Statistics for each of the estimated model parameters from the vocalization recordings is provided in Table 2 ( see Materials and methods for details ) . The model AMPS ( red ) has a lowpass shape and power-law scaling similar to the original vocalization sequence AMPS ( Fig 2B and 2F , black ) with an RMS error of 3 . 9 dB ( for frequencies between 1–100 Hz ) . Although the model follows the natural sound AMPS for low and intermediate modulation frequencies , it deviates at high modulation frequencies ( Fig 2B , >100 Hz for infant; Fig 2F , >40 Hz for rat pup ) . In humans , this model disparity is partly explained by periodic modulations generated by the vocal fold vibrations [12] that contribute to the perceived vocal pitch and , though critical for identifying speech source attributes such as gender , they are not essential for speech intelligibility [23] . This result indicates that our model captures much of the general AMPS shape of natural vocalization sequences , particularly the power-law scaling trend . By synthetically altering the model parameters we further explore how each temporal cue shapes the AMPS . First , we assess the contribution of vocalization amplitude variability by assigning a fixed amplitude to each model vocalization pulse ( Fig 2A and 2E , green ) while keeping all other parameters fixed . The pulse amplitudes are chosen so that the fixed amplitude model envelope and the original model envelope have matched variance . This manipulation has minimal effect on the AMPS ( Fig 2B and 2F , green ) since it maintains the lowpass shape and power-law scaling similar to the original vocalization sequence . Secondly , we manipulated the inter-vocalization intervals , defined as the time difference between consecutive vocalization onset times , Δtn = tn+1 − tn , to determine whether timing variability between vocalizations contributes to the power-law scaling . When we impose a constant inter-vocalization interval of 1 second ( a and e , magenta ) the modulation spectrum exhibits harmonic structure with 1Hz fundamental component that reflects the periodic structure of the inter-vocalization intervals . However , the resulting spectrum and the peak amplitude of the harmonics still follow the f−2 modulation spectrum trend ( b and f , magenta ) , which suggest that the exact structure of the inter-vocalization intervals are not the critical parameters accounting for this behavior . Thirdly , temporal variation in vocalization durations is explored by replacing the pulse model approximation of each natural vocalization with a Dirac impulse that has a fixed duration of zero seconds ( Fig 2A and 2E , blue ) . Removing the variation in vocalization duration results in a flat AMPS ( Fig 2B and 2F , blue ) that no longer exhibits scaling . The last manipulation conserves variations in the inter-vocalization intervals and amplitudes , indicating that these features alone are not sufficient to account for the lowpass trend with scaling at high frequencies whereas vocalization duration is critical . To further explore the impact of vocalization durations we synthetically manipulated the duration distribution to determine how it contributes to scaling . We replaced the empirically measured durations with samples drawn from either a uniform ( orange ) , exponential ( light blue ) , or gamma ( dark green ) distribution with matched mean and variance ( Fig 2C and 2G ) . As can be seen , the resulting AMPS is largely unaffected by the model distributions used as long as the vocalization durations have the same mean and variance ( Fig 2D and 2H; as described subsequently ) . The measured RMS error between the simulated model AMPS with different duration distributions and the actual AMPS for modulation frequencies between 1–100 Hz was relatively small ( between 3–4 dB for all of the distributions ) . This indicates that the AMPS shape is largely independent of the type of distribution used to model the vocalization durations . It is conceivable that scaling emerges due to serial correlations and co-variation between the vocalization amplitudes , durations , and intervals . We assess these possibilities by examining the statistical structure of these three acoustic parameters for the infant and rat pup ( Fig 3 ) . The joint duration-amplitude distribution ( Fig 3A and 3F ) is relatively compact and these parameters exhibit a significant but weak correlation ( infant , 0 . 11±0 . 04; rat , r = 0 . 49±0 . 05; mean±SE; t-test , p<0 . 01; see Table 3 for additional vocalization statistics ) . The autocorrelation for the duration and amplitude time series has impulsive structure , indicating minimal serial correlation for the infant and rat pup vocalization sequences ( infant , Fig 3C and 3D; rat pup , Fig 3H and 3I ) . Furthermore , the inter-vocalization intervals follow an approximately exponential distribution as expected for a Poisson point process ( Fig 3B and 3G ) , although there is a short latent period ( ~150 ms , infant; ~30 ms , rat pup ) in the interval distribution indicating a brief silent period between consecutive vocalizations . Inter-vocalization intervals are weakly correlated with the vocalization duration and amplitude parameters ( Table 3 ) . Finally , upon treating the vocalization onset times as a renewal point process , we find that these are uncorrelated as evident from the impulse structure of the point process autocorrelation ( Fig 3E and 3J ) . These analyses indicate that vocalization durations , amplitudes , and inter-vocalization intervals are distributed in a largely independent and serially uncorrelated fashion . To gain further insight on how each envelope parameter contributes to the scaling behavior in the AMPS , we derive the model AMPS in closed form by computing the power spectral density of the stochastic envelope model . Given that the estimated vocalization model parameters are weakly correlated ( Fig 3 and Table 3 ) , we assume independence of the model parameters to simplify the derivation . The model AMPS is ( Materials and methods , Vocalization model AMPS derivation ) Sxx ( f ) =λ∙E[An2]∙E[sin2 ( πDnf ) π2f2]=λ∙ ( μA2+σA2 ) ∙E[sin2 ( πDnf ) π2f2] ( 2 ) where E[∙] is the expectation operator , μA2 and σA2 are the amplitude mean-squared and variance , and E[An2]=μA2+σA2 is the second-order moment of An . This result demonstrates that although the rate of vocalizations ( λ ) and amplitude statistics ( μA2+σA2 ) both affect the overall AMPS by a multiplicative gain factor , they do not depend on f and therefore do not affect the AMPS shape . Instead , the AMPS shape is primarily determined by the distribution of vocalization durations ( term containing E[∙] ) . Since , as shown above , the exact duration distribution used has minimal impact on the AMPS shape ( Fig 2D and 2H ) we use a uniform distribution to simplify the analytic derivation . The AMPS is then evaluated in closed form as ( Materials and methods , Vocalization model AMPS derivation ) Sxx ( f ) =λ ( μA2+σA2 ) 2π2f2[1−T2sinc ( 2πT2f ) −T1sinc ( 2πT1f ) T2−T1] ( 3 ) Despite the simplifying assumptions , the analytic solution captures the general AMPS structure including the 1/f2 trend and the flat low frequency region for a human infant and rat pup vocalizations ( Fig 4; actual AMPS , black; simulated AMPS , red; analytic solution AMPS , dotted blue ) . Next , we evaluated the limiting AMPS behavior for these two regimes . For low frequencies ( f → 0 ) , it can be shown by applying L'Hospital's rule that: Sxx ( 0 ) =λ ( μA2+σA2 ) 3 ( T22+T1T2+T12 ) ( 4 ) which is the limiting value in the flat low frequency AMPS region observed in Fig 4 . By comparison , in the limiting case where the modulation frequency is large ( i . e . , f → ∞ ) : Sxx ( f ) =λ ( μA2+σA2 ) 2π2∙1f2 ( 5 ) so that the AMPS behaves as a power-law for high f with a power-law exponent of α = 2 . We find this dual regime lowpass structure is evident in all of the vocalization sequences examined ( Fig 4 ) . Although the model can deviate from the data as a result of vocalization production mechanisms not related to the temporal edges created by the initiation of isolated vocalizations ( e . g . , vocal fold vibration ) , in all cases the model captures the general lowpass structure . Furthermore , the model captures nearly all of the variability associated with the 1/f2 trend since the residual error spectrum lacks 1/f2 structure ( Fig 4G ) and all of the measured vocalizations sequence AMPS deviated from the simulated model by at most 3 . 9 dB ( RMS error between model and data for frequencies between 1–100 Hz ) . This suggests that temporal edges are the main acoustic features accounting for the general scaling behavior . Next , we explore the mechanism by which temporal edges in isolated vocalizations contribute to power-law scaling and the dual-regime structure . We start by noting that the vocalization sequence AMPS is precisely the average AMPS of individual vocalization envelopes if the vocalization onset times are serially uncorrelated . Considering the rectangular pulse vocalization sequence model ( Eq 1 ) , the AMPS of each rectangular pulse ( pn ( t ) ) is: Spnpn ( f ) =An2∙Dn2∙sinc2 ( πDnf ) =An2∙sin2 ( πDnf ) π2f2 ( 6 ) Thus , although isolated vocalizations contain both temporal onsets and offsets , which contribute to the 1/f2 behavior , on their own individual isolated vocalizations deviate from the 1/f2 trend . Based on Eq 6 the individual vocalization envelope power spectrum is approximated as a sinc2 ( ∙ ) function with a spectrum amplitude proportional to the pulse amplitude squared and bandwidth that is inversely related to the pulse duration . This is evident from the power spectra ( Fig 5B ) of three exemplar rectangular pulses ( Fig 5A ) taken from the speech ensemble . The spectrum of a single pulse has a lowpass structure with oscillatory side-lobes that deviate from the 1/f2 trend ( Fig 5B , black curves ) although the peak amplitude of the side-lobes precisely follows the 1/f2 trend ( blue curves ) . We propose that the observed dual-regime 1/f2 structure arises from the collective averaging across an ensemble of isolated vocalizations in a sequence . As can be seen in Fig 3 , isolated vocalizations have variable durations which consequently produce different notch and side-lobe configurations in the frequency domain ( Fig 5B ) . Upon averaging the spectrum of each vocalization , notches and side-lobes interfere and cancel producing the 1/f2 regime . In contrast , the sinc2 ( ∙ ) main lobes average constructively producing the flat AMPS regime at low frequencies . Thus , the dual-regime vocalization sequence AMPS behavior including the 1/f2 trend emerge naturally from the collective averaging across an ensemble of isolated vocalizations of variable durations . As demonstrated in the simulations of Figs 2 and 5 and the closed form model derivations , the dual-regime lowpass structure of the vocalizations sequence AMPS likely arises through the superposition of spectra from isolated vocalizations each with a bandwidth that is inversely related to the vocalization duration . To determine the relationship between vocalization duration distribution and the transition point for the 1/f2 regime in the vocalization sequence AMPS , we derive the solution for the half power or cutoff frequency ( fc ) of the model AMPS ( Materials and methods , Vocalization model cutoff frequency derivation ) . The analytic solution yields: fc≈1π∙1E[Dn2]=1π∙1μD2+σD2 ( 7 ) where μD and σD2 are the duration mean and variance . This result indicates that the vocalization duration statistics are the primary determinants of the fc . Specifically , fc is inversely related to the square root of the second order moment of the vocalization duration distribution . That is , vocalizations with a longer average duration will tend to have a lower fc values while vocalizations with shorter durations will tend to have larger fc values . This result is consistent with the results of ( Fig 2A , 2B , 2E and 2F; blue curves ) where we synthetically manipulated and set the vocalization model durations to zero . In such a case , vocalization pulses approach an impulse while the fc approaches infinity and only the flat region of the AMPS is observed . This mathematical formulation is a statistical variant of the uncertainty principle for a vocalization ensemble , which requires that the signal duration in the time-domain be inversely related to its bandwidth in the frequency-domain [13] . Finally , we examine whether the measured durations from natural vocalization sequences can be used to predict fc and therefore the transition point between the lowpass and 1/f2 regimes . As seen in Fig 6 , the fc estimated with our analytic model is correlated with the empirically measured fc for the animal vocalization recordings examined ( Fig 6A; log ( fc ) vs . log ( fc , model ) , Pearson r = 0 . 76±0 . 24 , mean±SEM; bootstrap t-test , p<0 . 05 ) . Furthermore , measured fc for the six recordings are inversely related to the experimentally measured second-order duration moment ( Fig 6B; log ( fc ) vs . log ( E[Dn2] ) , Pearson r = -0 . 76±0 . 24 , mean±SEM; bootstrap t-test , p<0 . 05 ) as predicted by Eq 7 ( Fig 5 , dotted line ) . This supports the idea that there is an inverse relationship between the vocalization durations and fc that manifests as a tradeoff in time-frequency resolution .
The results describe for the first time a single physical cue that universally accounts for scale invariant phenomenon in the envelope of natural vocalization sequences from several animal recordings . We find that the ensemble of temporal boundaries or edges for isolated vocalizations is the principal determinant of power-law scaling relationship . In addition , we find a systematic inverse relationship between the average vocalization duration and frequency at which scaling behavior initiates ( fc ) . Edges are responsible for the observed 1/f2 scaling region of the AMPS whereas the timing between consecutive on and off edges , which determine the vocalization duration , are critical in determining the flat region of the AMPS and the cutoff frequency ( fc ) . These findings thus provide a new conceptual framework for characterizing the temporal statistics of natural vocalized sounds in terms of definable temporal cues within ongoing sound sequences . For example , one can conceptualize vocalization elements such as words and phonemes in speech as acoustic objects formed by temporal edges in the sound envelope and our study indicates that these are primary determinants of the temporal statistics captured in the AMPS of vocalization sequences . Moreover , temporal edges are perceptually salient [9 , 24] and serve as temporal boundaries for grouping acoustic objects [25] . Although we have not extended our analysis to broader categories of sounds , other natural sounds [9–12] also exhibit scaling . The models and conceptual framework introduced here may have broad applicability as sound sequences and music in general are composed of transient and time-varying acoustic elements that can be coarsely modeled by onsets and offsets . In vision , the spatial arrangement of object boundaries and the distribution of object size in opaque natural images all contribute to scale invariance [4–6] . In an analogous fashion , we have shown that vocalization boundaries consisting of edges in the time-domain likewise contribute to scaling in the acoustic realm . Importantly , isolated vocalizations are not sufficient since the 1/f2 trend arises from the collective averaging amongst an ensemble of vocalization with variable durations ( Fig 5 ) . Yet , unlike for natural scenes where the object size distribution needs to follow a power-law relationship , scaling for natural sounds does not depend critically on the exact vocalization duration distribution as long as the distributions have similar means ( Fig 2 and closed form solutions ) . Furthermore , we point out that vocalization sequence onset times and amplitudes statistics are not critical to this result as determined from the closed form solutions of the model and demonstrated in Fig 2 , where the model parameters where perturbed to constant values ( periodic case for onset times and constant amplitude ) . Thus , the combined findings from the model and empirical perturbations provide strong evidence that the temporal edge boundaries in vocalizations are responsible for 1/f2 phenomenon . In our analysis , we considered isolated vocalization sequences which have well-identified vocalizations and well-defined temporal boundaries . Whether similar results apply to more complex acoustic scenarios including natural soundscapes consisting of mixtures of vocalizations that are superimposed is unclear and needs to be determined . This is plausible given that images containing mixtures of translucent objects can also exhibit scale invariance [4 , 5] . Previous works have demonstrated that although scaling is observed in natural environmental sounds , such sounds tend to have a scaling exponent that is somewhat lower than for vocalizations ( scaling exponent closer to α = 1 ) [10 , 11 , 26] . One plausible hypothesis that needs to be considered is that background sounds often consist of mixtures of isolated sound , each of which has a well-marked onsets and offsets , so that the superposition of isolated acoustic objects could create phase distortions at the sound boundaries that distort temporal edges and ultimately have a whitening effect on the envelope AMPS , thus reducing the scaling exponent . Future studies need to explore how and if our findings can be generalized into a theoretical framework that applies to an even broader range of natural and man-made sounds . Although the results provide a concise explanation for the 1/f2 scaling region that is linked to the temporal boundaries in vocalized sounds our model is not intended to account for other forms of scaling or features of the AMPS . Future studies and models are needed to further elucidate the acoustic generation mechanisms responsible for distinct regions of the AMPS of natural vocalized sounds . For instance , 1/f scaling has been previously described for very low modulation frequency ( <0 . 1 Hz ) for speech and music [8] . One plausible explanation for this phenomenon is that inter-vocalization statistics in sound sequences , such as for speech , have self-similar fractal structure at very long time scales [27] that may be responsible for 1/f scaling for very low modulation frequencies . Our model also is not intended to account for other features of the modulation spectrum , for instance the presence of periodic modulations created through vocal vibration and which are clearly evident in our speech envelopes and AMPS ( Fig 1 ) . These fast-periodic modulations are visible in the speech and infant vocalizations and show up as an additive modulation component in the AMPS ( positive AMPS deflection above the expected model results ) . A recent study observed the presence of peaks in the modulation spectrum of speech and music in the vicinity of 3–5 Hz [28] and lacked 1/f2 structure described here . This difference is due to the fact that the calculation of the modulation spectrum in that study used modulation filters with logarithmic bandwidths that mimic neural modulation tuning functions [12] to estimate the modulation power . Applying such modulation filters magnifies the output power by a factor proportional to f , so that the flat region of the AMPS we describe increases proportional to f and the region containing the 1/f2 trend decreases proportional to 1/f . Consequently , a peak is observed in the modulation spectrum within the vicinity of the cutoff frequency ( fc ) where the transition between the flat and 1/f2 behavior is observed in our model . We have confirmed the observations of Ding et al . by estimating modulation spectrum with octave band filters or alternately multiplying the modulation spectrum by f as described ( S1 Fig ) . In both cases , the resulting modulation spectrum contain a primary peak in the vicinity of ~3 Hz as observed by Ding et al . , but we also observe a secondary peak within the vicinity of 100–300 Hz where vocal fold vibration is prominent . Ding et al did not observed such a peak because they characterized the modulation power spectrum only up to 32 Hz . The results have a number of implications for theories of coding by the brain since auditory neurons are exquisitely sensitive to temporal transitions with millisecond precision [29–31] and have been shown to produce an efficient neural representation that equalize the modulation power of natural sounds [12 , 14] . Similar strategies have been proposed in vision where neurons through edge detection equalize or “whiten” the spectrum of natural images enabling an equitable use of neural resources [3] . Mechanistically , two distinct temporal coding mechanisms could contribute to such efficient representation in audition . First , auditory neurons have excitatory-inhibitory ( on-off ) responses to temporal edges that effectively perform a smooth temporal derivative operation on the sound envelope [32–35] . In the time domain , this could facilitate temporal edge detection for important information bearing acoustic temporal elements , analogous to edge detection in vision [3 , 7] . In the frequency domain , such temporal derivative operation has a transfer function squared-magnitude H2 ( f ) = 4π2f2 that opposes and precisely cancels the 1/f2 scaling of natural sounds thus whitening the spectrum . Secondly , power equalization could be partly achieved through modulation filter bandwidth scaling as previously observed for auditory midbrain neurons [12] and perceptually [36] . For both neurons and perception , modulation filter bandwidths increase proportional to f . This bandwidth scaling magnifies the output power by f , partly canceling the 1/f2 power trend observed for natural sounds [12] ( as shown in S1 Fig ) . In combination , temporal edge detection and bandwidth scaling could provide mechanisms to equalize modulation power in vocalizations allowing for efficient information transfer and coding , analogous to principles in vision . The findings are also relevant for sound coding and hearing technologies . For instance , the stochastic framework could be used to improve coding , compression , and sound recognition algorithms . The findings could further be used to improve algorithms to enhance detection of transient sound elements [37] in order to facilitate recognition in hearing aid , cochlear implant , and other assistive hearing technologies .
|
The efficient coding hypothesis posits that the brain encodes sensory signals efficiently in order to reduce metabolic cost and preserve behaviorally relevant environment information . In audition , recognition and coding depends on the brain’s ability to accurately and efficiently encode statistical regularities that are prevalent in natural sounds . Similarly , efficient audio coding and compression schemes attempt to preserve salient sound qualities while minimizing data bandwidth . A widely observed statistical regularity in nearly all natural sounds is the presence of scale invariance where the power of amplitude fluctuations is inversely related to the sound amplitude modulation frequency . In this study , we explore the physical sound cues responsible for the scale invariant phenomenon previously observed . We demonstrate that for animal vocalizations , including human speech , the scale invariant behavior is fully accounted by the presence of temporal acoustic edges that are largely created by opening and closing of the oral cavity and which mark the beginning and end of isolated vocalizations . The findings thus identify a single physical cue responsible for the universal scale invariant phenomenon that the brain can exploit to optimize coding and perception of vocalized sounds .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"acoustics",
"vocalization",
"linguistics",
"engineering",
"and",
"technology",
"signal",
"processing",
"social",
"sciences",
"neuroscience",
"animal",
"signaling",
"and",
"communication",
"animal",
"behavior",
"computational",
"neuroscience",
"bandwidth",
"(signal",
"processing)",
"zoology",
"coding",
"mechanisms",
"behavior",
"bioacoustics",
"speech",
"physics",
"speech",
"signal",
"processing",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"acoustic",
"signals"
] |
2018
|
Origins of scale invariance in vocalization sequences and speech
|
Landscapes exhibiting multiple secondary structures arise in natural RNA molecules that modulate gene expression , protein synthesis , and viral . We report herein that high-throughput chemical experiments can isolate an RNA’s multiple alternative secondary structures as they are stabilized by systematic mutagenesis ( mutate-and-map , M2 ) and that a computational algorithm , REEFFIT , enables unbiased reconstruction of these states’ structures and populations . In an in silico benchmark on non-coding RNAs with complex landscapes , M2-REEFFIT recovers 95% of RNA helices present with at least 25% population while maintaining a low false discovery rate ( 10% ) and conservative error estimates . In experimental benchmarks , M2-REEFFIT recovers the structure landscapes of a 35-nt MedLoop hairpin , a 110-nt 16S rRNA four-way junction with an excited state , a 25-nt bistable hairpin , and a 112-nt three-state adenine riboswitch with its expression platform , molecules whose characterization previously required expert mutational analysis and specialized NMR or chemical mapping experiments . With this validation , M2-REEFFIT enabled tests of whether artificial RNA sequences might exhibit complex landscapes in the absence of explicit design . An artificial flavin mononucleotide riboswitch and a randomly generated RNA sequence are found to interconvert between three or more states , including structures for which there was no design , but that could be stabilized through mutations . These results highlight the likely pervasiveness of rich landscapes with multiple secondary structures in both natural and artificial RNAs and demonstrate an automated chemical/computational route for their empirical characterization .
RNAs are deeply involved in gene expression , gene regulation , and structural scaffolding and are forming the basis of novel approaches to control these processes [1–3] . Several of RNA’s natural and engineered roles rely on its ability to fold into and interconvert between multiple functional structures . Ribozymes , riboswitches , and protein-complexed RNAs transition between several states to detect and respond to small molecules and other macromolecules; to proceed through numerous steps of RNA splicing reactions; to initiate , catalyze , and proof-read protein translation; to activate logical circuits in cells; and to package , release , and replicate RNA viruses [4–9] . The number of structures and equilibrium fractions that constitute these ‘dynamic structure landscapes’ are linked to the biological function of the RNA ( Fig 1a ) . Rationally dissecting and re-engineering these landscapes depends on knowledge of the alternative states of an RNA’s structural ensemble [10 , 11] . Empirical portraits of such landscapes are missing for the vast majority of natural and engineered RNAs and it is unclear whether a rich multi-state landscape is a property specially selected by evolution or an intrinsic feature of RNA that can arise without explicit design or selection . Watson-Crick RNA secondary structure landscapes have been computationally predicted using dynamic programming techniques for decades [12 , 13] and many RNAs are predicted to form multiple structures at equilibrium , with ‘non-native’ helices reaching populations of 25% or greater [12–14] . In some cases , the alternative structures—and not the dominant structure—harbor motifs that are recognized by protein or small molecule partners [15–17] . However , difficulties in treating non-canonical interactions render these predictions inaccurate . Indeed , some studies have suggested that conformational switches are special hallmarks of biological function rather than an intrinsic feature of generic RNA sequences [18–21] . Unfortunately , the few experimental techniques that can validate or refute multi-state models are costly and difficult . For example , single molecule methods have been successful at revealing rarely populated RNA states but have not provided enough information to infer their structures [22] . Powerful insights have come from advances in nuclear magnetic resonance ( NMR ) spectroscopy [23] but require focused technical expertise , expensive infrastructure , and RNA targets with limited structural heterogeneity . In contrast , RNA chemical mapping , or footprinting , is a simple class of techniques that can achieve single-nucleotide resolution structural data for any RNA [24 , 25] . For RNAs with multiple states , however , these chemical mapping data give ensemble averages over all states , leading to a dramatic loss of information compared to what would be needed to resolve the RNA’s dynamic structure landscape and to make testable predictions ( Fig 1b , see also refs . [20 , 26] ) . While developing experiments that couple systematic mutagenesis with chemical mapping ( mutate-and-map , or M2 ) , we observed that single mutations can produce dramatic changes in chemical mapping data throughout an RNA sequence , with several mutations often giving the same alternative pattern [27 , 28] . We hypothesized that these perturbed patterns correspond to the reweighting of the structural ensemble of the RNA so that alternative component states dominate the chemical mapping data ( Fig 1c ) . We have recently shown that such alternative structures can be inferred after expert inspection of M2 measurements and extensive compensatory rescue studies in an E . coli 16S ribosomal RNA domain [29] . Reasoning that such landscape dissection might be fully automated through the use of blind source separation algorithms , we have now developed an analysis framework called the RNA Ensemble Extraction From Footprinting Insights Technique ( REEFFIT; see Fig 1c ) . Here , we sought to use M2-REEFFIT to determine whether complex landscapes could arise in artificial RNAs without explicit natural selection or design . Prior techniques for detecting alternative structures , including covariation-based methods [19] and the recent RING-MaP method [20] , have not been benchmarked in cases with well-characterized landscapes and may be biased towards or against alternative structures . Therefore , we developed three tests for M2-REEFFIT landscape dissection . First , a benchmark on simulated M2 data for 20 natural non-coding RNA sequences provided ‘gold standard’ reference results allowing unambiguous assessment of accuracy . Second , we applied M2-REEFFIT to four experimental test cases involving biological and artificial RNAs that had been previously characterized in detail by NMR or chemical mapping . For these cases , M2-REEFFIT recovered the landscapes defined through prior expert analysis . Third , we developed a validation approach based on stabilizing and experimentally testing predicted structures by multiple mutations . With these benchmarks and methodological developments , we used M2-REEFFIT to demonstrate that an imperfectly designed riboswitch and a randomly generated RNA sequence each form at least three structures , including states that could not be predicted from computational modeling alone .
To leverage the signals of stabilized alternative structures present in M2 measurements , we developed a new analysis framework , REEFFIT , to simultaneously infer multiple structures and their population fractions across the mutant RNAs ( see Methods ) . We first applied M2-REEFFIT to infer structure landscapes for an in silico benchmark of 20 sequences drawn from the Rfam database of non-coding RNA molecules . We simulated M2 data by randomly choosing up to 200 suboptimal structures in the ensembles of the wild type sequence and of all variants that mutate a single nucleotide to its complement , as are typically probed in M2 experiments . We then simulated these structures’ respective SHAPE reactivity profiles using known reactivity distributions . To mimic inaccuracies in available energetic models , we re-weighted the ensemble randomly by introducing Gaussian noise ( mean of 0 kcal/mol and standard deviation of 1 kcal/mol ) to the predicted free energies . The resulting landscapes ranged from single dominant structures ( e . g . cases RF00027 and RF00301 ) to more complex scenarios with two or three states ( e . g . cases RF01300 and RF00051 ) to multi-state RNAs with more than three states present at non-negligible fractions ( e . g . cases RF01274 and RF01125; Fig A in S1 Text gives five examples examined in detail in the Supporting Results , Fig B in S1 Text gives results for the entire benchmark , and the figures in S2 Text contains detailed Figs for all benchmark cases ) . Over the entire 20-RNA benchmark , REEFFIT was able to consistently detect the presence of dominant and alternative helices ( Supporting Results and Figs A-D in S1 Text ) . We set a criterion for helix detection that the fitted population should be larger than the population error estimated from bootstrapping , i . e . the signal-to-noise ratio should be greater than one . With this criterion , REEFFIT achieved 94 . 6% sensitivity over helices present with at least 3 base pairs and at least 25% population , corresponding to a false negative rate ( FNR ) of 5 . 4% . The false discovery rate ( FDR ) was low as well , at 9 . 7% ( see Table 1 and Table A in S1 Text; full precision-recall curves given in Fig B in S1 Text ) . Without data , the error rates were substantially worse , by three-fold and two-fold , respectively ( FNR of 18 . 0% and FDR of 23 . 3% ) . Values for base-pair-level error rates were similar to helix-level error rates ( Table B in S1 Text ) . Error rates for wild type sequences alone ( i . e . , excluding single-nucleotide mutants ) were higher but also showed a strong improvement in REEFFIT ensembles compared to ensembles modeled with no data ( FNR of 15 . 2% , FDR of 15 . 2% for REEFFIT; FNR of 24 . 5% and FDR of 25 . 4% without data , see Table C in S1 Text ) . Use of only wild type data and no mutants , as would be carried out in conventional chemical mapping measurements , gave high error rates similar to landscapes modeled without data ( FNR of 22 . 7% and FDR of 25% , see Table D in S1 Text and an example in Supporting Results and Fig D in S1 Text ) , confirming the necessity of M2 data for accurate landscape dissection . Further systematic checks and use of ViennaRNA [30] initial models and benchmark evaluation using helix-wise RMSD are given in the Supporting Results in S1 Text and Tables E , F , and G in S1 Text . Success in the above in silico benchmark suggested that REEFFIT would accurately recover prior analyses of experimental M2 data sets , such as the MedLoop RNA ( Fig 2a–2d ) . This RNA was previously designed to exhibit a single , stable 10 base pair helix with a 15 nucleotide loop ( MLP-A , Fig 2d ) [27] . A few mutations were observed to give a clearly distinct chemical mapping pattern and predicted computationally to fold into an alternative structure ( MLP-B in Fig 2a; see e . g . G4C in M2 data ) but not expected to be strongly populated in the wild type sequence . Automated REEFFIT modeling of the MedLoop M2 data recovered both the dominant structure MLP-A and the alternative state MLP-B , and bootstrapped uncertainties gave bounds on the frequency of MLP-B in the wild-type sequence ( 4±4% , see mutant-wise state fractions in Fig Fa-Fc in S1 Text ) . Thus , REEFFIT was able to explain rearrangements of mutants into an alternative state , but at the same time did not over-predict its presence in the wild-type sequence ( Fig 2c ) . As a more difficult test , we applied the algorithm to the 126–235 region of the E . coli 16S ribosomal RNA . In protein-bound ribosome crystals , this domain forms a four-way junction with helices P1a-c , P2a-b , P2* , P3 , and P4a-b , but its solution state has been controversial [29] ( Fig 2e–2h ) . Conventional SHAPE-guided analysis suggested loss of P2a and P4 , and formation of alternative helices alt-P1d and alt-P4 [31] ( Fig 2h , green structure ) . However , M2 and compensatory rescue experiments gave no evidence for the SHAPE-based model but instead recovered the dominant solution structure to be the holo-ribosome conformation , except for a register shift in P4a ( called shift-P4a ) ; the crystallographic P4 register was also detected as a 20±10% ‘excited state’ [29] . In agreement with this detailed analysis , automated REEFFIT analysis of the 126–235 RNA M2 data also returned helices P1a-c , P2* , P2b , and P3 with high population fractions ( >80% , Fig 2g ) . Importantly , REEFFIT recovered an admixture of P4 ( 21±16% ) and shift-P4 ( 60±23% ) in a 3:1 ratio , in agreement with prior analysis ( Fig 2g , magnification ) ; RNAstructure calculations or use of wild type SHAPE data alone assigned negligible probability or highly uncertain population fractions to these helices , respectively ( Fig 2g ) . The population of alternative helices alt-P1d and alt-P4 were found to have low populations and high errors ( see Supporting Results in S1 Text ) . Further , a refined REEFFIT analysis including data for compensatory rescue double mutants recovered , with conservative error estimates , prior expert analysis ( see Supporting Results and Fig E in S1 Text ) , illustrating the automation of modeling of even a complex RNA structure landscape . Encouraged by REEFFIT’s performance in previous test cases for chemical mapping , we sought tests involving fully independent experimental characterization by NMR . We first investigated a bistable RNA sequence whose landscape was dissected by Höbartner and colleagues by decomposing its NH…N 1H NMR spectrum into a weighted sum of two states forming different hairpins ( here called BST-A and BST-B ) [32] . As expected , M2 measurements gave clear evidence for two distinct chemical mapping profiles , reflecting the bistable nature of this RNA ( Fig 2i–2l ) . One profile , consistent with BST-A appeared in variants with mutations in the 5´ end of the sequence , such as U4A . Another profile , consistent with BST-B , was revealed by mutations in the 3´ end such as C23G , which would destabilize the BST-A hairpin ( Fig 2i ) . Both structures were recovered by REEFFIT analysis , with population weights of 73 ± 11% and 26 ± 9% , for BST-A and BST-B respectively , in agreement with the previously reported fractions measured by NMR ( 70 ± 5% and 30 ± 5% , respectively ) and correcting the erroneous weights predicted with no data ( 99% and <1% , respectively ) . We then challenged M2-REEFFIT with a more complex test case: a Vibrio vulnificus adenosine deaminase ( add ) mRNA riboswitch ( Fig 3a–3d ) that , in response to the ligand adenine , exposes start site segments ( Shine-Dalgarno sequence and/or AUG start codon ) to promote mRNA translation . In a detailed NMR study , spectra in ligand-free conditions fit well to a model with two states , apoA ( ~30% , with helices P1 , P2 , P3 , P4 , and P5 ) and apoB ( ~70% , with P1B , P2B , P3 , an extended P4 called P4B , and P5 ) . Addition of adenine ligand resulted in spectra dominated by a state holo with P1 , P2 , P3 , and P5 and perturbed chemical shifts consistent with adenine binding to the aptamer [17] . While three states gave the simplest model for the prior data , more complex multi-state models were consistent as well and would be generally predicted from RNA secondary structure calculations [12 , 13] . We focused on whether M2-REEFFIT could recover the base pairs of P1 , P2 , P1B , P2B , P3 , P4 , P4B , and P5 , which were unambiguously determined through NOE spectroscopy and model construct comparisons . Even in the absence of adenine , our M2 measurements of the RNA suggested the presence of at least two distinct structures that protect or expose the mRNA start site ( AUG at nts 120–122 ) and , in an anticorrelated manner , expose or protect segments in the aptamer region ( e . g . , nts 53–60 and 66–72 ) , respectively ( M2 data in Fig 3a ) . In addition to observing these different states upon mutation , addition of 5 mM adenine induced clear changes in the M2 data , including protections in the aptamer and consistent exposure of the AUG start codon in most mutants ( M2 data in Fig 3a , bottom; ) . While consistent with this region’s unpaired status without and with ligand in prior NMR studies , our studies indicate a more dramatic switch in this region for more complete add riboswitch constructs and will be reported elsewhere . REEFFIT analysis gave excellent fits to these add riboswitch M2 data ( Fig 3b ) , automatically detecting the presence of P1 ( 14±10% ) , P2 ( 37±15% ) , P3 ( 86±18% ) , P4 without extension ( 69±10% ) , the extension P4B at lower population ( 36±6% ) , and P5 ( 94±18% ) in the absence of adenine , in agreement with the apoA and apoB model from NMR ( Fig 3c–3d ) ; and recovering a holo state with P1 , P2 , P3 , P4 , and P5 dominating in 5 mM adenine ( see Supporting Methods in S1 Text for treatment of ligand-bound structures ) . In ligand-free conditions , the REEFFIT analysis also gave several alternative helices in the P1/P2 region , including P1B and P2B ( 20±7% ) , consistent with the NMR-detected features in the apoB structure . The M2 data were critical in making these detections; RNAstructure calculations and use of wild type SHAPE data alone assigned negligible ( <5% ) probability to P1 , P1B , P2B , and the possibility of P4B shortening to P4 ( Fig 3c–3d ) . Coarse clustering of REEFFIT structures returned four states in which the NMR-modeled apoA was recovered as the cluster medioid of one state , ADD-A , and apoB was recovered as a medioid in another , ADD-B , albeit with an additional helix , P4B ( Fig 3c ) . Structures with alternative helices to P1 , P2 , P1B , and P2B in the 5´ region clustered into states ADD-C and ADD-D . The population fractions of these helices , as well as a set of P6 and P7 helices not detected in NMR experiments , were low; these features were appropriately flagged as uncertain from bootstrapping analysis ( e . g . , 21±15% for the most populated helix of this kind , P6 ) . Overall , the M2-REEFFIT results successfully recovered NMR-detected helices for this adenine riboswitch sequence , including heterogeneous structure in the 5´ region , dynamics in the P4 region , and rearrangements on adenine addition . After validation of M2-REEFFIT on diverse computational and experimental test cases , we used the method to estimate whether complex landscapes might arise in artificial RNA sequences without explicit design or selection . First , we analyzed the folding landscape of an imperfect RNA switch , ‘Tebowned’ , that was designed to convert between two states upon flavin mononucleotide ( FMN ) binding ( Figs 4 and 5 ) in early rounds of a riboswitch design puzzle in the Eterna massive open laboratory [33] . While the chemical mapping pattern of this RNA changed upon binding FMN , the measurements for the unbound state did not match the desired unbound structure , particularly near nucleotide A30 ( red arrows in Fig 4a ) ; this region should have been paired but instead was measured to be reactive . A priori , we could not distinguish whether this discrepancy was due to an incorrect balance of the two target bound and unbound structures or if there were other unexpected structures involved . To elucidate the discrepancy , we acquired M2 data for the Tebowned RNA . We first used REEFFIT to fit the M2 data using only the two desired structures , but this fit did not capture several features , such as the exposure at A30 ( Fig G in S1 Text ) . In contrast , a global ensemble fit adequately captured these features ( Fig 4a–4b , and Fig Gb in S1 Text ) . REEFFIT automatically clustered the modeled ensemble into three states: TBWN-A ( 56 ±16% ) and TBWN-B ( 27 ± 12% ) , matching the desired switch structures , and a third structure TBWN-C present at 17 ± 11% fraction ( red arrows in Fig 4a and Fig Ha-Hc in S1 Text ) . The unexpected state TBWN-C exhibits an apical loop around nucleotide A30 , explaining the observed discrepancies in this region for the wild type RNA , and harbors a purine-rich symmetric loop that may be significantly stabilized compared to the energy assumed in current nearest-neighbor models [34] . Other helices were discovered to be populated at non-negligible fractions in the analysis , but were deemed uncertain ( signal-to-noise ratios less than 1 ) from bootstrapping analysis . The REEFFIT-inferred populations of TBWN-B and TBWN-C were low ( <30% ) . Therefore , to further test the presence of at least three states , we sought to compare their modeled REEFFIT reactivity profiles to actually measured reactivities for these states . To achieve this comparison , we designed mutants that specifically stabilized helices in each of the TBWN-A , TBWN-B , and TBWN-C structures ( Fig 5a–5c ) . For each state , the reactivities of these state-isolating mutants agreed with each other within experimental error and were approximated well by REEFFIT’s predicted profile , providing independent confirmation of the modeled ensemble , including the unanticipated third state TBWN-C . Additional evidence for the accuracy of REEFFIT’s predictions was revealed by each of the stabilizing mutants’ FMN binding affinities: the TBWN-B and TBWN-A/TBWN-C mutants enhanced and worsened ligand binding , respectively , as expected ( Fig 5d , and Supporting Results and Fig I in S1 Text ) . We emphasize that the TBWN-C state would have been difficult to propose and then validate without automated REEFFIT analysis , given its negligible predicted population in secondary structure prediction calculations without M2 data ( Fig 4c ) . As a second test case with a previously unknown structural ensemble , we tested whether randomly generated or scrambled RNA sequences tend to fold into multiple disparate structures at equilibrium—a long-standing hypothesis fundamental to understanding RNA evolution , put forth by several in silico studies [35 , 36] and an experimental study that could not deconvolve the structures [37] . We carried out M2-REEFFIT for a randomly generated sequence , called here the M-stable RNA ( Fig 6 ) . Based on simulations , the structural ensemble of the construct was expected to consist mainly of a simple hairpin ( P1 in MST-A in Fig 6d , see top triangle of the Fig 6c ) but with at least two other structures becoming more stable than MST-A upon single mutations . The experimental M2 measurements were complex , with different mutants giving disparate protection patterns even in segments that appeared highly reactive ( and seemingly unstructured ) in the wild type RNA . As a first check on the number of states , REEFFIT fits assuming only 2 or 3 states missed many features observed in the data , including extended segments of changed chemical reactivity in several mutants ( Fig J in S1 Text ) . However , the REEFFIT global ensemble fit successfully modeled the M-stable data and suggested an ensemble with many more weakly populated helices than RNAstructure’s estimate ( compare bottom and top halves of Fig 6c ) . For visualization , we clustered these heterogeneous component structures into three states , MST-A , MST-B , and MST-C ( see Fig 6d–6f and Fig Hd-Hf in S1 Text ) . Analogous to the case of the Tebowned switch , we tested the REEFFIT prediction of these alternative states by designing mutations to stabilize the medioid structures of each cluster ( Fig 6d–6f ) . These mutants gave reactivities in agreement with predictions for MST-A and MST-C , supporting the inference of those structures . For MST-B , the state-stabilizing mutants gave reactivity profiles that did not exactly match each other , suggesting residual heterogeneity of structure; the profiles were nevertheless closer to the REEFFIT-predicted MST-B profile than wild type reactivities . These results corroborate the M2-REEFFIT model that the M-stable random RNA has a complex landscape with at least three structures , and likely significantly more heterogeneity , detectable upon unbiased nucleotide mutation .
The structure landscapes of natural and newly designed RNA molecules underlie their biological behaviors , but these landscapes’ complexities are largely uncharacterized [11] . Current experimental techniques used to probe these ensembles at nucleotide resolution require significant infrastructure investment and expert intuition . We have presented M2-REEFFIT , an unbiased strategy based on readily acquired chemical mapping measurements that detects dominant and alternative states of RNA structure landscapes in the ensemble perturbations induced by single-nucleotide mutations , with conservative estimates based on bootstrapping . Due to the challenges inherent in estimating a full ensemble of secondary structures rather than a single best-fit model , we have invested significant effort into benchmarking M2-REEFFIT and its uncertainty estimates . We confirmed the accuracy of REEFFIT and its uncertainty estimates in M2 simulation data with known ‘ground truth’ landscapes and in experimental data of RNAs whose landscapes were previously studied by chemical mapping or NMR experiments . We then applied REEFFIT to investigate if artificial sequences could present complex landscapes of alternative structures without specific design in two model systems whose ensemble behaviors were refractory to prior tools . We first traced a structural discrepancy in an artificial flavin-mononucleotide-binding switch to a significant population of an unexpected state . We then tested in silico predictions for the complex structural landscape of random sequences by obtaining the first experimentally derived landscape model of a randomly generated sequence , the M-stable RNA . For all cases , we tested REEFFIT’s predictions through independent experiments , including comparison to independent experimental methods and effects of stabilizing mutations and ligand binding . Beyond the inference of structural landscapes of single sequences , the high-throughput nature of M2-REEFFIT makes inference of the landscapes of multiple sequences related through their presumed or engineered function possible . For example , when dissecting structural features of sequences selected for binding efficiencies to molecular partners through in vitro selection or when analyzing the role of conserved structural motifs in diverse RNA sequences , a full portrait of their landscapes may reveal the interplay of several states that is critical for function . As an example , one of our experimental test cases , the Tebowned FMN switch , illustrates that information relevant for function can be extracted from knowing the full landscape of the RNA . In this case , we have a ‘functional’ readout of a structural motif ( the FMN binding motif ) that is present at different fractions in different mutants of the M2 experiment . A full view of the ensemble yields an extra state , TBWN-C , whose interplay with the state that contains the binding motif , TBWN-B , is necessary for understanding FMN binding of the riboswitch and its variants . Currently , uncertainties in the reactivities and energies of RNA motifs lead to RMSD errors in M2-REEFFIT state population fractions on the order of 10–15% , rendering the detection of states with lower population fractions difficult; these uncertainties may become poorer for longer RNA domains . Nevertheless , we expect that rapidly growing databases of rigorously standardized reactivity data [38 , 39] and of energetic parameters [40] for diverse RNA motifs will reduce these uncertainties . Furthermore , M2 experiments performed with other chemical probes , such as dimethyl sulfate , should provide powerful cross-validation data sets for testing inferred landscapes . Through the presented and related chemical mapping technologies , we therefore expect to have more routine visualization of the rich structural landscapes that appear to be pervasive in both functional RNAs and the generic sequences from which they evolve .
The MedLoop RNA mutate-and-map data were obtained as part of the experimental pipeline of the Eterna massively parallel open laboratory [27] , [41] . The MedLoop RNA and its mutants were generated through in vitro transcription of a pool of DNA constructs purchased from CustomArray . The RNA was probed with 1-methyl-7-nitroisatoic anhydride ( 1M7 ) in a folding buffer ( see Table I in S1 Text for detailed folding conditions ) using the MAP-seq protocol and sequenced . in a MiSeq sequencer . The resulting reads and were analyzed with the MAP-seeker software [41] . The Bistable hairpin , Tebowned switch , and M-stable RNAs as well as their respective complementary single-nucleotide mutants were constructed using PCR assembly , in vitro transcription , and probed with 1M7 as described previously [42] . Briefly , an assembly consisting of at most 60-nucleotide primers was designed to synthesize an in vitro transcription sample by PCR . DNA was purified with AMPure XP beads ( Agencourt , Beckman Coulter ) and in vitro transcribed for 3 hours . The resulting RNA was purified with AMPure XP beads , heated for 3 minutes at 90°C , cooled at room temperature for 15 minutes , and folded in folding buffer at room temperature for 1 hour ( see Table I in S1 Text for detailed folding conditions for each RNA ) . Because we sought to probe the ensemble of the RNA with minimal interference from the 3´ unpaired sequence that we use as the primer binding site , we folded the RNA in the presence of the fluorescent primer attached to the oligo ( dT ) beads ( Ambion ) that we regularly use for purification . Folding in this condition sequesters any additional single stranded regions that may interfere with our sequence of interest . The RNA was then subjected to 1M7 mapping ( 5 mM final concentration ) , purified with the oligo ( dT ) beads , and reverse transcribed for 30 minutes at 42°C . Unmodified RNA controls were also included in the experiment . RNA was then degraded using alkaline hydrolysis and cDNA was purified , eluted in Hi-Di Formamide spiked with a fragment analysis ladder ( ROX 350 standard , Applied Biosystems ) , and electrophoresed in an ABI 3150 capillary electrophoresis sequencer . The add adenine riboswitch M2 data was obtained similarly but folded under different conditions , in the NMR buffer of the previous study ( 50 mM KCl and 25 mM K3PO4 , 10 mM MgCl2 , pH 6 . 5 ) with or without 5 mM adenine for bound and unbound conditions , [17] For NMR folding conditions , we adjusted the 1M7 incubation time to 15 minutes instead of 3 minutes to account for the low pH . For the Tebowned FMN titrations , dimethyl sulfate ( DMS ) was used in lieu of 1M7 since it yields a readily seen signal change across FMN concentrations [33] . Electrophoretic traces were aligned , baseline subtracted , and normalized with the HiTRACE MATLAB toolkit [43] . 1M7 modification traces were quantified , background subtracted , and corrected for attenuation using 10X dilutions , the unmodified controls , and the pentaloop hairpins added at the ends of the constructs as reference [44] . For the Tebowned RNA , no pentaloop hairpins were added and we relied instead on the HiTRACE background subtraction routine overmod _and_background_correct_logL using unmodified controls . The lifft function from the LIFFT package [45] in HiTRACE was used to calculate the FMN binding dissociation constants for REEFFIT comparisons . Mutants with low signal were flagged as low quality and were not taken into account for the analysis . For a given mutate-and-map data set , REEFFIT infers the expected reactivity profile of each structure , the combination of structure population fractions ( also denoted here as structure weights ) , and sequence-position-wise noise levels that best fit the data ( see Figs 1b and 1c ) using prior information on known chemical reactivity distributions , a weak prior based on an approximate secondary structure energetic model , and a well-defined likelihood function . To achieve this fit , we expanded Gaussian factor analysis , a standard blind-source separation technique , to include position-specific , non-negative , non-Gaussian priors for the expected reactivities and factor weights that depend on both structure and sequence position . Local perturbations due to mutations , including the release of base pairing partners induced by the single nucleotide changes , were also included in the statistical model ( see section below “Handling local perturbations” ) . The model including these variables and parameters was fitted by REEFFIT using a maximum a posteriori ( MAP ) approximation . ( A Bayesian simulation inference method sampling over the posterior distribution gave indistinguishable results at significantly greater computational expense and is not presented in detail here . ) . In this section , we present the basic idea behind the model . Then , in the following sections we introduce the priors used for each variable and parameter in the model . Finally , we incorporate these priors and present inferential steps used to fit the model to the data . In a broad sense , for m chemical mapping measurements of n-nucleotide sequences , REEFFIT models the data , denoted here as Dobs ∈ ℜm×n , with a set of r secondary structures . The data are modeled as linear combinations of the structures’ reactivity profiles , denoted here as a matrix D ∈ ℜr×n , with a weight matrix W ∈ ℜm×r plus Gaussian noise ( see Fig 1b ) . Then , for each measurement j we have: Djiobs= ∑s ∈ structuresWjsDsi+ϵiϵi∼N ( 0 , Ψi ) ( 1 ) The rows of the weight matrix W correspond to the population fractions of the structures in each measurement . Therefore , the corresponding weights for each measurement define a probability distribution and are non-negative and add up to one: ∑s ∈ structuresWjs =1 , ∀j=1 , … , m∀j , s Wjs≥0 ( 2 ) In REEFFIT , D is a set of hidden variables since the isolated reactivity profiles of each structure are not typically available . We can impose a prior on each of these hidden profiles depending on their corresponding modeled secondary structure . Because a nucleotide’s chemical reactivity is reduced upon base pairing , a reasonable prior would force Dsi to be small if i is paired in structure s and higher if it is unpaired . The needed priors are derived empirically from distributions of the reactivities of paired and unpaired nucleotides in the RNA Mapping Database ( RMDB ) of RNAs with known crystallographic structure [39 , 46] . Let RMDBU and RMDBP be the empirical RMDB unpaired and paired reactivity distributions respectively and d be some reactivity , we can then define a prior likelihood for Dsi as: RMDBsi ( d ) = { RMDBU ( d ) if i is unpaired in sRMDBP ( d ) if i is paired in s ( 3 ) Here , we equate reactivities to values given by SHAPE modifiers; to handle other modifiers , e . g . dimethyl sulfate , RMDBU and RMDBP can be replaced with the respective estimated distributions using unpaired and paired data of that modifier . To simplify statistical inference , we approximated the RMDB reactivity distributions for paired and unpaired residues with exponential distributions , which has been found to be suitable for distributions of reactivities in unpaired regions [47] . We denote these approximations as: RMDB*P ( d ) =λPexp ( −λPd ) RMDB*U ( d ) =λUexp ( −λUd ) ( 4 ) Fitting these models to reactivity distributions for paired and unpaired residues in the RMDB gave scaling parameters of λP = 0 . 5 and λU = 0 . 2 . We denote RMDB*si as the resulting approximation for RMDBsi for structure s in position i , and λsi as the corresponding scaling parameter , that is: λsi= { λP if i is paired in sλU if i is unpaired in s ( 5 ) Systematic experimental perturbations used to alter the RNA's structural ensemble may induce local changes that cannot be captured by the linear combination of the weights W and the hidden reactivities D . This is the case in M2 experiments , where mutations induce local perturbations in the underlying reactivities of each structure . To model these perturbations , we add a set of random variables that take the values of the change in reactivity of D at perturbed positions in order to account for the data . That is , we add a ΔCsji variable for all sequence positions i that are in a set of perturbed sites for structure s in mutant j , perturbed ( s , j ) . These perturbed sites are defined as positions lying at most one nucleotide away from the site of a mutation in mutant j or from a base pair that would be disrupted due to a mutation in structure s in mutant j . To simplify notation , we define C as containing the values of these perturbation variables at the relevant positions and mutants: Csji={ ΔCsji if i ∈perturbed ( s , j ) 0 otherwise ( 6 ) We set the prior distributions of ΔCsji to a Gaussian approximation of differences in reactivities in M2 experiments available in the RMDB: ΔCsji~N ( 0 , 1/λperturbed ) ( 7 ) Fitting a normal distribution to the RMDB reactivity differences gave λperturbed = 2 . Since REEFFIT potentially fits hundreds or even thousands of structures ( see section below , “Building the structural ensemble and reducing model complexity” ) , we set a sparsity prior on the weight matrix W . Standard Laplacian ( l1 ) sparsity cannot be imposed because the per-measurement weights are probability distributions and are constrained to sum to unity for each variant . We therefore imposed a “smooth sparsity” regularizer λR using a Gaussian distribution ( l2 ) , where weights found to be unimportant for the model are reduced to low , but typically non-zero , values . To further avoid over-fitting weights , we encode a penalty that disallows dramatic deviations of computationally-predicted ΔΔG values for each structure s between the wild type sequence and each mutant to values calculated from the RNAstructure package . For mutant j , let ΔGjs be the energy for structure s calculated by the efn2 program in the RNAstructure package [48] , kB the Boltzmann constant , and T the temperature at which the experiments were performed . We denote the RNAstructure weights at ( j , s ) as: W0 , js=exp ( −ΔGjskBT ) ∑s′exp ( ΔGjs′kBT ) ( 9 ) We then impose a Gaussian prior on wild type weights that drives their wild-type/mutant weight differences to be close to the RNAstructure values through a parameter λΔ . Let wild type weights for each structure s are denoted as WWTs then: |WWTs−Wjs|~N ( |WWT0 , s−W0 , js| , 1/λΔ ) ( 10 ) Values for λΔ and λR were obtained using a cross-validation strategy ( see section below , “Building the structural ensemble and reducing model complexity” ) . These priors do not impose the probability distribution constraints of eq ( 2 ) on the measurement weights; instead , the probability distribution constraints are enforced during the optimization of the posterior function ( see below ) . After defining the variables for modeling the data and their respective priors , we can write the complete REEFFIT model . Let j = 1 , … , m , index over measurements . Then we write the model as: Djiobs= ∑s ∈ structuresWjs ( Dsi+Csji ) +ϵiϵi∼N ( 0 , Ψi ) Dsi∼RMDB*si;ΔCsji~N ( 0 , 1/λperturbed ) Wjs~N ( 0 , 1/λR ) ;|WWTs−Wjs|~N ( |WWT0 , s−W0 , js| , 1/λΔ ) ∑s ∈ structuresWjs =1∀j , s Wjs≥0 ( 11 ) Here , the hidden variables D and their perturbations ΔCsji are encoded in C , while the parameters to estimate are W and the variances Ψi . It is important to note that here , as in similar factor analysis models , all noise covariance matrices are assumed to be diagonal; that is , the measurements are independent of each other [49 , 50] . This assumption holds in the case of multiple chemical mapping measurements , since each measurement is carried out in different capillaries in capillary electrophoresis [51] or is based on Poisson distributed counts derived from separate single molecules in deep sequencing [41 , 52] . We used a hard expectation maximization ( EM ) algorithm to obtain MAP estimates for the values of the hidden variables and the model parameters . In hard EM optimization , the values for the hidden variables obtained in the E-step are maximum a posteriori estimates rather than expectations used in soft EM [50] . In standard factor analysis a soft EM is typically used: the E-step can be obtained in closed form by calculating the sufficient statistics for the likelihood function , which happen to be the first two moments of the posterior distribution of D: E[D|Dobs] and E[DDT|Dobs] [49 , 50 , 53] . However , the non-Gaussian form of our priors for each Dsi , precludes a closed form for these statistics and we instead used a hard EM strategy . This strategy yielded results that were comparable with a much more computationally expensive soft EM procedure that used Markov Chain Monte Carlo ( MCMC ) to approximate E[D|Dobs] and E[DDT|Dobs] in the E-step . For the M-step , we incorporated probability distribution constraints on W by casting the posterior maximization as a quadratic problem and solving it numerically . Given the REEFFIT factor analysis model ( 11 ) we want to calculate MAP estimates for W and each Ψi given the hidden variables D . The posterior function can be written as: p ( W , Ψ , D , C|Dobs ) ∝Likelihood×Prior of D×Prior of C×Prior of W=∏ i∈positionsj∈measurements1 ( 2πΨi ) 12exp ( −12Ψi ( Djiobs−∑s∈structuresWsj ( Dsi+Csji ) ) 2 ) ∏s∈structuresλsiexp ( −λsiDsi ) ∏i∈perturbed ( s , j ) 1 ( 2πλperturbed ) 12exp ( −12λperturbedCsji2 ) ∏s∈structures1 ( 2πλR ) 12exp ( −12λRWjs2 ) ∏s∈structures1 ( 2πλΔ ) 12exp ( −12λΔ ( | WWTs−Wjs |−| WWT0 , s−W0 , js | ) 2 ) ( 12 ) Corresponding to a log-posterior ( obviating the normalizing factor ) : logp ( W , Ψ , D , C|Dobs ) =−nm2log ( 2π ) +∑ i∈positionsj∈measurements−log ( Ψi ) 2−12Ψi ( Djiobs−∑s∈structuresWjs ( Dsi+Csji ) ) 2+∑s∈structureslog ( λsi ) −λsiDsi+∑i∈perturbed ( s , j ) −12log ( 2πλperturbed ) −12λperturbedCsji2+∑s∈structures−12log ( 2πλR ) −12λRWjs2+∑s∈structures−12log ( 2πλΔ ) −12λΔ ( | WWTs−Wjs |−| WWT0 , s−W0 , js | ) 2 ( 13 ) For the E-step , we estimate optimal hidden variable values for D and C . For each position i , finding the hidden variable for each structure s , Dsi that maximizes log p by differentiating and setting to zero , gives the following linear equations: ∑j∈measurementss′∈structuresWjsWjs′ ( Ds′i+Cs′ji ) =∑j∈measurementsWjsDjiobs+λsiΨi , ∀s∈structures ( 14 ) For measurement j , if position i lies in perturbed ( s , j ) , solving for Csji in the same manner gives the equation: ∑s′∈structuresWjsWjs′ ( Ds′i+Cs′ji ) −ΨiλperturbedCsji=WjsDjiobs∀s∈structures∀j∈measurements ( 15 ) For each position i , compiling the Eqs ( 14 ) and ( 15 ) for all structures and measurements results in a linear system . Solving these n systems ( one per sequence position ) , we obtain optimal values for D , C , which we name D* , C* . For the M-step , we calculate W by maximizing log p in each measurement enforcing the probability distribution constraints in Eq ( 2 ) . We can cast this maximization as a set of quadratic programs . Let Djobs be the data and Wj be the set of weights for measurement j , then , to estimate the optimal weights for j , we solve: argmaxWj logp ( Wj , Ψ , D* , C*|Djobs ) subject to ∑s∈structurestWjs=1 and Wjs≥0 ( 16 ) We solve the resulting quadratic programs using the CVXOPT python library [54] and denote the resulting weight matrix estimate as W* . Re-estimation of the variances Ψi in the M-step is also given by optimizing log p , but over Ψi: Ψi*=1m∑j∈measurements ( Djiobs−∑s∈structuresWjs* ( Dsj*+Csji* ) ) 2 ( 17 ) The E and M step optimization procedures are then repeated until the maximum difference of the induced base-pair probability matrices of the previous and current iteration is less than 1% . In our benchmarks , we have observed that usually 10 to 20 EM iterations are required for convergence . We note that our log posterior function is not convex and therefore our EM procedure does not necessarily converge to a global optimum and is sensitive to initial conditions . Nevertheless , testing different initial conditions with RNAstructure and ViennaRNA gave important improvements and similar results in our in silico benchmark ( see Supporting Results in S1 Text above ) . To initialize the variables Ψi , we choose the empirical variance of position i across all chemical mapping measurements , consistent with the variance calculation performed when using M2 z-scores as pseudo-energy bonuses for secondary structure prediction [28] . For initial estimates of W , we use RNAstructure to calculate the energies of each structure in each mutant as in Eq ( 9 ) . To calculate uncertainties for W that are robust to outliers and high-reactivity values , we re-fit the model in bootstrapped datasets by sampling columns ( i . e . sequence-positions ) with replacement of Dobs per bootstrap iteration . The final estimate of W is then the average weight matrix of all of the replicates; in all fits shown we report uncertainties as bootstrap standard deviations . Throughout this work , we used 100 bootstrap iterations . In the absence of bootstrapping , REEFFIT can also provide error estimates based on the Fisher information matrix approach , but those values are generally underestimates of the uncertainties . In most realistic scenarios for de novo RNA landscape modeling , it is not known a priori what set of structures would best model the data . To select an initial set of structures , we obtain a set of suboptimal structures for each sequence in the multi-dimensional chemical mapping experiment: for M2 experiments , the suboptimal structures of all variants involved have to be taken into account . We obtain at most 200 suboptimal structures of each mutant’s structural ensemble using the AllSub program , with default parameters ( 5% maximum energy difference from the MFE structure ) from the RNAstructure program suite ( version 5 . 5 ) [48] . The fits presented herein use all of the structures sampled in this manner . To reduce model complexity , we collapsed position-wise hidden reactivity variables that corresponded to different structures but formed part of the same structural motif . For example , if for position i we have 100 hidden reactivity variables that either correspond to a particular helix or an interior loop in the same part of the sequence , then we collapse the hidden variables into two variables , a helix and an interior loop variable , rather than 100 hidden reactivity variables . This variable collapse encodes the assumption that identical structural motifs will exhibit similar chemical reactivities independent of the structural context . Specifically , let sm be a structural motif ( the structural motifs that we take into account are: helices , bulges , x-way junctions , interior loops , dangles , single-stranded regions between helices , and hairpin loops ) and let motif ( i , sm ) be the structures that have the sm motif at position i . Then , we constrain the values for all variables in the set {Dsi|∀s ∈ motif ( i , sm ) }to be the same , collapsing the |motif ( i , sm ) | variables into only one variable . Because the number of structures sampled typically exceeds the number of measurements ( in most cases analyzed here we obtained ensembles of over 200 structures ) , this model simplification is essential to prevent over-fitting . In cases where the same sub-motif may have different reactivities in different structures ( e . g . upon ligand-binding ) , this information can be used to expand the number of fit parameters for that motif , as we carried out for the add or Tebowned riboswitches ( see below ) . To select the regularization parameters λΔ and λR we used a cross-validation approach . Since chemical mapping data are structured ( i . e . we cannot assume that reactivities in all positions come from the same population given the structural context of each nucleotide ) , we have to adapt the cross-validation technique accordingly . We employed a strategy frequently used in structured data contexts such as spline smoothing , where the i-th subsample in a k-fold cross-validation is the sequence {i , i+k , i+2k , …}[55] . This strategy samples the structured data uniformly across the nucleotide sequence , maintaining the assumptions necessary for cross-validation . In each fold of the cross validation , we obtain structure weights using the data from the training samples . With these weight estimates , we then obtain best predictions of the positions in the test sets with these weights . The cross-validation error is then the mean square error of the predicted and the observed test data [56] . We calculated an optimal value for λΔ and λR using our in silico , ab initio benchmark of 20 Rfam members , minimizing the 10-fold cross-validation error across all the benchmark . The optimal values , λΔ = 5 and λR = 0 . 26 were then used for all fits presented herein .
The REEFFIT programs and their source code are available at http://rmdb . stanford . edu/tools/docs/reeffit/ , along with software documentation and tutorials . M2 capillary electrophoresis data for the Bistable , add riboswitch , M-stable , and Tebowned RNAs have been deposited in the RMDB ( RMDB IDs BSTHPN_1M7_0000 . rdat , ADDSCHW_1M7_0000 , ADDSCHW_1M7_0001 , MSTBL_1M7_0000 , TBWND_1M7_0000 , and TBWND_1M7_0001 ) . M2-seq data for the MedLoop are part of the EteRNA cloud lab , rounds 72 ( RMDB ID ETERNA_R72_0000 , project name “MedLoop” ) . RDAT files for the simulated datasets can be downloaded from http://purl . stanford . edu/zr287dq2666
|
RNA is a versatile macromolecule that underlies core natural processes throughout living systems and new strategies to re-engineer these systems . In several cases , this versatility is due to the ability of an RNA molecule to adopt multiple conformations: the full description of the molecule involves a ‘landscape’ of alternative structures that are present at equilibrium and whose interconversions are critical for function . However , there is little work on empirically probing these landscapes due to size , resolution , and infrastructure constraints of available experimental methods . We show herein that RNA landscapes can be characterized rapidly through the mutate-and-map ( M2 ) methodology when coupled with a novel blind-source separation method , REEFFIT . We present extensive computational and experimental benchmarks supporting the use of M2-REEFFIT to detect alternative RNA secondary structures accurately without ‘over-predicting’ them . We then experimentally address a basic question in RNA biophysics—do landscapes involving multiple secondary structures require explicit selection or can they arise in artificial RNA sequences without selection ? For both an artificial flavin mononucleotide riboswitch and a randomly generated sequence , M2-REEFFIT and further mutation tests find evidence for alternative , unexpected states , suggesting that rich landscapes are common and will become better appreciated with use of the developed technology .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods",
"Data",
"set",
"and",
"software",
"availability"
] |
[] |
2015
|
Rich RNA Structure Landscapes Revealed by Mutate-and-Map Analysis
|
Feeding behavior is one of the most essential activities in animals , which is tightly regulated by neuroendocrine factors . Drosophila melanogaster short neuropeptide F ( sNPF ) and the mammalian functional homolog neuropeptide Y ( NPY ) regulate food intake . Understanding the molecular mechanism of sNPF and NPY signaling is critical to elucidate feeding regulation . Here , we found that minibrain ( mnb ) and the mammalian ortholog Dyrk1a target genes of sNPF and NPY signaling and regulate food intake in Drosophila melanogaster and mice . In Drosophila melanogaster neuronal cells and mouse hypothalamic cells , sNPF and NPY modulated the mnb and Dyrk1a expression through the PKA-CREB pathway . Increased Dyrk1a activated Sirt1 to regulate the deacetylation of FOXO , which potentiated FOXO-induced sNPF/NPY expression and in turn promoted food intake . Conversely , AKT-mediated insulin signaling suppressed FOXO-mediated sNPF/NPY expression , which resulted in decreasing food intake . Furthermore , human Dyrk1a transgenic mice exhibited decreased FOXO acetylation and increased NPY expression in the hypothalamus , as well as increased food intake . Our findings demonstrate that Mnb/Dyrk1a regulates food intake through the evolutionary conserved Sir2-FOXO-sNPF/NPY pathway in Drosophila melanogaster and mammals .
Neuropeptides regulate a wide range of physiological processes in animals . In mammals , NPY is widely distributed in the brain and involved in various physiological functions including food intake . In the mammalian brain , the hypothalamus is the center for controlling food intake . The hypothalamic injection of NPY in the rat brain induces hyperphagia and obesity . In the hypothalamus , the arcuate nucleus ( ARC ) that contains orexigenic NPY and AgRP expressing neurons and anorexigenic POMC neurons senses hormonal levels of insulin and leptin and regulates food intake [1] . In Drosophila , sNPF , a functional homolog of NPY produced in sNPFnergic neurons of the fly brain , regulates food intake and growth [2] . Recently , we reported that sNPF and sNPF receptor ( sNPFR1 ) regulate body growth through evolutionary conserved ERK-mediated insulin signaling in Drosophila and rat insulinoma cells [3] . Drosophila Minibrain ( Mnb ) and its mammalian ortholog Dual specificity tyrosine-phosphorylation-regulated kinase 1a ( Dyrk1a ) are highly expressed in the neural tissues [4] , [5] , [6] . The Dyrk1a gene has been implicated in Down Syndrome ( DS ) [5] , [7] and the expression level of Dyrk1a is increased in DS patients and Ts65Dn mice , a mouse model of Down syndrome [4] , [8] . Mutations of mnb and Dyrk1a in Drosophila and mammals show neural phenotypes like defects in neuroblasts proliferation and brain development [6] , [9] . Human patients with truncated mutations in the Dyrk1a gene also show microcephaly [10] , [11] . To date , however , the effects of mnb and Dyrk1a upon food intake have not been described . FoxO1 modulates food intake by regulation of orexigenic Argp and anorexigenic Pomc genes in the hypothalamus of mice . In the ARC of hypothalamic neurons , FoxO1 is localized in the nuclei during fasting and in the cytoplasm by feeding [12] . Sirtuin1 ( Sirt1 ) , the mammalian ortholog of Drosophila Silent information regulator 2 ( Sir2 ) , in the ARC also regulates food intake [13] . The Sirt1 protein level increases during fasting . Sirt1 inhibition by the hypothalamic knock-out in the AgRP neurons decreases food intake [14] . In N43 hypothalamic cells , pharmacological inhibition of Sirt1 increases anorexigenic POMC expression but co-treatment with Sirt1 inhibitor and FoxO1 siRNA does not [15] , suggesting that Sirt1-mediated FoxO1 deactylation is involved in the regulation of POMC mRNA and food intake . In this study , we identified mnb and Dyrk1a as target genes of sNPF and NPY signaling , respectively , and describe a molecular mechanism of how Mnb and Dyrk1a regulate food intake in Drosophila and mice .
To find genes affected by sNPF signaling , we performed a DNA microarray analysis using the Affymetrix Drosophila Genome 2 . 0 Array GeneChip with mRNA extracted from Drosophila neuronal BG2-c6 cells treated with sNPF peptide . Among the 159 genes with at least a two-fold change , mRNA of mnb increased 34-fold compared to the control ( Table S1 ) . To test whether the expression of mnb is dependent on sNPF signaling in vivo , we examined the expression levels of mnb in sNPF and sNPFR1 mutants . When sNPF was overexpressed in sNPFnergic neurons with the sNPF-Gal4 driver [16] ( sNPF>sNPF , sNPF>2XsNPF ) , mnb mRNA increased 4 to 5-fold compared with the sNPF-Gal4 . mRNA of mnb decreased by less than half when sNPF was inhibited ( sNPF>sNPF-Ri ) or by an sNPF mutant ( sNPFc00448 ) ( Figure 1A and Figure S1A ) . When sNPFR1 was overexpressed via a sNPFR1-Gal4 driver ( Figure S2 ) ( sNPFR1>sNPFR1 ) , mnb mRNA was increased 3-fold compared with the sNPFR1-Gal4 control . When sNPFR1 was inhibited ( sNPFR1>sNPFR1-Ri ) or suppressed ( sNPFR1>sNPFR1-DN ) , mnb mRNA was decreased by more than 50% ( Figure 1A and Figure S1A ) . Like mnb mRNA , Mnb proteins were also increased in sNPF or sNPFR1 overexpression with the sNPF-Gal4 or sNPFR1-Gal4 driver , ( sNPF>2XsNPF , sNPFR1>sNPFR1 ) while reduced in an sNPF mutant ( sNPFc00448 ) or sNPFR1 inhibition ( sNPFR1>sNPFR1-Ri ) compared with the sNPF-Gal4 or sNPFR1-Gal4 control ( Figure S3A ) . However , the numbers of Mnb expression neurons ( asterisks ) are consistent in the sNPFR1-Gal4 control , sNPFR1 overexpression ( sNPFR1>sNPFR1 ) , sNPFR1 inhibition ( sNPFR1>sNPFR1-Ri ) , and an sNPFc00448 mutant ( Figure S3B–S3F ) . These results indicate that sNPF-sNPFR1 signaling regulates mnb mRNA and protein expression in Drosophila . To understand how Mnb protein may interact with the sNPFR1 receptor , we immunostained fly adult brains with Mnb and sNPFR1 antibodies . The Mnb antibody produced strong and weak staining in neuronal cells ( Figure 1H , 1K , red ) while the sNPFR1 receptor antibody stained many neurons ( Figure 1I , 1L , green ) . Among the strongly stained Mnb neurons , cell bodies of symmetrically localized median neurons behind the antennal lobe show overlap with the antibody against sNPFR1 ( Figure 1J , 1M , arrows ) . At least ten neuronal cell bodies in median neurons were stained with the both antibodies . This coincidence suggests that at least part of Mnb function may be regulated by sNPF-sNPFR1 signaling . Since sNPF signaling regulates food intake and growth , and growth is regulated by ERK-mediated insulin signaling [3] , we hypothesized that sNPF may regulate food intake through the mnb gene . To assess this hypothesis , we used the CAFÉ assay [17] to measure feeding in mnb mutant adults . Because homozygous mnb deletion mutants ( mnbd305 and mnbd419 ) generated by the imprecise excisions of the P-element ( Figure S4A ) are lethal ( as are homozygous Dyrk1a mutant mice ) we analyzed mnb overexpression and hypomorphs generated by RNAi . mnb overexpression in sNPFR1 neurons ( sNPFR1>mnb ) increased cumulative food consumption compared to the sNPFR1-Gal4 control whereas inhibiting mnb ( sNPFR1>mnb-Ri ) decreased cumulative food consumption ( Figure 1C ) , indicating that mnb expression in sNPFR1 neurons can regulate food intake . Likewise , we measured the amount of food intake by the amount of digested dye from colored food . Overexpression of mnb in sNPFR1 neurons ( sNPFR1>mnb and sNPFR1>2Xmnb ) increased consumed dye up to 57% compared with that of the sNPFR1-Gal4 control whereas mnb inhibition ( sNPFR1>mnb-Ri ) or the mnb mutant ( mnbG1767 ) decreased this intake by 30% ( Figure 1B and Figure S1B ) . As expected , levels of mnb mRNA and protein were markedly reduced by mnb inhibition and by the mnbG1767 mutant relative to the sNPFR1-Gal4 and w- controls ( Figure S4B , S4C ) . Since sNPFR1 signaling in the insulin producing cells ( IPCs ) regulates body growth through insulin signaling [3] , we examined the effect of mnb in IPCs upon food intake . However , food intake was not affected by mnb overexpression in IPCs driven via Dilp2-Gal4 ( Dilp2>mnb and Dilp2>2Xmnb ) or by mnb inhibition in IPCs ( Dilp2>mnb-Ri ) ( Figure 1B ) . Expression of mnb in sNPFR1 neurons but not in IPCs ( Figure 1D–1G ) is sufficient to regulate food intake . To determine the consequences of mnb control upon food intake we measured the body weight of young adults from mutant and control . Overexpression of mnb in sNPFR1 neurons ( sNPFR1>mnb ) increased body weight relative to that of sNPFR1-Gal4 controls , similar to the effect seen when sNPFR1 is overexpressed ( sNPFR1>sNPFR1 ) . On the contrary , body weight is decreased when mnb is repressed in sNPFR1 neurons ( sNPFR1>mnb-Ri ) and mnbG1767 mutant ( Figure S4D ) . The amounts of food intake in the mutants were similar when they were normalized to body mass or to the number of flies ( Figure S4E ) . Since mnb is involved in neural development [6] , [9] , we restricted mnb expression in the adult stage using the tub-GAL80ts inducible system [18] and tested food intake . mnb overexpression ( sNPFR1-Gal4+tubGal80ts>mnb , sNPFR1-Gal4+tubGal80ts>2Xmnb ) and mnb inhibition ( sNPFR1-Gal4+tubGal80ts>mnb-Ri ) flies were cultured in the 22°C permissive temperature until adulthood to suppress sNPFR1-Gal4 expression by the tubGal80ts . Then , these adult flies were shifted to the 30°C restrictive temperature in which the tubGal80ts cannot suppress sNPFR1-Gal4 . In the permissive condition , the mnb overexpression and mnb inhibition flies did not change the amount of food intake compared with the control flies ( sNPFR1-Gal4; tub-Gal80ts ) ( Figure S5A ) . However , in the restrictive condition , the mnb overexpression increased food intake compared with the control and the mnb inhibition suppressed food intake ( Figure S5B ) . These results indicate that the food intake phenotype of mnb mutants is not due to developmental effects . To study how sNPFR1 regulates mnb expression , we treated Drosophila central nervous system-derived BG2-c6 cells [19] with synthetic sNPF peptide , which changed sNPF and sNPFR1 expression slightly ( Figure S6A ) . Consistent with our initial observations and with patterns in genetically manipulated flies , sNPF treatment increased mnb mRNA more than 5-fold compared to the control when measured by quantitative PCR ( Figure 2A ) . Then , we tested whether the induction of this mnb mRNA is mediated by ERK , as we have previously observed for the induction of Drosophila insulin like peptides ( Dilps ) by sNPF [3] . However , ERK inhibitor PD98059 treatment of the sNPF peptide-treated cells did not suppress the mnb expression . On the other hand , sNPFR1 is a G-protein coupled receptor ( GPCR ) , and the second messenger of GPCRs is cAMP or Ca++ which respectively activates PKA or PKC [20] . Thus , we treated BG2-c6 cells with the protein kinase A ( PKA ) inhibitor H89 or with protein kinase C ( PKC ) inhibitor Chelerythrine Chloride ( CC ) . H89 decreased both basal and sNPF-induced mnb expression level but the PKC inhibitor CC showed no effect ( Figure 2A ) . sNPF signaling appears to control mnb expression through PKA , not through ERK or PKC . Consistent with this interpretation , BG2-c6 cells treated with sNPF showed increased levels of cAMP in a time-dependent manner , peaking at 15 min ( Figure S6B ) . To find the Gα subunit of the sNPFR1 G-protein heterotrimer , we examined Gαs and Gαi , both of which modulate cAMP [20] . When transfected into BG2-c6 cells Gαs siRNA inhibited sNPF-induced cAMP whereas transfection with Gαi siRNA did not ( Figure 2C ) , suggesting that Gαs is a Gα subunit of sNPFR1 that can modulate the cAMP-PKA pathway in Drosophila neuronal cells . Next , we examined the activation of the cAMP responding element binding protein ( CREB ) , which is a PKA down-stream transcription factor [21] . sNPF stimulated the phosphorylation of CREB in control cells whereas Gαs siRNA transfection suppressed this sNPF dependent activation of CREB ( Figure 2E ) . In addition , Gαs siRNA transfection completely blocked the induction of mnb by sNPF , but Gαi siRNA transfection did not ( Figure 2G ) . These data indicate that Gαs is a key Gα subunit of the sNPFR1 G-protein as it regulates mnb expression . Taken together , these findings demonstrate that sNPF signaling effectively regulates mnb expression through the Gαs-cAMP-PKA-CREB pathway in Drosophila neuronal cells . To compare the functional conservation of sNPF-sNPFR1-PKA-CREB-mnb signaling with the signaling of mammalian NPY , we conducted similar experiments with mouse GT1-7 hypothalamic cells [22] . NPY treatment increased Dyrk1a mRNA while the PKA inhibitor H89 strongly suppressed NPY-induced Dyrk1a expression ( Figure 2B ) . NPY signaling activates Dyrk1a expression through PKA , much like the PKA mediated mnb expression by sNPF in fly neuronal cells . Next , we measured the cAMP level in the NPY treated GT1-7 cells . As expected , cAMP level increased time-dependently and peaked at 15 min ( Figure S6C ) . Five NPY receptors ( NPYR1 , 2 , 4 , 5 , and 6 ) mediate the NPY signal [23] . Among them , NPYR1 , 2 , and 5 receptors are broadly expressed in the mouse nervous system and mediate NPY-induced food intake [24] . We treated GT1-7 cells with chemical inhibitors against these receptors: BIBO3304 for NPYR1 , BIIE0246 for NPYR2 , and CGP71683 for NPYR5 . The NPYR1 inhibitor BIBO3304 substantially decreased the NPY-induced cAMP level; little effect was seen for the inhibitors of NPYR2 and NPYR5 ( Figure 2D ) . Thus , NPY appears to activate the cAMP-PKA pathway mainly through NPYR1 in GT1-7 hypothalamic cells . Next , we measured the CREB activation . As expected , inhibiting PKA or NPYR1 suppressed the NPY-induced activation of CREB ( Figure 2F ) , confirming that NPY signal is mediated through NPYR1-cAMP-PKA-CREB . In addition , the NPYR1 inhibitor strongly suppressed NPY-induced Dyrk1a expression; this was not seen with the inhibitors of NPYR2 and NPYR5 ( Figure 2H ) . Taken together , these findings indicate that NPY signaling regulates Dyrk1a expression mainly through the NPYR1-cAMP-PKA-CREB pathway in mouse hypothalamic cells . Importantly , this signal transduction pathway is conserved between fly neuronal cells and mammalian hypothalamic cells . To study genetic interactions among sNPFR1 , Gαs , PKA , CREB , and mnb genes , we suppressed Gαs , PKA , CREB , and mnb by RNAi and Dominant Negative ( DN ) forms in neurons that simultaneously overexpressed sNPFR1 . Each of these suppression genotypes reduced the level of mnb mRNA compared with sNPFR1-Gal4 and UAS controls ( Figure 3A and Figure S7A ) . In contrast to the strong induction of mnb produced by sNPFR1 overexpression alone ( sNPFR1>sNPFR1 ) , mnb induction was inhibited in genotypes where sNPFR1 overexpression occurred with each of the suppression constructs ( sNPFR1>sNPFR1+Gαs-Ri , sNPFR1>sNPFR1+PKA-DN , sNPFR1>sNPFR1+CREB-DN , sNPFR1>sNPFR1+mnb-Ri ) ( Figure 3B ) . In sNPFR1 neurons of flies , as in isolated cells , Gαs , PKA , and CREB may work downstream of sNPFR1 to regulate mnb expression . The consequences of these interactions are also seen in terms of food intake . Gαs , PKA , CREB , and mnb suppression mutant flies have reduced food intake compared to those of the sNPFR1-Gal4 and UAS controls ( Figure 3C and Figures S1C , S7B ) . Furthermore , increased food intake of sNPFR1 overexpression was suppressed by co-inhibition of Gαs , PKA , and CREB , respectively ( Figure 3D ) . These results suggest that the sNPFR1 may regulate food intake through Gαs , PKA , CREB , and mnb . Based on promoter analysis of mnb genes from twelve Drosophila species , we found a conserved cAMP responding element ( CRE ) site ( Figure S8 ) . Interestingly , the promoters of human Dyrk1a and mouse Dyrk1a genes contain CRE [25] . To test whether CREB binds to the promoter of the mnb gene , we performed the chromatin immunoprecipitation ( ChIP ) -PCR analysis with the CREB antibody in sNPF treated Drosophila neuronal BG2-c6 cells . CREB binding was enriched at the sNPF treated promoter region of the mnb gene by 3-fold compared to the Act5C and sNPF non-treated controls ( Figure 3E ) . Together these in vivo and in vitro findings indicate that sNPF-sNPFR1-Gαs-PKA-CREB pathway controls expression of the mnb target gene and regulates food intake in Drosophila . A possible avenue through which Mnb regulates food intake could involve Sirt1/Sir2 . Notably , Dyrk1a kinase phosphorylates Sirt1 in HEK293T cells [26] , and activated Sirt1 deacetylates FoxO1 to modulate the activity of this transcription factor in the rat hypothalamus [15] . Accordingly we determined if these interactions were present and associated in mouse hypothalamic GT1-7 cells . In cells transfected with Dyrk1a or treated with NPY , phosphorylation of Sirt1 was increased as detected by immunoprecipitation with Sirt1 antibody , followed by immunobloting with phospho-threonine ( pThr ) antibody . Sirt1 phosphorylation was reduced by Dyrk1a siRNA or Dyrk1a siRNA with NPY ( Figure 4A ) . In addition , FoxO1 acetylation was reduced in cells transfected by Dyrk1a or treated with NPY , while FoxO1 acetylation was increased by Dyrk1a siRNA , Dyrk1a siRNA with NPY , or Dyrk1a transfection coupled with the Sirt1 inhibitor EX527 ( Figure 4C ) . Importantly , NPY mRNA itself was increased in cells transfected with Dyrk1a or treated with NPY peptide , and NPY mRNA was decreased by Dyrk1a siRNA , Dyrk1a siRNA with NPY , or Dyrk1a overexpression in the presence of Sirt1 inhibitor ( Figure 4B ) . In mouse hypothalamic GT1-7 cells , Dyrk1a phosphorylates Sirt1 and this activated Sirt1 appears to deacetylate FoxO1 which in turn positively regulates expression of NPY . To study genetic interactions among mnb , Sir2 , and dFOXO in an animal model , we manipulated Sir2 and dFOXO in the Drosophila mnb overexpression genotype . When mnb , Sir2 , and dFOXO were overexpressed in sNPFR1-Gal4 neurons ( sNPFR1>mnb , sNPFR1>Sir2 , sNPFR1>dFOXO ) ( Figure S9A ) , sNPF mRNA and food intake were increased compared to sNPFR1-Gal4 and UAS controls ( Figure 4D and 4E , Figure S7B and S7C ) . Conversely , when mnb , Sir2 , and dFOXO were inhibited in sNPFR1 expressing neurons ( sNPFR1>mnb-Ri , sNPFR1>Sir2-Ri , sNPFR1>dFOXO-Ri ) ( Figure S9B ) , the expression levels of sNPF and food intake were decreased or similar to those of sNPFR1-Gal4 and UAS controls ( Figure 4D and 4E , Figure S7B and S7C ) . Finally the level of sNPF mRNA and food intake were reduced in adults when Sir2 or dFOXO were inhibited in sNPFR1 neurons that overexpressed mnb ( sNPFR1>mnb+Sir2-Ri , sNPFR1>mnb+dFOXO-Ri ) compared with flies only overexpressing mnb ( sNPFR1>mnb ) . These data suggest that mnb may regulate sNPF expression and food intake through Sir2 and dFOXO . Since fasting can stimulate food intake , we tested whether an acute period of food deprivation affected the expression of mnb and sNPF of adult flies . Levels of mnb and sNPF mRNA increased 2-fold after 12 h starvation ( Figure 4F ) . We propose that dFOXO contributes to this expression of sNPF in starved flies . We identified a common dFOXO consensus binding site ( RWWAACA ) in the sNPF promoter from twelve Drosophila species ( Figure S10 ) and performed a chromatin immunoprecipitation ( ChIP ) -tiled gene array analysis with dFOXO antibody in fed and starved adult flies . dFOXO binding was enriched at the promoter region of sNPF gene more than 3-fold in the starved flies compared to the Act5c and fed controls ( Figure 4G ) . These results suggest that the dFOXO transcriptional factor regulates sNPF mRNA expression by direct binding to its promoter in Drosophila , as seen for FoxO1 regulation of NPY expression in mice [27] . Overall , these results from mouse hypothalamic GT1-7 cells and Drosophila indicate that the Mnb/Dyrk1a-Sir2-FOXO pathway positively regulates sNPF/NPY expression and food intake . The positive feedback regulation of sNPF signaling we have described to this point must occur alongside a system to negatively regulate sNPF signaling . Insulin , one of several anorexigenic hormones , inhibits food intake through AKT-mediated FoxO1 inactivation in the hypothalamus [27] . In Drosophila , neuronal overexpression of Dilps negatively regulates larval food intake [28] . To understand the inhibitory mechanism of insulin on food intake , we analyzed the phosphorylation of FOXO and the expression of NPY and sNPF . In the mouse hypothalamic GT1-7 cells , insulin treatment increased FoxO1 phosphorylation and decreased NPY mRNA while insulin combined with AKT inhibitor co-treatment slightly decreased FoxO1 phosphorylation and increased NPY expression ( Figure 5A , 5B ) . Likewise , in fly neuronal BG2-c6 cells , insulin with AKT inhibitor co-treatment increased sNPF mRNA ( Figure 5C ) . Thus , in both models AKT-mediated insulin signaling increased FOXO phosphorylation and suppressed NPY or sNPF expression . We extended these results with analyses of Drosophila with insulin and insulin receptor transgenes . Compared to Dilp2-Gal4 and sNPFR1-Gal4 controls , sNPF mRNA and food intake were decreased when Dilp2 was overexpressed in insulin producing cells ( Dilp2>Dilp2 ) and when insulin receptor ( InR ) was overexpressed in sNPFR1 expressing neurons ( sNPFR1>InRWT ) ( Figure 5D , 5E ) . On the other hand , sNPF expression and food intake were increased when InR was suppressed by a dominant negative construct expressed in sNPFR1 neurons ( sNPFR1>InRDN ) ( Figure 5D , 5E ) . Fasting may contribute to sNPF expression and the propensity for food intake because fasting in the adult reduces the expression of several Dilps ( Figure 5F ) , as previously observed to occur in Drosophila larvae [29] . Taken together , the results from mouse and Drosophila neuronal cells and from adult flies indicate that the insulin signaling negatively regulates sNPF/NPY expression and food intake . To evaluate these Mnb/Dyrk1a-Sir2-FOXO-NPY interactions and consequences in a mammalian animal model , we examined FoxO1 acetylation and NPY expression in the hypothalamus of transgenic mice containing the human Dyrk1a BAC clone ( hDyrk1a TG ) . As expected , in the Western blot , Dyrk1a in the hypothalamus was increased in hDyrk1a TG mice compared to controls ( Figure 6A ) . On the other hand , FoxO1 in the hypothalamus was less acetylated in hDyrk1a TG mice compared to controls ( Figure 6C ) . Hypothalamic NPY mRNA as well as serum NPY levels were elevated in in hDyrk1a TG mice compared to controls ( Figure 6B ) . Thus , mammalian Dyrk1a appears to regulate FoxO1 acetylation and NPY expression in the mouse hypothalamus , as we have observed for this system in Drosophila sNPFR1 neurons . To assess whether mammalian Dyrk1a also regulates food intake as seen for the homolog mnb of Drosophila , we monitored food intake in seven-week-old hDyrk1a TG mice . Daily food consumption was increased in the transgenic mice compared to littermate controls ( Figure 6D ) and the average food intake of hDyrk1a transgenic mice was elevated by 15% ( Figure 6E ) . Correspondingly , the hDyrk1a transgenic mice presented slightly increased mass ( Figure S11 ) . Dyrk1a thus appears to regulate food intake through the expression of NPY mediated by FOXO in a molecular pathway that is evolutionarily conserved in Drosophila .
The production of sNPF and NPY in sNPFnergic and hypothalamic neurons of flies and mammals respectively , is increased during fasting . These neuropeptides are secreted to produce paracrine and endocrine effects [24] but also feedback upon their synthesizing neurons where they respectively induce mnb and Dyrk1a gene expression through the PKA-CREB pathway ( Figure 6F ) . This Mnb/Dyrk1a kinase phosphorylates and activates the Sir2/Sirt1 deacetylase , which in turn deacetylates and activates the FOXO transcription factor . Among its many potential targets , FOXO then increases sNPF/NPY mRNA expression . Negative controls modulate the positive feedback of sNPF/NPY . Feeding activates the insulin receptor-PI3K-AKT pathway . FOXO becomes phosphorylated and transcriptionally inactivated by translocation to the cytoplasm [30] . In this state the induction of sNPF/NPY by FOXO is decreased . Because sNPF and NPY are orexogenic , their positive feedback during fasting should reinforce the propensity for food intake whereas the negative regulation of sNPF and NPY mRNA during feeding condition would then contribute to satiety ( Figure 6F ) . FOXO family transcriptional factors are involved in metabolism , longevity , and cell proliferation [31] . FOXO is in part regulated in these processes by post-transcriptional modifications including phosphorylation and acetylation [30] . In many model systems , the ligand activated Insulin-PI3K-AKT pathway phosphorylates FOXO to inactivate this transcription factor by moving it to the cytoplasm . The cytoplasmic localization of FOXO is mediated by 14-3-3 chaperone proteins in Drosophila and mammals [32] , [33] . FOXO may also be acetylated , as is FoxO1 of mice , by the CREB-binding protein ( CBP ) /p300 acetylase and this inhibits FOXO transcriptional function by suppressing its DNA-binding affinity . Such FoxO1 acetylation can be reversed by SirT1 to help activate the FoxO1 transcription factor [34] . Here we describe for Drosophila how dFOXO in sNPFR1 neurons regulates the expression of sNPF and food intake ( Figure 4D , 4E ) . This mechanism parallels how hypothalamic FoxO1 regulates food intake through its control of orexigenic NPY and Agrp in rodents [12] , [27] . Post-transcriptional modification of FOXO is central to these controls in both animals . sNPF and NPY expression is increased when FOXO is deacetylated by Sir2/Sirt1 , while sNPF and NPY are decreased when FOXO is phosphorylated via the Insulin-PI3K-AKT pathway . Post-transcriptional modifications of FOXO proteins play a critical role for controlling food intake through the sNPF and NPY expression in flies and rodents . Mnb/Dyrk1a has been described to participate in olfactory learning , circadian rhythm , and the development of the nervous system and brain [6] . Mnb and Dyrk1a proteins contain a nuclear targeting signal sequence , a protein kinase domain , a PEST domain , and a serine/threonine rich domain . The kinase domains are evolutionary well-conserved from flies to humans [35] . In Down syndrome ( DS ) , chromosome 21 trisomy gives patients three copies of a critical region that includes the Mnb/Dyrk1a; trisomy of this region is associated with anomalies of both the nervous and endocrine systems [36] . DS patients often show high Body Mass Index due to the increased fat mass . Children with DS have elevated serum leptin coupled with leptin resistance , both of which contribute to the obesity risk common to DS patients [37] , [38] . We now observe a novel function of Mnb/Dyrk1a that may underlay this metabolic condition of DS patients . Mnb/Dyrk1a regulates food intake in flies and mice . This is controlled by sNPF/NPY-PKA-CREB up-stream signaling and thus produces down-stream affects upon Sir2/Sirt1-FOXO-sNPF/NPY . Fasting not only increases the expression of mnb , but also of sNPF , suggesting that Mnb kinase activates a positive feedback loop where Sir2-dFOXO induces sNPF gene expression . Notably , fasting increases Sirt1 deacetylase activity and localizes FoxO1 to the nucleus in the orexogenic AgRP neurons of the mouse hypothalamus [15] . Increased dosage of Dyrk1a in DS patients may reinforce the positive feedback by NPY and disrupt the balance between hunger and satiety required to maintain a healthy body mass . Insulin produced in the pancreas affects the hypothalamus to regulate feeding in mammals [1] . Insulin injected into the intracerebroventrical of the hypothalamus reduces food intake while inhibiting insulin receptors of the hypothalamic ARC nucleus causes hyperphasia and obesity in rodent models [39] , [40] . Here we saw a similar pattern for Drosophila where overexpression of insulin-like peptide ( Dilp2 ) at insulin producing neurons decreased food intake while food intake was increased by inhibiting the insulin receptor in sNPFR1 expressing neurons ( Figure 5E ) . Likewise , during fasting , serum insulin and leptin levels are decreased in mammals [1] , as is mRNA for insulin-like peptides of Drosophila [29] , [41] ( Figure 5F ) . Thus , the mechanism by which insulin and insulin receptor signaling suppresses food intake is conserved from fly to mammals in at least some important ways . Previously , we reported how sNPF signaling regulates Dilp expression through ERK in IPCs and controls growth in Drosophila [3] . Here , we show that sNPF signaling regulates mnb expression through the PKA-CREB pathway in non-IPC neurons and controls food intake ( Figure 1B , 1D–1G ) . Since sNPF works through the sNPFR1 receptor , sNPFR1 in IPCs and non-IPCs neurons might transduce different signals and thereby modulate different phenotypes . Four Dilps ( Dilp1 , 2 , 3 , and 5 ) are expressed in the IPCs of the brain [42] . Interestingly , levels of Dilp1 and 2 mRNA are reduced in the sNPF mutant , which has small body size [3] , but here we find only Dilp3 and 5 mRNA levels are reduced upon 24 h fasting . Likewise , only Dilp5 is reduced when adult flies are maintained on yeast-limited diets [43] . In addition , Dilp1 and 2 null mutants show slight reduced body weights but Dilp3 and Dilp5 null mutants do not [44] . These results suggest that Dilp1 and 2 behave like a mammalian insulin growth factor for size regulation while Dilp3 and 5 act like a mammalian insulin for the regulation of metabolism . However , in the long term starvation , Dilp2 and Dilp5 mRNA levels are reduced and Dilp3 mRNA expression is increased [45] . During fasting , sNPF but not sNPFR1 mRNA expression was increased in samples prepared from fly heads ( Figure 4F and Figure S9C ) , which increases food intake . On the other hand , in feeding , the high level of insulin signaling reduced sNPF but not sNPFR1 mRNA expression and suppressed food intake ( Figure 5D and 5E , Figure S9D ) . Interestingly , in the antenna of starved flies , sNPFR1 but not sNPF mRNA expression is increased and induces presynaptic facilitation , which resulted in effective odor-driven food search . However , high insulin signaling suppresses sNPFR1 mRNA expression and prevents presynaptic facilitation in DM1 glomerulus [46] . These results indicate that starvation-mediated or insulin signaling-mediated sNPF-sNPFR1 signaling plays a critical role in Drosophila feeding behavior including food intake and food search even though the fine tuning is different . In this study , we present a molecular mechanism for how sNPF and NPY regulate food intake in Drosophila and mice . We describe a system of positive feedback regulation for sNPF and NPY signaling that increases food intake and a mode of negative regulation for sNPF and NPY by the insulin signaling that suppresses food intake . Modifications of the FOXO protein play a critical role for regulating sNPF and NPY expression , resulting in the control of food intake .
Drosophila melanogaster were cultured and at 25°C on standard cornmeal , yeast , sugar , agar diet . Wild-type Canton-S , w- , and UAS-CREB-DN were obtained from the Bloomington stock center . sNPFc00448 was obtained from the Harvard stock center ( Exelixis stock collection ) . UAS-sNPF , UAS-2XsNPF , UAS-sNPF-Ri , UAS-sNPFR1 , UAS-sNPFR1-DN and sNPF-Gal4 transgenic flies were described in our previous reports [2] , [3] , [16] . The sNPFR1-Gal4 construct was generated from a 2 . 5 kb genomic DNA fragment of the 5′-untranslated region of the sNPFR1 gene . The full-length coding sequence of Drosophila minibrain-H ( mnb , CG 42273 ) was subcloned into the pUAS vector to generate the pUAS-mnb construct . sNPFR1-Gal4 and UAS-mnb transgenic flies were obtained by the P-element-mediated germ line transformation [47] . mnbG1767 , an EP line for minibrain , was purchased from the GenExel , Inc . ( KAIST , Korea ) . UAS-sNPFR1-Ri ( VDRC9379 ) , UAS-mnb-Ri ( VDRC28628 ) , UAS-Sir2-Ri ( VDRC23201 ) and UAS-FOXO-Ri ( VDRC106097 ) were obtained from the Vienna Drosophila RNAi Center ( VDRC ) . Dilp2-Gal4 , UAS-Gαs-Ri , UAS-PKA-DN ( a dominant-negative form of PKA ) , UAS-Sir2 transgenic flies were described previously [42] , [48] , [49] , [50] , [51] . To express these UAS lines , UAS-Gal4 system was used [52] . For minimizing the genetic background effect among tested Drosophila lines , all stocks were crossed with w- and then crossed to the second ( w-; Bc , Elp/CyO ) or third ( w-; D/TM3 , Sb ) chromosome balancers , respectively . For making double mutants , w-; T ( 2:3 ) ApXa/CyO; TM3 was crossed with the flies containing UAS-X transgene to produce w-; UAS-X/CyO; +/TM3 . Then , w-; +/CyO; UAS-Y/TM3 flies generated by the similar way were crossed with w-; UAS-X/CyO; +/TM3 to produce w-; UAS-X/CyO; UAS-Y/TM3 . Drosophila BG2-c6 cells established by the single colony isolation of primary cells derived from the third instar larval central nervous system . This cell line synthesizes acetylcholine and expresses insect neuron specific glycans and a RNA-binding protein Elav [19] . BG2-c6 cells purchased from the Drosophila Genomics Resource Center ( DGRC , Indiana University ) were maintained at 26°C in Schneider medium supplemented with 10% bovine calf serum . Immortalized GT1-7 mouse hypothalamic neurons [22] were cultured in 4 . 5 g/l glucose Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% fetal bovine serum , 2% of l-glutamine , 100 µU/ml penicillin and 100 µg/ml streptomycin in 5% CO2 at 37°C . The culture medium was changed every 2–3 days . Before peptide treatments , cells were starved for 8 h in the serum-free medium containing 0 . 5% BSA and pretreated with a chemical inhibitor or vehicle . PKA inhibitor H89 ( 10 µM , Calbiochem ) , ERK-specific kinase MEK inhibitor PD98059 ( 10 µM , Calbiochem ) , PKC inhibitor Chelerythrine chloride ( 1 µM , Sigma ) were used . NPY1R inhibitor BIBO3304 ( 10 nM ) , NPY2R inhibitor BIIE0246 ( 50 nM ) , NPY5R inhibitor CGP71683 ( 1 µM ) and Sirt1 inhibitor EX527 ( 10 µM ) were purchased from Tocris . Then , cells were treated with 100 nM synthetic 19 amino acids sNPF2 or 100 nM human NPY 1–36 peptide ( Sigma ) . For transfection , cells were cultured in the growth medium without antibiotics and transfected with small interfering RNA ( siRNA ) using Lipofectamine 2000 reagent ( Invitrogen ) . Gαs and Gαi siRNA constructs were designed by the BLOCK-iT RNAi Designer and Dyrk1a siRNA was purchased from Invitrogen . The sequences of siRNA are caggauauucuucggugccguguuu for Gαs and cggcgggauacuaucuaaauucgcu for Gαi . The BLOCK-iT Fluorescent Oligo , which is a fluorescent-labeled dsRNA oligomer , was used as the non-targeting siRNA control . For the overexpression mouse Dyrk1a , a full-length mDyrk1a cDNA was cloned to pCDNA3 . 1 ( Invitrogen ) . We measured food intake of Drosophila in two ways . The CAFE assay [17] was performed with 3 day-old adult male flies . Twelve hours before the assay , ten flies were placed in the CAFE device [17] containing 5% sucrose solution in calibrated glass micropipettes ( VWR , West Chester , PA ) . At time zero , the micropipettes with 5% sucrose solution were replaced and the amount of liquid consumed was measured every 6 h . A colorimetric food intake assay was modified from published methods [2] , [53] . Since flies had most fed color food in the crop during first 30 min and started to excrete from 1 h ( Figure S4C , S4D ) [54] , flies were starved in PBS-containing vials for 2 h and fed for 30 min in vials containing 0 . 05% Bromophenol Blue dye and 10% sucrose in yeast paste . Then , the flies were frozen , homogenized in PBS , and centrifuged twice for 25 min each . The supernatant was measured at 625 nm . Each experiment consisted of 20 flies , and the assay was repeated at least three times . Dyrk1a transgenic mice expressed the human Dyrk1a BAC clone in the C57BL/6 background [55] . Seven weeks-old male Dyrk1a TG and littermate control mice were used in the experiments ( n = 7 ) . The mice were housed individually in the standard plastic rodent cages . They were maintained at 22±2°C in a room with a 12-hour light/dark cycle and habituated to frequent handling . Food intake and body weight were measured within 30 min before the light turned on and off . Drinking water was available at all times . Food intake data were corrected with body weight . Animal care and all experiments were conducted according to KRIBB Guidelines for the Care and Use of Laboratory Animals and Inje University Council . Twenty w- female flies were starved overnight and fed for 2 h for the physiological synchronization . Then , starvations for the experiment were started . The heads from starved flies were collected for the Quantitative RT-PCR analysis . The experiments were repeated three times . Eggs laid by five female flies for 6 h at 25°C were cultured to avoid over-crowding and lack of nutrition . For weight of individual fly , over 50 three day-old adult male flies were measured with the balancer ( METTLER AJ100 ) and divided with the number of flies . At least three experiments were performed in each assay . Adult heads from 20 flies were collected for RNA preparation . Total RNA was extracted using the easy-BLUE ( TM ) reagent ( iNtRON biotechnology ) . All RNA samples were treated with RNase-free DNase ( Promega ) . cDNA was synthesized using a SuperScript III First-Strand Synthesis System ( Invitrogen ) . For quantitative RT-PCR analysis , ABI Prism 7900 Sequence Detection System ( Applied Biosystems ) and SyberGreen PCR Core reagents ( Applied Biosystems ) were used . mRNA levels were expressed as the relative fold change against the normalized rp49 mRNA . The comparative cycle threshold ( Ct ) method ( User Bulletin 2 , Applied Biosystems ) was used to analyze the data . All experiments were repeated at least three times . The statistical significance was tested by Microsoft Excel-based application for the student t-test statistical analysis . Primers used in the RT-PCR analyses were listed in Table S2 . Minibrain antiserum was generated by the immunization of rabbits with the synthetic peptide ( CQHRVRNWPTNGNQ ) corresponding to the N-terminal sequence ( 75–88 ) of the Minibrain-H protein . Antiserum against sNPFR1 was generated by the immunization of rat with the synthetic peptide ( GEAIGAGGGAELGRRIN ) corresponding to the C-terminal sequence ( 585–600 ) of the sNPFR1 protein . For immunostaining , adult brain from newly eclosed flies ( 3 day old ) was dissected in PBS , fixed in 4% paraformaldehyde , and blocked in 5% BSA and 5% normal goat serum . Primary antibodies were incubated two days at 4°C and secondary antibodies were incubated for 2 h at room temperature . The tissues were mounted in the DABCO solution ( 70% glycerol , 2 . 5% DABCO , Sigma , St Louis , MO ) and fluorescence images were acquired by FluoView confocal microscope ( Olympus ) . sNPF ( 1∶200 ) , sNPFR1 ( 1∶200 ) , and Minibrain ( 1∶200 ) primary antibodies , and anti-rat IgG Alexa 488 , anti-rabbit IgG Alexa 488 , or Alexa 594 ( 1∶200 , Molecular Probes ) and anti-guinea pig Cy5 ( 1∶200 , Jackson ImmunoResearch ) secondary antibodies were used . The cells were lysed by the Lysis buffer ( Cell signaling ) containing NaF , PMSF and Na3VO4 . Total cell lysates were immunoprecipitated with Sirt1 antibody ( Cell signaling ) and protein A-agarose ( Pierce ) . The immunoprecipitates were washed three times with Lysis buffer and solubilized in the SDS sample buffer ( 63 . 5 mM Tris-HCl; pH 6 . 8 , 2% w/v SDS , 10% glycerol , 50 mM DTT , 0 . 01% w/v bromphenol blue ) . Western blot analyses were performed as described previously [2] . Phospho-CREB , phospho-Threonine , FoxO1 ( 1∶1000 , Cell signaling ) , Ac-FKHR ( 1∶1000 , Santa Cruze ) , β-actin ( 1∶3000 , Abcam ) primary antibodies , and horseradish peroxidase-conjugated anti-rabbit IgG ( 1∶5000 , Santa Cruze ) and anti-mouse IgG secondary antibody ( 1∶5000 , Sigma ) were used . Intracellular cAMP was measured with the cAMP Biotrak Enzyme Immunoassay Kit ( GE Healthcare ) by the manufacturer's instruction . Briefly , samples were incubated with anti-cAMP antibody , which was immobilized in the secondary antibody coated micro-plates . Following enzyme substrate conversion , an optical density was measured at 450 nm with microplate reader ( Fluostar Optima , BMG labtech ) . cAMP concentration was expressed as the cAMP pM per mg of protein and converted to the fold change relative to the basal control value . About 250 of 3-day-old W[DAH] female flies were collected after 12 h starvation . Then , flies were homogenized and cross-linked in 1X PBS containing 1% formaldehyde . The ChIP protocol was performed as described in Teleman et al . [56] . Immunoprecipitation was performed using Dynal protean G beads ( Invitrogen ) and anti-dFOXO antibody ( a gift from Heather Broihier ) . Purified DNA was amplified and labeled following Affymetrix ChIP Assay Protocol . Drosophila Tiling 2 . 0R Array was used to detect dFOXO binding enrichment . For ChIP-PCR analysis , about 108 of BG2-c6 cells were treated with sNPF2 peptide as described above . sNPF-treated and untreated cells were cross-linked with 1% formaldehyde . After immunoprecipitation with the CREB antibody ( Cell signaling ) and Protein A Sepharose CL-4B ( GE Healthcare ) , quantitative RT-PCR analysis was performed using input DNA and immunoprecipitated DNA for the CREB binding site in the mnb promoter region and the 3rd axon of Actin5C . Values in the paper are presented as means ± s . e . m . Statistical significant of all data were evaluated by the One-way ANOVA test ( GraphPad Prism software ) . P<0 . 05 was accepted as statistically significant .
|
Feeding behavior is one of the most essential activities in animals . Abnormal feeding behaviors cause metabolic syndromes including obesity and diabetes . Neuropeptides regulate feeding behavior in animals from nematode to human . Here , we presented molecular genetic evidences of how neuropeptides regulate food intake using fruit fly and mouse model systems . Drosophila short neuropetide F ( sNPF ) and the mammalian functional homolog neuropeptide Y ( NPY ) are produced from neurons in the brain of fruit fly and mouse , respectively . These neuropeptides turned on the minibrain , in mammals also called Dyrk1a , a target gene through the PKA-CREB pathway . Then , this Mnb/Dyrk1a enzyme activated Sir2/Sirt1 enzyme , which activated FOXO transcriptional factor , turning on the expression of a sNPF/NPY target gene . The increased sNPF/NPY increased food intake in fruit flies and mice . On the contrary , increased food intake induced insulin and activated insulin signaling . When insulin signaling is activated , FOXO transcriptional factor inhibited expression of a sNPF/NPY target gene . The inhibited sNPF/NPY reduced food intake . These findings indicate that FOXO transcription factor acts as a gatekeeper for fasting–feeding transition by regulating sNPF/NPY expression in Drosophila and mammals .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"animal",
"genetics",
"gene",
"regulation",
"neuroscience",
"gene",
"function",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"molecular",
"genetics",
"biology",
"mouse",
"molecular",
"biology",
"signal",
"transduction",
"phenotypes",
"heredity",
"genetics",
"molecular",
"cell",
"biology",
"behavioral",
"neuroscience",
"neurophysiology",
"genetics",
"and",
"genomics"
] |
2012
|
Minibrain/Dyrk1a Regulates Food Intake through the Sir2-FOXO-sNPF/NPY Pathway in Drosophila and Mammals
|
In order to be transmitted , a pathogen must first successfully colonize and multiply within a host . Ecological principles can be applied to study host-pathogen interactions to predict transmission dynamics . Little is known about the population biology of Salmonella during persistent infection . To define Salmonella enterica serovar Typhimurium population structure in this context , 129SvJ mice were oral gavaged with a mixture of eight wild-type isogenic tagged Salmonella ( WITS ) strains . Distinct subpopulations arose within intestinal and systemic tissues after 35 days , and clonal expansion of the cecal and colonic subpopulation was responsible for increases in Salmonella fecal shedding . A co-infection system utilizing differentially marked isogenic strains was developed in which each mouse received one strain orally and the other systemically by intraperitoneal ( IP ) injection . Co-infections demonstrated that the intestinal subpopulation exerted intraspecies priority effects by excluding systemic S . Typhimurium from colonizing an extracellular niche within the cecum and colon . Importantly , the systemic strain was excluded from these distal gut sites and was not transmitted to naïve hosts . In addition , S . Typhimurium required hydrogenase , an enzyme that mediates acquisition of hydrogen from the gut microbiota , during the first week of infection to exert priority effects in the gut . Thus , early inhibitory priority effects are facilitated by the acquisition of nutrients , which allow S . Typhimurium to successfully compete for a nutritional niche in the distal gut . We also show that intraspecies colonization resistance is maintained by Salmonella Pathogenicity Islands SPI1 and SPI2 during persistent distal gut infection . Thus , important virulence effectors not only modulate interactions with host cells , but are crucial for Salmonella colonization of an extracellular intestinal niche and thereby also shape intraspecies dynamics . We conclude that priority effects and intraspecies competition for colonization niches in the distal gut control Salmonella population assembly and transmission .
The Salmonella enterica serovars are important pathogens that cause disease ranging from a self-limiting gastroenteritis to persistent systemic infections . The human-adapted Salmonella enterica Typhi and Paratyphi serovars are the causative agents of typhoid fever , and penetrate the intestinal epithelium to disseminate to systemic tissues [1] . Approximately 1–6% of infected patients become chronic carriers and serve as the reservoir of disease , remaining asymptomatic while excreting Salmonella in their stool [1] , [2] . S . Typhimurium causes a typhoid-like disease in mice , but also infects a wide-range of mammalian hosts , including livestock [3] , [4] . S . Typhimurium is a major cause of foodborne diarrheal disease in humans , but can also cause invasive non-typhoidal Salmonella ( NTS ) disease in immunocompromised individuals [5] , [6] . NTS can persist in the gastrointestinal tract and be excreted in feces in certain patients [7] , with elevated levels of NTS fecal shedding associated with antibiotic therapy [8] . Surprisingly little is known about Salmonella fecal shedding dynamics , particularly during persistent infection . However , this aspect of the Salmonella life cycle is fundamentally important for understanding transmission to new hosts . Transmission of this enteric pathogen occurs via the fecal-oral route . During invasive disease with host-adapted serovars , Salmonella invade the Peyer's patches ( PP ) in the small intestine and breach the epithelium . Trafficking through the blood and lymphatics results in systemic dissemination of the pathogen to the mesenteric lymph nodes ( mLN ) , bone marrow , spleen , liver , and gallbladder [9] . It is thought that systemic Salmonella in gallbladder bile secretions reseed the small intestine to be transmitted in feces [1] , [10] . However , the fate of the initial invading Salmonella in the intestine and whether they contribute to fecal shedding has not been determined . A deeper understanding of the within-host population biology of Salmonella infections is crucial for determining treatment strategies and preventing spread . The mammalian host can be viewed as an ecosystem , with different tissues functioning as interconnected habitats . In this landscape , pathogens develop into population structures based on processes of dispersal , diversification , environmental selection , and coevolution within the host [11] . During host-to-host spread , each individual acts as an independent ecosystem , and a pathogen must adapt to a new environment in order to be successfully transmitted . Principles in ecology can thus be applied to explain and predict the resulting infection dynamics [11] , [12] . Since host-adapted Salmonella serovars first enter the gastrointestinal tract before spreading to systemic tissues , we hypothesized that distinct groups of communities would assemble within these two host compartments . In population ecology , this is referred to as a subpopulation , or a local group of individuals that interact within a certain habitat [13]–[17] . A metapopulation then consists of a collection of subpopulations with various interactions and rates of dispersal between their habitats . Indeed , studies utilizing tagged isogenic strains have revealed formation of metapopulations in other systemic infections . Due to differing replication rates and dispersal routes within host tissues , independent pathogen subpopulations form during Listeria monocytogenes , Yersinia pseudotuberculosis , and uropathogenic Escherichia coli infections [18]–[21] , although the impact of these subpopulations on transmission is unknown . Wild-type isogenic tagged Salmonella ( WITS ) strains have been developed to resolve the early kinetics of acute infection in the susceptible C57BL/6 mouse background . In the streptomycin-treated diarrhea model , WITS were applied to generate a mathematical model describing replication and immune clearance of Salmonella in the cecal lymph node 24 hours post-infection [22] . Analysis of an intravenous model of infection revealed that concomitant death and rapid bacterial replication resulted in the formation of independent WITS subpopulations in the liver and spleen , although hematogenous mixing led to the homogenization of these systemic communities after 48 hours [23] . A study of early dissemination determined that founder bacteria initiated infection independently in Peyer's patches and systemic compartments 4 days post-infection [24] . However , the WITS technique has not been utilized to dissect the spatiotemporal population dynamics during chronic infections . It is not known whether different subpopulations of Salmonella form during persistent infection , or how they contribute to the pool of Salmonella that is ultimately shed in the feces . Furthermore , it is important to determine whether Salmonella that are carried long-term in systemic tissues and/or in the gallbladder contribute to fecal shedding in the presence of a previously established intestinal subpopulation . The effect of an established intestinal subpopulation on subsequent super-infections is also unclear . However , this scenario could arise in endemic regions and outbreaks , and therefore has implications on human disease and livestock husbandry . It is also unclear whether humans can be co-infected with multiple Salmonella strains due to difficulties in obtaining consistent patient samples , but this scenario could arise in endemic regions and outbreaks . Studies in ecology have determined that immigration order dictates community structure through a priority effect , in which early colonization affords one member an advantage over future colonizers [25]–[27] . These competitive interactions are often mediated by resource availability [26]–[28] . Darwin's naturalization hypothesis posits that challenging species are more successful in habitats in which their close relatives are absent [29] , as the more closely related they are , the more strongly they will compete for the same resources . Following this logic , we hypothesized that different subpopulations of Salmonella will compete for colonization of niches important for fecal shedding . In this study , we employed tagged isogenic S . Typhimurium strains in a mouse model of persistent systemic infection . We show that a Salmonella metapopulation structure forms during persistent infection , with distinct subpopulations in intestinal and systemic tissues . We further found that established subpopulations of intestinal Salmonella colonize crucial extracellular niches in the cecum and colon that are required for fecal shedding . Systemic Salmonella from the gallbladder , as well as challenging strains from other infected donor mice , are excluded from the distal gut niche in a novel observation of intraspecies colonization resistance by an enteropathogen . Salmonella hydrogenase , an enzyme that mediates acquisition of microbiota-derived hydrogen [30] , is required to exert priority effects in this crucial transmission niche . In addition , we demonstrate that maintenance of this intraspecies colonization resistance is dependent on the Salmonella pathogenicity islands SPI-1 and SPI-2 during persistent infection .
To define the Salmonella population structure that arises during chronic infection , we employed a previously established tagged strain approach using a mixture of barcoded , phenotypically equivalent S . Typhimurium strains [23] . These Salmonella wild-type isogenic tagged strains ( WITS ) each carry a unique 40 base pair tag in between the malX and malY pseudogenes , are equally fit , and have been applied to studies of acute infection [23] . Utilizing these previously published sequence tags , we constructed 8 WITS strains in the S . Typhimurium SL1344 background ( W1–W8; Table S1 ) and confirmed each strain to be equally fit when grown in broth culture ( Figure S1A ) . 129X1/SvJ mice , which possess a wild-type Nramp1 allele and can be persistently colonized with S . Typhimurium [31]–[33] , were orally inoculated with 108 colony forming units ( CFU ) of an equal mixture of strains W1–W8 ( Figure S1B–C ) . Total WITS CFU were enumerated by plating ( Figure S1D ) and qPCR was performed to determine the WITS abundances in systemic ( spleen , liver , gallbladder ) and intestinal ( PP , small intestine , cecum , colon , feces ) sites after 35 days of infection . Individual mice had WITS profiles that were distinct from other animals , with certain WITS comprising the majority of Salmonella found within infected tissues that varied on a mouse-by-mouse basis ( Figure 1A ) . However , combined analysis of all infected mice revealed that all 8 WITS strains were represented in every tissue compartment ( Figure 1A ) and there was no statistically significant difference between the relative abundances of the WITS strains in each of the tissues , indicating all 8 WITS are equally represented in vivo ( Table S2 , one-way ANOVA and Kruskal-Wallis tests ) . A control experiment in which 4 of the 8 WITS were underrepresented in the inoculum resulted in their subsequent underrepresentation within infected tissues ( Figure S2 ) , indicating that these 4 WITS did not have any fitness advantage during infection . The WITS compositions in systemic and intestinal tissues were compared in order to determine whether Salmonella subpopulations arose after 35 days of persistent infection , a time after which the bacteria have breached the intestinal epithelium and have spread systemically to the liver and spleen . The strain composition within individual mice varied depending on the site of infection ( Figure 1A ) . In order to quantify potential differences in WITS abundances , we utilized a Bray-Curtis dissimilarity statistic , which has been commonly used in community abundance analyses in ecology and studies of the microbiota [34]–[36] . This calculation was applied to our model to obtain population-level distance values of WITS compositions in different sites . Bray-Curtis values were calculated between the WITS relative abundances of two tissues ( see Materials and Methods ) , in which a score of 0 indicates an identical WITS profile in both organs and a score of 1 indicates completely dissimilar populations . A dissimilarity matrix was calculated for all tissue comparisons ( Table 1 ) . The subpopulation in the liver closely matched that of the spleen with a low mean dissimilarity score of 0 . 248 ( Figure 1B , Table 1 ) , which is consistent with these environments being highly connected by migration pathways through the bloodstream and/or lymphatics . In addition to colonizing systemic sites , Salmonella persisted within intestinal tissues for 35 days . However , in contrast to the spleen and liver , which contained an average of 3–4 WITS , the intestinal tissues were colonized by 1–2 strains ( Figure 1A ) . This suggested that while there was some bottlenecking in dissemination to systemic sites , stronger selection pressures likely existed within intestinal tissues . Further analysis of the WITS profiles indicated that the strain compositions in proximal gut tissues ( PP and small intestine ) were dissimilar from those present in distal gut tissues ( cecum and colon , Figure 1A ) , with dissimilarity scores of 0 . 416–0 . 575 ( Figure 1B , Table 1 , Figure S3A ) . In contrast , the WITS composition in the cecum and colon were very similar with a score of 0 . 101 ( Figure 1B , Table 1 ) , which was significantly lower than the dissimilarity scores observed in the proximal gut ( Figure S3A ) . Together , these data suggest that during persistent infection , different subpopulations of Salmonella form between proximal and distal gut tissues . It is thought that Salmonella in the liver and gallbladder reseed the intestinal tract via bile , followed by subsequent shedding in the feces . If the bile ducts provided highly connected migration pathways between these sites , the WITS profiles should be similar between systemic and intestinal tissues . Although not all of the mice were colonized by Salmonella in the gallbladder ( Figure 1A ) , the WITS profiles in the gallbladder were most similar to the compositions of the spleen and liver from these mice ( Table 1 , Figure S3B ) . In contrast , the WITS compositions in the gallbladder were very different from the composition within the intestinal tissues ( Figure 1B , Table 1 , Figure S3B ) . In addition , the WITS compositions in the distal gut were distinct from those in the systemic tissues with high dissimilarity scores >0 . 816 ( Figure 1B , Table 1 ) . Collectively , analysis of the WITS compositions in various compartments within each infected mouse demonstrate that spatially delimited Salmonella subpopulations form during persistent infection , with systemic organs containing populations that are distinct from those in intestinal tissues . Since host-to-host transmission requires high levels of Salmonella shed in the feces [4] , [33] , we wished to elucidate the kinetics and population dynamics of Salmonella shedding . Fecal samples were collected at various time points throughout the 35-day infection period ( Figure 2A ) . An average of 6–7 WITS were present in feces after one day of infection , indicating some initial bottlenecking effects in the oral infection route may have occurred ( Figure 2A–B ) . However , even greater dynamic changes in WITS compositions were observed at early time points in infection , with different strains shed at 7 and 14 days post-infection compared to day 1 ( Figure 2A ) . Importantly , there was a dramatic decrease in the number of strains detected in the feces to an average of 1–2 WITS , which did not change during the 35-day infection ( Figure 2A–B ) . Importantly , the sharp decrease in the number of strains shed in the feces on day 7 correlated with an increase in total fecal Salmonella CFU ( Figure 2B ) , suggesting that clonal expansion of dominant WITS strains was responsible for increased fecal shedding . To ascertain the tissue compartment that served as the source of clonal Salmonella expansion , we compared the WITS relative abundance profiles of the feces to both systemic and intestinal tissues to identify similarities . Although Salmonella initially invade the PP , the WITS compositions in the PP compared to the feces were significantly different at 35 days post-infection ( Figures 2C , Table 1 ) . In addition , the compositions of the Salmonella populations within systemic sites compared to the population composition in the feces were even more dissimilar ( Table 1 ) . This further corroborated our earlier finding that distinct Salmonella subpopulations arose between systemic and intestinal compartments . Furthermore , we did not observe an increase in the number of WITS strains present during increased fecal shedding ( Figure 2B ) , which would be expected to occur if increased reseeding of systemic Salmonella was the source . Instead , these analyses revealed that the WITS profiles in both the cecum and colon very closely matched the composition of Salmonella shed in the feces ( Figure 2B; Table 1 ) . Importantly , the dissimilarity values between the distal gut sites and the feces were significantly lower than that of any other tissue compartment analyzed ( Figure 2C , Table 1 , Figure S3 ) . Taken together , our results indicate that a clonal expansion of cecal and colonic Salmonella is responsible for the increases in fecal shedding . The results of our WITS experiment demonstrated that distinct subpopulations formed in systemic and intestinal tissues by 35 days post-infection ( Figure 1 ) . However , even though high Bray-Curtis scores were computed between systemic and distal gut tissues , values were <1 indicating there were small percentages of shared WITS in these sites . One limitation of our mixed inoculum approach was that we could not discern the directionality of dissemination . For example , it could not be determined whether WITS present in the distal gut were part of the initial population or if they arrived secondarily by seeding the intestinal tract from systemic sites . In order to determine the relative contribution of systemic and intestinal strains to fecal shedding , it required a strategy to mark Salmonella in these different sites within the host . To address this , we developed a co-infection model that employed isogenic marked strains rapidly identifiable by differential plating on antibiotics . We used the parental streptomycin-resistant SL1344 strain that has a missense mutation in hisG , which is not required for virulence , and an isogenic SL1344-kanR strain containing a kanamycin resistance cassette inserted at this site ( hisG::aphT ) . These strains are equally fit in single and in mixed infections in mice inoculated by oral or IP routes [33] . In our co-infections , each mouse received 108 of one strain by oral inoculation and 103 of the isogenic strain by intraperitoneal ( IP ) injection . The IP route bypasses the gastrointestinal tract , such that Salmonella colonize systemic tissues first [32] . To confirm that successful reseeding occurs in our model , Salmonella shedding and tissue burdens were compared in control mice that received single IP infections or those that received single oral infections . Systemic IP-delivered Salmonella reseeded the small intestine , where they reached the same range of fecal shedding levels by 14 days post-infection as mice infected orally ( Figure S4 ) . However , the oral inoculation route resulted in >1 , 000-fold more Salmonella fecal CFUs 1 day post-infection compared to the IP route , and reached peak fecal shedding levels more rapidly ( Figure S4A ) . Thus , in the co-infection model , the oral strain establishes an infection in the gut before the systemic strain reaches the intestine , allowing us to test the strength of priority effects in Salmonella population assembly . Mice injected IP with a single Salmonella strain shed this strain in the feces as soon as 1 day post-infection ( Figure S4A ) . This was in contrast to what occurred in mice that had been co-infected orally with an isogenic WT strain ( Figure 3 ) . The systemic strain was detected in the feces of only 5 of the 54 mice throughout the 30 days of infection ( Figure 3A–C ) . Importantly , shedding of the systemic strain only occurred on a single day and did not persist . Since mice shed variable levels of Salmonella [33] , [37] , we wondered whether this would influence the ability of the IP strain to be shed . Surprisingly , the oral strain was exclusively shed in the feces of low ( <104 CFU/gram ) , moderate ( <108 CFU/gram ) , and super ( ≥108 CFU/gram ) shedder mice ( Figure 3A–C ) . In addition , when the reciprocal combination of strains ( oral: SL1344-kanR , IP: SL1344 ) was used the same result was obtained throughout 60 days of infection ( Figure 3 , Figure S5A ) . Taken together , these results indicate that the established intestinal strain prevents colonization of the cecum and colon by Salmonella disseminating from systemic tissues . We next wondered what the composition of the Salmonella strains were within systemic tissues of mice that had been co-infected for 30 days . In contrast to the cecum and colon , the IP and oral strains were both present within systemic tissues after 30 days of co-infection . The spleen and liver were comprised of similar abundances of both strains ( Figure 4 ) , indicating that intestinal Salmonella effectively disseminated to systemic sites . Although the orally inoculated strain was present in the gallbladder , the IP strain comprised >80 . 44% of the total Salmonella population in this organ ( Figure 4 ) . In addition , the IP strain was present as a minority of the population present in the PP ( 19% ) , small intestine ( 30% ) , and mLN ( 38% ) ( Figure 4 ) . The IP strain was not detected in the cecum and colon in 25 out of 28 mice , and comprised <8% in the remaining animals ( Figure 4 ) . Strikingly , the oral strain remained dominant in the cecum and feces during the 60-day infection ( Figure S5 ) . Thus , our results from the co-infection model and the WITS analyses suggest that Salmonella that are established in the cecum and colon prevent systemic subpopulations from colonizing important niches that are required for fecal shedding . One possible explanation for the dominance of the oral strain in the distal gut and feces could be that there is insufficient reseeding of systemic Salmonella into the gastrointestinal tract . To test this possibility , we utilized an established gallstone model of infection , in which S . Typhimurium biofilm formation on gallstones increased reseeding and subsequent fecal shedding by 1 , 000-fold [38] . We fed mice a lithogenic diet for 10 weeks to induce gallstone formation that resulted in 1–9 stones/mouse as confirmed by ultrasound imaging ( Figure S6 ) . In contrast , mice on a standard diet never developed gallstones ( Figure S6 ) . As previously demonstrated , mice with gallstones that were infected with 103 S . Typhimurium by IP injection shed >1 , 000-fold higher levels of bacteria 7 days post-infection compared to control mice ( Figure S7 ) . To determine whether increased levels of S . Typhimurium in the gallbladder would allow systemic bacteria to colonize the cecum and/or colon , mice with diet-induced gallstones were co-infected orally with SL1344 and IP with SL1344-kanR . By 14 days post-infection , mice with gallstones had a mean Salmonella gallbladder burden >10 , 000-fold higher than mice without gallstones ( Figure 5A ) . This represented an increase in systemic Salmonella , as the gallbladders were exclusively colonized by the IP strain ( Figure 5B ) . In addition , mice with diet-induced gallstones had significantly higher levels of S . Typhimurium in the small intestine , indicating that increased numbers of systemic bacteria had reseeded this site ( Figure 5B ) . Despite this drastic increase in the levels of systemic Salmonella reseeding the intestine , the established intestinal strain remained dominant in the cecum , colon , and feces ( Figure 5C ) . Taken together , our data suggest that in the presence of an established Salmonella strain , systemic Salmonella are excluded from colonizing crucial transmission niches in the distal gut . Although the presence of gallstones increased the numbers of S . Typhimurium in the gallbladder to 103–107 CFU/organ as well as subsequent reseeding of the small intestine , it is possible that these levels were insufficient to compete with the established intestinal subpopulations ( Tables S3 , S4 ) . Indeed , we have measured the levels of Salmonella in gastrointestinal sites and found that there is a range of 101–108 total CFU ( Table S3 ) . To address this issue , we performed sequential infections in which resident intestinal Salmonella were challenged with a high oral dose of a second strain . First , mice were inoculated with 103 SL1344 by IP injection to establish a systemic infection . This initial Salmonella strain was detected in the feces after 5–7 days and was persistently shed for 35 days ( Figure 6A ) . These mice were then super-infected with 108 SL1344-kanR orally . Although , the orally inoculated strain was detected in the feces 1 day post-infection ( dpi ) , it was not detected in the feces for the remaining 7 days post-oral inoculation ( 35–42 dpi , Figure 6A ) . The challenging oral strain was not detected in any systemic or intestinal tissues by 7 days post-challenge ( 42 dpi , Figure S8A ) . This demonstrates that super-infecting strains are excluded from colonizing the intestine in the presence of a resident , persistent intestinal Salmonella infection , regardless of the route of inoculation . Collectively , our results suggest that there is intraspecies competition for a transmission niche in the distal gut . Based on our evidence of intraspecies competition for a distal gut niche , we proposed that the dominance of established Salmonella in the cecum and colon is attributed to priority effects that govern distal gut colonization and subsequent fecal shedding . To test this notion , we performed sequential S . Typhimurium infections to evaluate the duration and strength of these competitive interactions . Mice were infected with 108 SL1344 orally , and fecal shedding of Salmonella was monitored . All of the mice continued to shed Salmonella over the 102 days of infection ( Figure 6B ) . After 102 days , the mice were inoculated orally with 108 CFU of a second competing strain , SL1344-kanR . The competing strain was detected in the feces during the first 3 days post-infection ( Figure 6B ) . However , by 14 days post-challenge ( 116 dpi ) , 28 of the 45 mice were no longer shedding the competing Salmonella strain ( Figure 6B ) . Finally , by 35 days post-challenge ( 137 dpi ) , the competing strain was not detected in the feces ( Figure 6B ) , intestinal compartments , or systemic tissues of co-infected mice ( Figure S8B ) . The reciprocal strain combinations were also tested: mice were first infected with 108 SL1344-kanR orally for 60 days before subsequent challenge with 108 SL1344 , in which the competing strain was cleared from the feces by 20 days post-challenge ( Figure S9A ) . Thus , this colonization resistance against the same Salmonella species was maintained during the chronic stages of infection . We next sought to determine whether the levels of the initial oral strain ( SL1344 ) in the colon and in the feces would influence the clearance kinetics of the second competing oral strain ( SL1344-kanR ) . One day after the second oral inoculation , the percentage of the competing strain varied depending on the level of shedding of the resident strain . For example , in mice that were shedding >108 CFU/g feces ( super shedder mice ) , the competing SL1344-kanR strain comprised 4 . 88% of the total population on the first day post-secondary inoculation ( Figure S9B ) . In contrast , for low and moderate shedder mice , the competing strain comprised 42 . 02% and 35 . 58% of the total population , respectively ( Figure S9B ) . These differences remained significant 5 days after infection with the second competing SL1344-kanR strain . However , by days 10 and 14 , the second SL1344-kanR strain was no longer detected in the feces of any of the mice ( Figure S9B ) . These results indicate that more robust and rapid priority effects are exhibited in mice that are colonized with higher colonic Salmonella loads . Finally , to determine whether we would see the same intraspecies priority effects in the distal gut during host-to-host transmission , we utilized our previously established model of transmission from an infected , super shedder mouse to uninfected mice in the same cage [33] . In this experiment , the donor mouse was orally infected with SL1344-kanR and was shedding >108 CFU/g at 14 days post-inoculation ( Figure 6C ) . As a positive control for host-to-host transmission , the donor mouse was co-housed with uninfected mice . Similar to our previous results , naïve mice began shedding SL1344-kanR within 24 hours and continued to shed even after the donor was removed ( Figure S10A ) . In contrast , recipient mice that had been infected for 14 days with SL1344 required 10 days of cohousing before low levels of the donor strain ( <0 . 02% of all Salmonella ) were detected in the feces ( Figure 6C ) . In addition , shedding of the donor strain in the previously infected recipient mice was transient , and was not detected in the feces or tissues 10 days post-cohousing ( Figure 6C , Figure S8C ) . Similar results were obtained when a SL1344 super shedder was cohoused with SL1344-kanR infected mice . Cohousing for 12 days was required before the donor strain could be detected in the feces of recipient SL1344-kanR mice ( Figure S10B ) . The super shedder donor was left in the cage for an additional 6 days before removal , but consistent with previous findings , the donor strain was not detected in the feces of recipient mice by day 23 post-cohousing ( Figure S10B ) . Together , these experiments show that priority effects determine Salmonella population assembly in intestinal transmission niches , where established subpopulations exert colonization resistance against incoming challengers . Since the established subpopulation of Salmonella in the cecum and colon exerts colonization resistance , we proposed that their removal would allow challengers to occupy vital transmission niches . To test this idea , mice co-infected with 108 SL1344 orally and 103 SL1344-kanR IP for 7 days were then treated with a single dose of kanamycin . Kanamycin is not absorbed systemically and thus was used to ablate the extracellular , kanamycin-sensitive bacteria in the gastrointestinal tract . Within 24 hours of antibiotic administration , fecal shedding of the established intestinal SL1344 decreased by ∼5 logs ( Figure 7A , left ) . Concomitant with the decrease in the established strain , over 107 CFU of the systemic SL1344-kanR strain was shed per gram of feces ( Figure 7A , left ) . By 4 days post-antibiotic treatment , the systemic strain was exclusively shed in the feces ( Figure 7A , left ) and was transmitted to naïve recipients ( Figure 7A , right ) . Thus , priority effects arose during the first 7 days of infection , which coincided with the clonal expansion in the distal gut and feces observed in the WITS studies ( Figure 2 ) . Based on these findings , we hypothesized that Salmonella strains were competing for limited nutrient or spatial resources within the cecum and colon , which inhibited the ability of systemic strains to colonize the distal gut . We tested this notion by gavaging co-infected mice ( SL1344 oral , SL1344-kanR IP ) with 5 mg streptomycin in order to disrupt the microbiota and make more of these resources available [37] , [39]–[41] . Both Salmonella strains are streptomycin-resistant , and previous studies have shown that streptomycin treatment of infected mice increases Salmonella fecal shedding to super shedder levels [33] , [37] . We observed that all streptomycin-treated mice became super shedders , yet the increase in fecal Salmonella CFU reflected expansion of the oral SL1344 strain ( Figure S11 ) . This indicated that disrupting the microbiota with streptomycin treatment was insufficient to permit shedding of the systemic strain , as the newly available resources were likely immediately utilized by established intestinal Salmonella . Furthermore , since kanamycin does not enter mammalian cells , these results collectively indicate that established intestinal Salmonella occupy an extracellular transmission niche in the distal gut and exclude the bacteria that are reseeding the intestine from systemic sites . Our data show that intraspecies priority effects govern Salmonella population assembly in the distal gut . Since our previous results demonstrated that clonal expansion and priority effects in the cecum and colon could occur by 7 days post-infection ( Figure 2 , 7A ) , we hypothesized that nutrient acquisition was very important during this stage of colonization . Indeed , ecological theory has implicated competition for nutrients as an important determinant in priority effects and community structure [28] . S . Typhimurium hydrogenase ( hyb ) is a key mediator of cecal ecosystem invasion and is required to consume a microbiota-derived metabolite [30] . In the un-inflamed gut of conventional mice with complex microbiota , hydrogenase enzymes facilitate consumption of hydrogen ( H2 ) intermediates in a SPI1- and SPI2-independent manner [30] . Similarly , we show here that Hyb is important for gut colonization and fecal shedding in 129SvJ mice with an intact conventional microbiota ( Figure S12 ) . To test the role of Hyb in intraspecies priority effects , co-infections were performed in which all mice were injected IP with 103 wild-type ( WT ) SL1344 bacteria and one group of mice was co-inoculated orally with 108 ΔhybΔSPI1ΔSPI2 isogenic mutant S . Typhimurium while control mice were co-inoculated orally with 108 WT SL1344-kanR . The hyb mutation was constructed in a ΔSPI1ΔSPI2 background to assess the need for hydrogenase in the context of a non-inflamed gut . The relative levels of each strain in the feces were monitored over 15 days of co-infection ( Figure 7B–C ) . Importantly , the Salmonella shed in the feces at 4 and 7 days contained systemic WT bacteria and by 15 days post-infection were entirely comprised of the systemic WT strain in mice that received ΔhybΔSPI1ΔSPI2 orally ( Figure 7B ) . Total levels of fecal Salmonella were significantly lower in the ΔhybΔSPI1ΔSPI2 co-infection group compared to controls ( Figure S13A ) , which corresponds to the decreased fecal shedding of ΔhybΔSPI1ΔSPI2 mutants during single oral infections ( Figure S12B ) . The strain compositions in the feces of these mice throughout infection indicated that the increase in total fecal CFU on day 15 reflected rapid reseeding and shedding of the systemic WT strain concomitant with declining levels of the oral ΔhybΔSPI1ΔSPI2 strain ( Figure 7C ) . Taken together , we have demonstrated that the hydrogenase mutant was unable to effectively invade the cecal and colonic niche ( Figure 7D ) , thereby nullifying any priority effects and allowing systemic Salmonella to colonize the distal gut with subsequent transmission in feces ( Figure 7C–D ) . To determine whether intraspecies colonization resistance was still dependent on the maintenance of the extracellular intestinal niche during persistent infection , co-infected mice ( oral: SL1344 , IP: SL1344-kanR ) were treated with a single dose of kanamycin 42 days post-infection . Within 24 hours of antibiotic administration , fecal shedding of the established intestinal SL1344 was decreased by ∼6 logs concomitant with a ∼5 log increase in the systemic SL1344-kanR strain ( Figure 8A , left ) . The systemic strain was exclusively shed in the feces 2 days post-antibiotic treatment ( Figure 8A , left ) , and comprised the entire Salmonella population in the cecum and colon after 7 days ( Figure 8A , right ) . These data thus indicate that the extracellular niche in the cecum and colon is required to maintain intraspecies colonization resistance during persistent infection , which actively inhibits successful fecal shedding of systemic Salmonella . To gain more insight into how S . Typhimurium competitively excludes incoming challengers from colonizing the distal gut niche , we tested the potential role of the key virulence factors Salmonella Pathogenicity Islands SPI1 and SPI2 , which encode type III secretion systems that deliver effector proteins required for persistence in host tissues [32] , [42]–[44] and fecal transmission [33] . Co-infections were performed in which mice simultaneously received 103 WT SL1344 by IP and 108 isogenic ΔSPI1ΔSPI2 mutant bacteria orally . In the ΔSPI1ΔSPI2 co-infected mice , the IP-injected WT bacteria were not present in significant numbers at day 7 ( Figure 8B ) . However , by day 25 , 21 . 43% of all fecal Salmonella were WT bacteria , and by day 70 , 98 . 86% were WT S . Typhimurium ( Figure 8B ) . In addition , the total fecal Salmonella CFU in the control ( SL1344 oral , SL1344-kanR IP ) and the ΔSPI1ΔSPI2 co-infected mice were similar ( Figure S13B ) , which is consistent with our result that the systemic WT strain reseeded and replicated within the intestinal tract once the ΔSPI1ΔSPI2 mutant was cleared ( Figure 8C ) . Indeed , examination of strain abundances in intestinal tissues after 70 days of co-infection confirmed that the systemic IP strain had predominantly colonized the mLN , small intestine , cecum , and colon while the initial ΔSPI1ΔSPI2 mutant was cleared from these sites ( Figure 8D ) . These studies demonstrate that SPI1 and SPI2 are required for the established intestinal Salmonella population to maintain active colonization resistance against systemic reseeding bacteria .
Microbial fecal shedding by chronically infected hosts is the major source of new infection and disease for many enteropathogenic microbes . However , very little is known about the dynamics of Salmonella subpopulations within mammalian hosts and what their relative contributions are to host-to-host transmission . Community assembly theory provides a framework for understanding infection processes , and in this study , we defined the S . Typhimurium metapopulation structure that arose during persistent infection . We then applied ecological principles that govern community assembly to determine the contribution of different Salmonella subpopulations to fecal shedding . Our tagged strain approach revealed that distinct S . Typhimurium subpopulations arose within different host tissues , resulting in a metapopulation structure with variable migration between sites . After 35 days of infection , the WITS compositions between the liver and spleen closely matched each other , suggesting that robust migration pathways in the blood and lymphatics exist between these tissues . Previous studies of acute infection in susceptible C57BL/6 mice determined that hematogenous spread 48 hours post-infection resulted in S . Typhimurium mixing between the spleen and liver [23] . Expanding our WITS analyses to include a more comprehensive set of infected tissues , we determined that Salmonella in systemic sites were distinct from subpopulations in the intestinal tract . Interestingly , we found that the WITS profiles in the PP and small intestine were also dissimilar from those in the cecum and colon . This likely represents stochastic invasion of the PP by a subset of individual WITS strains , while different subsets initiate separate infection foci in other tissues . Indeed , a recent study of early infection dynamics determined that PP invasion fueled spread to the mLN while an independent pool of bacteria initiated splenic and hepatic infection [24] . Our work suggests that these initial colonization dynamics shape the metapopulation structure that arises and is maintained throughout persistent infection . Quantifying the differences in systemic , proximal and distal gut sites with Bray-Curtis dissimilarity scores , we were able to gain new insights into the importance of the distal gut as a transmission niche . Surprisingly , we have shown that systemic Salmonella can only colonize the distal gut upon clearance of the established intestinal subpopulation with an oral kanamycin treatment . In contrast , treatment with streptomycin , to which the SL1344 strain is resistant , was insufficient to permit shedding of the systemic strain . This suggests that disrupting microbiota-mediated colonization resistance does not create new niches for systemic bacteria to colonize . Previous studies found that administration of ciprofloxacin killed extracellular Salmonella and permitted tolerant bacteria within dendritic cells of the cecal lymph node to colonize the cecum [45] , [46] . Although this fluoroquinolone treatment also ablated systemic Salmonella [45] , these studies all highlight the intensely competitive dynamics between Salmonella within the distal gut . Competition for gut colonization was also reported with E . coli K12 strains in germ-free mice , although differences in colonization ability was due to varying fitness costs of antibiotic resistances [47] . Our study with isogenic strains support the idea that intraspecies competition for nutrients excludes systemic bacteria from colonizing the distal gut , in which established Salmonella has saturated a required niche . Intraspecies priority effects have recently been described for commensal species of Bacteroides [48] and E . coli [49] , but our findings with an enteropathogen that causes persistent systemic infection is novel . There may also be evidence of these competitive interactions during Yersinia enterocolitica microcolony formation within intestinal tissues , in which previously infected PP were less likely to be super-infected [21] . However , it remains unknown whether colonization can proceed if established Yersinia are eliminated . It is possible that this may be unique to Salmonella rather than a broad enteropathogen phenomenon , as this colonization resistance was not seen between isogenic Campylobacter jejuni strains in a transmission study involving chickens [50] . Host-adapted Salmonella serovars infect the gastrointestinal tract before disseminating to systemic sites such as the gallbladder , which has been classically thought to be the source of Salmonella transmitted in feces [51] . However , the contribution of systemic reseeding in the presence of an established Salmonella intestinal tract infection had never been investigated . We show that an established intestinal strain persisted in the cecum and colon , even when gallstone formation increased gallbladder levels of S . Typhimurium >10 , 000-fold . It is interesting to speculate that intraspecies colonization resistance may occur in other hosts that are persistently infected by Salmonella . For example , humans can carry S . Typhi for long periods of time possibly in the gallbladder [52] . Although gallbladder removal sometimes cures patients , over 20% of carriers continued to shed S . Typhi and S . Paratyphi in their stool [53] , [54] , which indicates an alternative persistent reservoir . While circulating S . Typhi in Kathmandu are resistant to nalidixic acid and several fluoroquinolones , patient gallbladder isolates are more sensitive to nalidixic acid , gatifloxacin , and ofloxacin , indicative of a limited role in typhoid transmission [55] . The relative contributions of systemic versus intestinal populations of S . Typhi to transmission are not known . Perhaps the presence of fecal “showers” of S . Typhi [4] are due to reseeding bacteria from systemic sites that gain access to spatial and nutritional resources in the gut . In our co-infection model , the oral strain comprised 88 . 83% of fecal Salmonella after 60 days , which was lower than the 97 . 96% observed after 30 days ( p = 0 . 06 , unpaired Mann-Whitney ) . Though this was not a significant difference , it is possible that the intestinal strain may lose its dominance at even later time points , at which point systemic Salmonella may reseed from mesenteric lymph node macrophages [31] , [56] , [57] and/or the gallbladder [1] , [2] , [10] . Intraspecies Salmonella colonization resistance could be shaping typhoid epidemiology in endemic regions , but future work is required to determine whether this occurs in other Salmonella serovars besides Typhimurium . We have found that the clonal expansion of the intestinal subpopulation is responsible for increases in S . Typhimurium fecal shedding . The mechanisms by which this subpopulation expands and establishes intraspecies colonization resistance are likely multifactorial . S . Typhimurium fimbriae and adhesins are important for attachment to intestinal tissues [58]–[60] and may play a role in this intraspecies dynamic . Host immune responses contribute to Salmonella clearance [61]–[63] , and could also be involved in influencing intraspecies colonization resistance . However , intraspecies colonization resistance was observed at 14 days post-infection and lasted over 102 days in the context of co-infections , cohousing experiments , and sequential infections . This suggests that neither the innate nor the adaptive immune responses alone could be responsible for the exclusion of systemic reseeding Salmonella . Microbial communities undergo local diversification in different habitats within the host [12] , [64]–[68] , and we considered the possibility that genetic mutations could be responsible for Salmonella expansion and intraspecies colonization resistance . Previous studies with marked isogenic strains determined that spontaneous mutations alone do not shape S . Typhimurium colonization dynamics or fecal transmission during persistent infection in 129Sv mice . The dominance of a re-isolated strain was lost upon subsequent infection or passage in broth , and exhibited the same infectious dose ( ID50 ) as a culture-grown strain [24] , [33] . A study of systemic S . Typhimurium infection revealed that enhanced growth of bacteria were not due to the selection of mutants , but rather were transient phenotypic changes dependent on gene regulation [69] . Systemic Salmonella did not accumulate attenuating mutations during our experiments . This subpopulation adapted to the intestinal environment following ablation of the resident strain , and replicated to supershedder levels with rapid transmission to naïve mice . Salmonella transcriptional responses likely play an important role in expansion in the distal gut , and insight into these changes will elucidate other mechanisms by which priority effects are exerted . Our studies with a hydrogenase mutant revealed that Salmonella competition for a microbiota-derived nutrient is one mechanism by which a challenging systemic strain is excluded from the distal gut transmission niche . According to the monopolization hypothesis , rapid population growth upon colonization of a new habitat results in the effective monopolization of resources , resulting in a strong inhibitory priority effect [70] . Since Salmonella are mainly localized in extracellular regions of the distal gut [33] , it is tempting to speculate that other Salmonella factors required for nutrient acquisition play a role in intraspecies colonization resistance . The importance of nutrient acquisition in establishing priority effects could be applied to the development of novel therapies , in which targeting key metabolic pathways could potentially prevent pathogen colonization and transmission . We have found that SPI1 and SPI2 contribute to intraspecies colonization resistance up to 70 days post-infection . Importantly , co-infected mice that received 108 ΔSPI1ΔSPI2 orally shed significant levels of WT systemic Salmonella beginning 25 days post-infection , with no significant changes in the total fecal shedding of Salmonella . This suggests that as soon as nutrient and/or spatial resources are made available by the clearance of the initial ΔSPI1ΔSPI2 mutant , WT Salmonella spread from systemic tissues and rapidly expand within the intestinal tract . The T3SS encoded by these Salmonella pathogenicity islands deliver over thirty effectors with diverse functions [42]–[44] , [71] . These effectors could act on Salmonella directly , or create an environment that kills strains reseeding from systemic tissues . These mechanisms could involve Salmonella-induced inflammation and modulation of the host immune response [72] . Inflammation also disrupts the host microbiota and allows the pathogen to metabolize newly available nutrients 1 [73]–[76] . Future work will seek to determine which of these are involved in establishing priority effects and exerting intraspecies colonization resistance . Priority effects have long been known to shape community assembly in a variety of ecological systems , ranging from bacteria to larger eukaryotic organisms [27] , [64]–[66] , but this is the first time the phenomenon has been described for pathogen subpopulations during persistent infection within a host . In this landscape , the order in which S . Typhimurium arrive to the intestinal ecosystem dictates which bacteria are subsequently shed in the feces . The results presented herein demonstrate that colonization of distal gut tissues is a bottleneck for successful transmission , which subpopulations of Salmonella compete for . These studies may inform disease processes in host-adapted Salmonella serovars that cause invasive disease , yet are still transmitted fecal-orally . S . Typhimurium is a generalist pathogen that also infects livestock and humans , and thus our work has direct implications on public health [3] , [4] . Our findings also highlight the potential for the application of ecological principles to epidemiology in order to predict dominant circulating strains during outbreaks . This work also sheds light on potential mechanisms that influence human-to-human transmission of non-typhoidal diarrheal infections , which can also be invasive in certain patients [5] , [6] . A better understanding of these mechanisms might reveal novel therapeutic approaches , or even preventive measures in thwarting disease spread .
Experiments involving animals were performed in accordance with NIH guidelines , the Animal Welfare Act , and US federal law . All animal experiments were approved by the Stanford University Administrative Panel on Laboratory Animal Care ( APLAC ) and overseen by the Institutional Animal Care and Use Committee ( IACUC ) under Protocol ID 12826 . Animals were housed in a centralized research animal facility certified by the Association of Assessment and Accreditation of Laboratory Animal Care ( AAALAC ) International . 129X1/SvJ and 129S1/SvImJ mice were obtained from Jackson Laboratories ( Bar Harbor , ME ) . Male and female mice ( 5–7 weeks old ) were housed under specific pathogen-free conditions in filter-top cages that were changed weekly by veterinary personnel . Sterile water and food were provided ad libitum . Mice were given 1 week to acclimate to the Stanford Research Animal Facility prior to experimentation . The S . Typhimurium strains used in this study were derived from the streptomycin-resistant parental strain SL1344 [77] . A missense mutation ( hisG46 ) in SL1344 results in histidine auxotrophy [78] . The isogenic SL1344-kanR strain was created by replacing the hisG coding sequence with that of a kanamycin-resistance casette ( hisG::aphT ) using the methods of Datsenko and Wanner [33] , [79] . Genetic manipulations were originally made in the S . Typhimurium LT2 background before being transferred to SL1344 by P22 transduction . This methodology was also used to construct wild-type isogenic tagged Salmonella ( WITS strains: W1–W8 ) , in which a unique 40-bp signature tag and the kanamycin-resistance cassette were inserted between the malX and malY pseudogenes . Grant et . al . previously established this approach and published the unique 40-bg sequence tags of 8 WITS strains [23] , which were employed in this study ( Table S1 ) . Growth curves of W1–W8 in LB broth cultures were performed by optical density readings and plating for colony forming units ( CFU ) per milliliter ( Figure S1A ) . ΔSPI1ΔSPI2 ( orgA::tet , ssaV::kan ) was generated previously for use in other studies [80] . The Δhyb ( hypOhybABC::cm ) deletion was constructed as described by Maier et . al . , with P22 phage transduction to insert the deleted genomic region into the ΔSPI-1ΔSPI-2 strain ( [30] , Table S1 ) . All constructs were verified by PCR . All S . Typhimurium strains were grown at 37°C with aeration in Luria-Bertani ( LB ) medium containing the appropriate antibiotics: streptomycin ( 200 µg/ml ) , kanamycin ( 40 µg/ml ) , tetracycline ( 15 µg/ml ) and chloramphenicol ( 8 µg/ml ) . For mouse inoculation , an overnight culture of bacteria was spun down and washed with phosphate-buffered saline ( PBS ) before resuspension to obtain the desired concentration . Food was removed 16 hours prior to all mouse infections . In WITS experiments , mice were inoculated with an equal mixture of strains W1–W8 via oral gavage of 108 CFU in 100 µl PBS . For intraperitoneal ( IP ) infections , mice were injected with 103 CFU in 100 µl PBS as previously described [32] . In the co-infection model , mice drank an oral dose of 108 SL1344 in 20 µl PBS , then received an IP injection of 103 SL1344-kanR immediately afterwards . Co-infection experiments were repeated using the reciprocal combination of strains , SL1344-kanR ( oral ) and SL1344 ( IP ) , which had no effect on the trends observed . Individual mice were identified by distinct tail markings and tracked throughout the duration of infection . Between 2–3 fresh fecal pellets were collected directly into eppendorf tubes and weighed at the indicated time points . Pellets were resuspended in 500 µl PBS and CFU/gram feces were determined by plating serial dilutions on LB agar plates with the appropriate antibiotics . Low ( <104 CFU/gram ) , moderate ( <108 CFU/gram ) , and super shedder ( ≥108 CFU/gram ) mice were identified based on previously established criteria [33] , [37] . Following collection of fresh fecal pellets , animals were sacrificed at the specified time points . Blood was collected by cardiac puncture and animals were euthanized by cervical dislocation . Sterile dissection tools were used to isolate individual organs , which were weighed prior to homogenization . The entire gastrointestinal tract was removed , and the small intestine was immediately separated from the distal gut and transferred to a new sterile petri dish . Visible PP ( 3–6/mouse ) were isolated from the small intestine using sterile fine-tip straight tweezers and scalpels . PP , mLN , spleens , livers , and gall bladders were collected in 1 ml PBS . The small intestine , cecum , and colon were collected in 3 ml PBS . Homogenates were then serially diluted and plated onto LB agar containing the appropriate antibiotics to enumerate CFU/gram tissue . For co-infections with SL1344 and SL1344-kanR , several dilutions were plated to ensure adequate colonies ( >100 CFU per sample ) for subsequent patch plating to determine strain abundance . For WITS experiments , 300 µl of tissue homogenate was inoculated into LB broth containing streptomycin ( 200 µg/ml ) and kanamycin ( 40 µg/ml ) as a recovery method to enrich for low abundance strains . An UltroSpec 2100pro spectrophotometer ( Amersham Biosciences , Piscataway , NJ ) was used to obtain optical density readings of the resulting bacterial cultures . Genomic DNA ( gDNA ) was extracted from 2×109 S . Typhimurium from each sample in duplicate using a DNeasy blood and tissue kit ( Qiagen , 69506 ) as per the manufacturer's protocol for Gram-negative bacteria . All qPCRs were performed on an Applied Biosystems 7300 real-time PCR system . A 25 µl reaction contained 12 . 5 µl of FastStart SYBR Green Master Mix with Rox ( Roche , 04913914001 ) , 8 µl DNase/RNase-free water , 0 . 75 µl of forward and reverse ( 10 µM ) primers ( Table S1 ) , and 3 µl of gDNA ( 1–10 ng ) . Standard curves were generated using gDNA from each W1–W8 strain . Reaction conditions were 50°C for 2 min; 95°C for 10 min; 40 cycles of 95°C for 15 s and 60°C for 1 min; followed by a dissociation stage of 95°C for 15 s , 60°C for 1 min , 95°C for 15 s , and 60°C for 15 s . To determine presence of a WITS strain , the qPCR value had to be above a minimum threshold value . This measure of primer specificity was determined by a negative control matrix , in which a specific primer pair was tested on ∼11 . 25 ng of non-template gDNA from each of the other 7 WITS strains . To test primer sensitivity , detection limits were determined by test plates containing known CFU of each strain . Briefly , colonies were washed off the plates with PBS and gDNA was extracted from plates with varying abundances of WITS ( i . e . 1 CFU Strain A with 103–105 CFU Strain B ) . qPCR was performed and revealed a detection limit of 1 CFU/strain amidst over 4800 CFU from non-target strains . To verify that our method of broth recovery and qPCR analyses accurately rendered WITS abundances , we compared relative abundances of an equal mixture of culture-grown W1–W8 as determined by our qPCR strategy versus plating CFU of individual dilutions of each strain ( Fig . S1B ) . Plating onto selective LB agar containing streptomycin ( 20 µg/ml ) and kanamycin ( 40 µg/ml ) was used to determine strain abundances in co-infections , sequential challenges , and transmission experiments . In addition to patch plating a minimum of 100 CFU per sample , undiluted samples were plated on selective plates to increase detection limits . For super shedder mice , this permitted detection of a strain comprising just 0 . 00000001% of the total S . Typhimurium population . The strain relative abundances were determined for each tissue in all of the 19 mice infected with the 108 equal mixture of 8 WITS . The relative abundances of each WITS strain were analyzed by one-way ANOVA ( parametric ) and Kruskal-Wallis ( non-parametric ) tests in Prism statistical software . These analyses were performed for each tissue collected from infected mice . Non-significant P values indicated that a particular WITS was not under or over represented in any tissue type ( Table S2 ) . To further verify that certain WITS strains were not preferentially selected for , a control experiment was performed in which mice were orally infected with an inoculum comprised of a skewed WITS mixture ( Figure S2A ) . Underrepresented strains: W2 , W3 , W5 , W6 ( 4 . 17%–7 . 32% of inoculum ) , overrepresented strains: W1 , W4 , W7 , W8 ( 17 . 39%–20 . 94% of inoculum ) . Relative abundances of WITS in infected tissues were determined by qPCR after 35 days of infection . For each of the 8 WITS , defined bins were constructed for a range of strain relative abundances , with which the observed frequencies were used to generate histograms ( Figure S2B ) . Bray-Curtis dissimilarity scores were computed to quantitatively compare Salmonella population compositions in different sites . The relative abundance ( y ) of each WITS ( n ) was compared between two tissue sites i and j . The Bray-Curtis dissimilarity ( dBCD ) was calculated by:A value of 0 indicates an identical WITS composition between two sites , while a value of 1 signifies that two samples are completely dissimilar without any overlap in WITS representation . The lithogenic diet established by Crawford et . al . ( [38] ) was modified in our experiments to include the normal rodent diet ( Harlan , Teklad 2018 ) supplied in the Stanford Research Animal Facility . Mice were fed normal base chow supplemented with 1% cholesterol and 0 . 5% cholic acid ( Harlan , Teklad custom research diet ) for 10 weeks to induce cholesterol gallstone formation . Mice on control and lithogenic diets were anesthetized with isoflurane and shaved in the abdominal area for ultrasound imaging . A Vevo 2100 system ( VisualSonics ) was used to confirm gallstone formation . Mice were given 1 week to recover prior to infection with S . Typhimurium . For sequential infections in which the IP strain served as the initial strain , mice were first injected with 103 SL1344 and the infection was allowed to establish for 35 days . Following that time period , mice were challenged with an oral dose of 108 SL1344-kanR . In experiments with sequential oral infections , mice first received 108 SL1344 orally by drinking . A persistent infection was allowed to establish for 102 days before oral challenge with 108 SL1344-kanR . This sequential oral infection was performed with the reciprocal order of strains , in which SL1344-kanR was given as the initial strain and SL1344 given as the challenge strain . Mice were infected orally with either 108 SL1344 or SL1344-kanR and fecal shedding of Salmonella was monitored over 14 days prior to the start of the experiment . A SL1344-kanR super shedder donor was then cohoused with mice previously infected with SL1344 , in addition to a naïve uninfected mouse as a control . Cohousing was continued for 10 days before the super shedder donor was removed . The reciprocal cohousing experiments were performed in which a SL1344 super shedder donor was cohoused with mice previously infected with SL1344-kanR . The aminoglycoside was administered orogastrically in a single dose of 20 mg ( Sigma Aldrich , K4000 ) dissolved in 200 µl of water . Mice were transferred to new cages with autoclaved bedding , chow ( Harlan , Teklad 2018S ) , and water at the time of administration . Prism ( GraphPad ) was used to create all figures and perform all statistical analyses . Intergroup comparisons of Bray-Curtis dissimilarity values ( e . g . spleen-cecum versus colon-cecum ) were analyzed by paired t-tests . Comparisons of oral and IP strain abundances within the same group of mice were evaluated with Wilcoxon matched-pairs signed rank tests . Differences in CFUs and strain composition between groups were examined by unpaired nonparametric Mann-Whitney tests . Significance was defined by p≤0 . 05 .
|
Salmonella enterica serovars infect various mammalian hosts , causing disease ranging from self-limiting diarrhea to persistent systemic infections such as typhoid fever . Here we investigated the impact of an established intestinal S . Typhimurium population on fecal shedding in the presence of another challenging strain . This scenario arises during host-to-host transmission , as well as during chronic host-adapted infections when systemic Salmonella reseed the intestinal tract to be transmitted in feces . In a mouse model of persistent Salmonella infection , we found that distinct subpopulations formed in intestinal and systemic tissues . Expansion of the intestinal subpopulation was responsible for increases in fecal shedding , rather than increased secretion of systemic Salmonella . Furthermore , the Salmonella that initially colonized the gut excluded challengers from the cecum , colon , and feces . A challenging systemic strain could only be shed upon ablation of the established intestinal strain . This intraspecies colonization resistance requires Salmonella hydrogenase-mediated invasion of the distal gut and is maintained by the virulence effectors SPI1 and SPI2 . We describe novel observations indicating that Salmonella virulence effectors that have been shown to subvert the host immune response and microbiota , also play a role in intraspecies competition for colonization of transmission niches .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"ecological",
"niches",
"community",
"structure",
"microbiology",
"bacterial",
"diseases",
"bacterial",
"pathogens",
"salmonella",
"typhimurium",
"animal",
"models",
"of",
"infection",
"infectious",
"diseases",
"medical",
"microbiology",
"microbial",
"pathogens",
"salmonella",
"enterica",
"salmonella",
"microbial",
"ecology",
"typhoid",
"community",
"ecology",
"community",
"assembly",
"ecology",
"biology",
"and",
"life",
"sciences"
] |
2014
|
Intraspecies Competition for Niches in the Distal Gut Dictate Transmission during Persistent Salmonella Infection
|
Prostate cancer ( PCa ) is the most commonly diagnosed malignancy and the second leading cause of cancer-related death in American men . Androgen deprivation therapy ( ADT ) has become a standard treatment strategy for advanced PCa . Although a majority of patients initially respond to ADT well , most of them will eventually develop castration-resistant PCa ( CRPC ) . Previous studies suggest that ADT-induced changes in the immune microenvironment ( mE ) in PCa might be responsible for the failures of various therapies . However , the role of the immune system in CRPC development remains unclear . To systematically understand the immunity leading to CRPC progression and predict the optimal treatment strategy in silico , we developed a 3D Hybrid Multi-scale Model ( HMSM ) , consisting of an ODE system and an agent-based model ( ABM ) , to manipulate the tumor growth in a defined immune system . Based on our analysis , we revealed that the key factors ( e . g . WNT5A , TRAIL , CSF1 , etc . ) mediated the activation of PC-Treg and PC-TAM interaction pathways , which induced the immunosuppression during CRPC progression . Our HMSM model also provided an optimal therapeutic strategy for improving the outcomes of PCa treatment .
Prostate cancer ( PCa ) is the second leading cause of cancer-related death in American men [1 , 2] . Androgen deprivation therapy ( ADT ) has become standard therapy for the treatment of PCa . Although the majority of patients initially respond well to ADT , most patients will eventually become unresponsive , and the PCas recur within 1–3 years after ADT as castration-resistant prostate cancers ( CRPC ) [3] . Previous studies have demonstrated that androgen receptor ( AR ) -mediated signaling pathway plays a central role in CRPC cell survival and growth , constituting an attractive target for therapy [4 , 5] . MDV3100 ( enzalutamide ) , an FDA-approved drug , is a well-known AR antagonist that can effectively block androgen binding to AR , thereby preventing AR nuclear translocation and coactivator recruitment [6 , 7] . However , prostate cancer treatment with AR antagonists can also acquire resistance through AR mutations [8–10] , such as AR splice variants [11] and gene amplification . Therefore , there is an urgent need for the development of new therapeutic strategies . To date , immunotherapy represents an appealing option in prostate cancer treatment [12] . The FDA-approved vaccine ( sipeucel-T [13] ) and PD-1 inhibitor ( e . g . , pembrolizumab [14] ) have been used to treat advanced PCa in clinical trials . However , recent phase III trials showed multiple failures of immunotherapy in PCa [15–18] . Recent observations suggest that the microenvironment ( mE ) of PCa is immunosuppressive , which appears to be responsible for the failures of various agents targeting the immune system in PCa [15 , 19] . Escamila et al . found that tumor-associated macrophages ( TAM ) exert a negative impact on the treatment response of PCa after ADT [20] . ADT induced an increased expression of colony-stimulating factor 1 ( CSF1 ) in prostate cancer cells ( PCs ) , leading to a significant enhancement of TAM infiltration . The increased levels of IL10 , VEGF , and EGF in TAMs , in turn , promote treatment resistance by enhancing immune suppression and tumor proliferation . Moreover , in the prostate-specific Pten-/- mouse model , Akins and colleagues found that Treg ( Regulatory T cell ) expansion occurred following ADT and the frequency and function of CD8+ T cells ( CTLs ) increased at the early stage but reduced after a late time point [21 , 22] . In this study , we found that WNT5A activated AKT/AR signaling pathway and stimulated the PC proliferation , which might be a new mechanism of PC resistance to ADT . However , the precise cellular targets and the exact molecular mechanism of the immunity leading to CRPC remain unclear . Therefore , systematic understanding of the impact of androgen deprivation on the tumor-associated immune system will help to characterize novel cytokine networks and signal transduction pathways and develop more effective combined therapies for patients with advanced PCa . Taking above studies together , we hypothesize that 1 ) following ADT , the reactivation of AR signaling in PC cells and the altered immune mE contribute to the development of CRPC; 2 ) the communication between immune cells and PCs results in immune suppression and PCa progression; 3 ) targeting the immune-PC pathways mediated by cytokines after ADT may prevent CRPC development . In recent years , some mathematical approaches have been developed to model the tumor growth , angiogenesis , and drug resistance , providing a new perspective way in exploring the molecular mechanisms of cancer treatment resistance [23–26] . Peng et al . developed an ODE-based model to characterize the effect of castration on the immune system and to predict the efficacy of combined therapy with ADT and vaccines on PCs [27] . However , the role of the immune system in CRPC progression was rarely studied . Based on the hypothesis described above , we developed a predictive 3D Hybrid Multi-scale Model ( HMSM ) with various types of data for systematically understanding the immunity leading to CRPC progression . The HMSM model consists of a 3D agent-based model ( ABM ) and an ordinary differential equations ( ODEs ) model . The ABM is used for modeling tumor growth , angiogenesis , immune response in the prostate and lymph node compartments , and the ODEs model for dynamics of intracellular signal transduction . The HMSM model integrates key biological events spatially and temporally . Spatially , the simulated mE contains two components: prostate tumor space and lymph node . PCs and TAMs reside in the tumor space for tumor growth and angiogenesis , and CTLs and Tregs home in the lymph node and infiltrate to tumor bed once the initial immune response is activated . Temporally , we modeled the intracellular signaling dynamics ( minutes to hours ) ; cell division , apoptosis , migration , and immune infiltration ( hours to days ) ; drug response ( days to weeks ) , and CRPC progression and tumor growth ( weeks to months ) . After parameter tuning , the outcomes of our HMSM model in different conditions are fit with the experimental observations . Finally , we use this model to predict the effect of individual and combined treatments with WNT5A neutralization , CSF1R inhibition [20] , IL-2 neutralization [22] , and EGFR inhibition [28] on the development of CRPC . Our simulation indicates that suppression of Treg expansion with IL-2 antibody and blockade of PC-Treg and PC-TAM interactions appear to re-activate anti-tumor immune responses and to prevent CRPC occurrence . In summary , this study revealed the key cytokines/pathways-induced immunosuppression during CRPC progression and also provided an optimal therapeutic strategy for improving the outcomes of CRPC treatments .
To model CRPC progression , we first identified the cell-cell interactions between PC and Treg based on the transcriptomic data . We calculated 1 ) the significantly overexpressed ligands- and receptors-encoded genes from the GEO ( Gene Expression Omnibus ) dataset GSE38043 [29] and GSE46218 [30]; 2 ) the directionality of cell-cell communication of ligand-receptor pairs based on the prior information in public databases , such as iRefWeb [31] . The interactions between PC and Treg were mainly inferred from above two GEO datasets using the approach reported previously [31] . The dataset GSE38043 was generated from isolated Treg cells of CRPC patients ( 3 patients VS . 3 control ) . Student T-test was used to filter the significantly overexpressed genes with a p-value < 0 . 05 . In total , we filtered 18 ligand genes ( e . g . , WNT5A ) and 26 receptor genes ( e . g . , DCR2 , EGFR , etc . ) . The dataset GSE46218 was generated from prostate orthotopic xenograft models . We compared the gene expression profiles of castration-resistant prostate cancer and androgen-dependent prostate cancer , and obtained 23 overexpressed ligand-encoded genes and 39 overexpressed receptor-encoded genes from the castration-resistant PCa , such as FZD5 , BMP6 , TNFSF10 ( TRAIL ) , etc . The calculation procedure was shown in S1 Fig . We did further filtration analysis for the identified ligand- and receptor-encoded gene pairs and found potential pairwise interactions between PC and Treg: Treg→WNT5A→PC , and PC→TRAIL→Treg ( S2 Fig ) . All of the significantly overexpressed ligand- and receptor-encoding genes were listed in S1 and S2 Tables . To determine the cell-cell interaction inferred above , we treated castration-resistant prostate cancer cells 22RV1 with WNT5A and generated RNA-seq data . Our analysis shows that WNT5A treatment up-regulates a group of genes in 22RV1 cells , e . g . , AR , FZD5 , SKP2 , PKC , ERK , STAT3 , and TNFSF10 ( TRAIL ) , etc . ( S3 Fig ) . Further functional analysis of the significantly expressed genes shows that some important pathways are enriched , including PI3K/AKT/AR pathway , Ras pathway , MAPK pathway , JAK/STAT pathway , prostate cancer pathway , and WNT pathway , etc ( see the details in S3 Table ) . Thus , WNT5A appears to be a key factor in the activation of the survival and proliferation pathways in the castration-resistant PC cells . To further validate the results obtained from RNA-seq analysis and inferred WNT5A/TRAIL pathway loop , we treated 22RV1 cells with WNT5A and the gene expression and protein levels were measured using qRT-PCR and/or Western blot , including FZD5 , TNFSF10 ( TRAIL ) , BMP6 , AR , BMP6 , Skp2 , Foxo1 , and ERK . WNT5A receptor , FZD5 was significantly up-regulated at 1 hour after treatment ( Fig 1A ) . In addition , WNT5A stimulation induced a sharp increase of TNFSF10 ( TRAIL ) ( Fig 1B ) , which may further promote Treg expansion [32] . In addition , WNT5A led to a significant increase in the BMP6 level at 0 . 5 , 1 and 3 hours following treatment ( Fig 1C ) . This finding is consistent with the previous studies showing that WNT5A stimulates BMP-6 expression in metastatic prostate cancer ( CaP ) in the context of bone niche; and BMP-6 in turn stimulated the proliferation of CaP cells [33] . Most important , treating cells with WNT5A resulted in a dramatical and persistent increase in the transcript level of AR ( Fig 1D ) . The protein levels of Skp2 and FOXO1 were increased at 3 , 7 and 24 hours post-treatment ( Fig 1E and 1F ) . These findings are consistent with reports that Skp2 and FOXO1 activation are associated with AR transactivation and tumorigenesis [34 , 35] . Finally , increased ERK phosphorylation was observed at 0 . 5 and 7 hours ( Fig 1G ) , also consistent with the early studies that MEK/ERK axis may promote CRPC development , leading to early relapse [36 , 37] . Taken together , our experimental results demonstrated that WNT5A induced AR signaling activation and secretion of TRAIL , which potentially promotes CRPC development . We also identified the PC-TAM interactions based on the previous findings . Escamilla and coworkers found that CSF1 was significantly induced in the prostate cancer cells by ADT , leading to a significant increase in TAM [20] . TAM expresses elevated levels of VEGF , MMP-9 , IL10 , and EGF , thereby to promote the protumorigenic phenotype ( such as angiogenesis and immune suppression ) of macrophages [20 , 38] . Tang et al . reported that Treg expansion in Pten-/- mice after castration was mediated by IL-2 [22] . In order to validate the PC-TAM interactions inferred above , we performed coculture experiments ( Materials and Methods ) . The in vitro experiments were designed to test the interactions of induced M2 macrophages with LNCaP cells ( androgen-sensitive ) or 22RV1 . The RNA-seq data from the co-culture of M2 macrophages with LNCaP or 22Rv1 cells was used to validate the PC-TAM interactions . With a defined FC value > 1 . 3 ( fold change of presence TAM to absence TAM ) , we totally obtained 11 over-expressed ligand genes ( e . g . , TNFSF10 , VEGFA ) and 6 receptor genes from the co-cultured LNCAP cells; and 13 ligand genes ( TNFSF10 , SPP1 , etc . ) and 12 receptor genes ( e . g . , EGFR ) in the co-cultured 22RV1 cells . At the presence of TAMs , we found that 1 ) LNCaP positively expressed AR signaling axis; 2 ) 22RV1 secreted CSF1 and TNFSF10 ( TRAIL ) , which potentially induced TAM recruitment and polarization , and Treg proliferation . Similarly , we obtained 27 overexpressed ligand genes ( e . g . , IL10 ) and 30 receptor genes ( e . g . , CSF1R ) from M2 macrophages co-cultured with LNCAP cells , compared with the M2 cells without co-culture . Also , 31 ligand genes ( IL10 , TNFSF10 , and VEGFA , etc . ) and 46 receptor genes ( CSF1R , TGFBR1 , etc . ) were over-expressed in M2 macrophage co-cultured with 22RV1 cells . Fig 2A shows the top-ranked overexpressed ligand and receptor genes in these three types of cells ( S1 Data ) . As described in the above section , we determined the potential directional connections with high confidence scores ( from iRefWeb ) and obtained 5 ligand/receptor pairs between TAMs and 22RV1s ( Fig 2A ) , including the positive loop PC→CSF1→TAM and TAM→EGF→PC demonstrated by other researchers [20] . Combing the above findings , Fig 2B revealed the cell-cell interaction network between TAM , Treg , and 22RV1 . All the enriched genes corresponding to Fig 2A were presented in S4 Table . Taken together , our analyses show that two potential cell-cell interaction loops appear to involve in the development of CRPC . The first loop is the secreted WNT5A from Tregs and macrophages triggers the activation of signaling pathways of cell survival and proliferation ( e . g . , WNT5A signaling , PI3K/AKT/AR and MAPK pathways , etc . ) in androgen-resistant PCa cells . TRAIL secreted from PCs promotes Treg proliferation [32] . The second loop is ADT-induced CSF1 expression in the tumor cells stimulates TAM infiltration . Increased TAM activation leads to increased secretion of EGF and VEGF , which in turn activate AR signaling and promote angiogenesis , respectively . Combining the above information of cell-cell communications , we highlighted an integral system in the immune mE of prostate cancer that may lead to CRPC development ( Fig 3 ) . To test the fit of our HMSM model to the training data , in silico simulations under several contexts were evaluated using the experimental data from our laboratory , as well as the data from previously reported studies ( S2 Data ) . Firstly , we simulated the whole process of prostate tumor growth from the initial state to 6 wks after castration . The dynamic changes of TAM population , CSF1 expression , and TAM-secreted protumorigenic cytokines ( e . g . , IL10 , and VEGF , etc . ) in the simulated mE were predicted . Fig 6C shows that TAM population are increased by 1 . 826±0 . 2 folds at day 7 and continued to increase to 2 . 891±0 . 353 folds at the day 14 after castration . Fig 6D represents the simulated expression of CSF1 from prostate cancer cells . The expression of CSF1 in PCs is significantly increased after ADT , which is close to the measured results from the subcutaneous mouse model [20] . Moreover , the predicted expressions of IL10 and VEGF in TAMs have increased 2 . 98±0 . 171 and 1 . 54±0 . 078 folds at two days post-castration ( Fig 6E ) . We also predicted the effect of the CSF1R inhibitor on the distribution of the cell population after castration . As shown in Fig 6F , inhibition of CSF1R with PLX after ADT results in a 5-fold reduction of the TAM population in the early stage , compared with the results in castration only . The CSF1R blockade appears to inhibit macrophage proliferation , and lower TAM-induced expression of VEGF , which potentially delays the emergence of CRPC . The simulated results shown in Fig 6C–6F are consistent with the experimental data reported in [20] . To examine the dynamical changes of immune responses , we calculated the changes of the Treg population after castration in the simulated system . As shown in Fig 6G , the number of Treg cells is significantly increased in lymph nodes at 2 . 5 weeks and 5 weeks post-castration compared to that from pre-castration . This simulation result is close to the previous findings reported by Tang et al . [22] . In addition , Fig 6H shows that Treg expansion is prevented by IL2 neutralization , suggesting the increased IL2 after castration and immunization promotes Treg expansion [22] . Taken together , we found that the HMSM model fits the observed data very well under different contexts . To further validate the reliability of our HMSM model optimized above , we compared the simulation results with additional experimental data ( S2 Data ) generated from Pten-/- prostate cancer mouse model [21 , 27] . Fourteen weeks-old mice were castrated , and the relative changes of immune cells ( Treg and CD8+ T cells ) in tumor space were observed at the 2 . 5 weeks and 5 weeks after castration [21] . Fig 6I and 6J show that castration induces infiltration of Treg cells into the tumor area in the prostate tumor-bearing mice . However , the accumulation of functional CD8+ T cells in the prostate tumor is not long-lasting , evident at 2 . 5 weeks after castration but reduced at 5 weeks after castration . The measured CD8+ T cells at 2 . 5 weeks and 5 weeks after castration are around 2 . 05±0 . 25 and 1 . 75±0 . 125 folds of pre-castration , respectively [27] . Fig 6I indicates that the prediction of the CD8+ population ( 2 . 107±0 . 775 , and 1 . 7606±0 . 8141 folds ) in the HMSM model is consistent with the experimental observations . Moreover , the number of Tregs was significantly increased to 1 . 7±0 . 5 and 3 . 2±0 . 6 folds in the tumors at 2 . 5 and 5 weeks after castration , respectively [27] . HMSM model simulation shows that the predicted changes of the Treg population at 2 . 5 and 5 weeks after castration , and our predicted results are close to the experimental observations ( Fig 6J ) . In summary , our testing experimental data further confirms that the outputs of HMSM model are reliable . To identify the potential therapeutic targets of PCa in the immunosuppressive prostate cancer mE , we predicted the effects of single or combined treatments with castration on PCa growth using the established HMSM model ( S2 Data ) . Our experimental data revealed WNT5A was a potential factor associated with CRPC development . Therefore , we simulated the effect of WNT5A neutralization on PC growth with our in silico model . Recent studies indicate that CSF1R inhibitor ( PLX3397 ) [20 , 66] and IL-2 neutralization [22] revealed the effects for immune re-activation after ADT . In addition , early studies have reported the efficacy of EGFR inhibitors ( e . g . erlotinib , canertinib , and cetuximab , etc . ) in castration-resistant prostate cancer in vitro and in vivo , and claimed that EGFR inhibition might improve the outcome of patients with CRPC [28 , 67 , 68] . Therefore , we mainly tested the anti-tumor effects of four representative agents in HMSM , including anti-WNT5A antibody , PLX3997 , Anti-IL-2 antibody ( IL-2 neutralization ) , and EGFR inhibitor . Fig 6K shows the predicted outcomes from single or combined treatments relative to pre-castration . Prostate tumor cells were reduced sharply at the first 2 weeks after castration and then re-expanded continuously ( silver curve ) . The combined treatment with castration and a single agent ( Anti-WNT5A , PLX , Anti-IL-2 , or EGFR inhibitor ) yields a better treatment response than that from the castration only group . Comparing with Anti-IL-2 and EGFR inhibitor , a combination of castration with Anti-WNT5A or PLX yields better anti-tumor responses , indicating that blockade of the PC-Treg or PC-TAM interaction may effectively reduce tumor cell growth . In addition , the poor response was observed in the combined treatment group with castration plus EGFR inhibitor , compared with the other combined treatment groups . The optimal prediction outcome was achieved from the treatment group with a combination of PLX , Anti-WNT5A , and Anti-IL-2 ( red curve ) after castration , revealing that the activation of both Treg and TAM appears to contribute to CRPC development . Moreover , we compared the predicted results with the experimental observations reported previously . Fig 6L shows that the tumor growth rebounded approximately 3 times after castration , paralleling the emergence of CRPC observed in the clinical setting [20] . The addition of CSF1R inhibitor PLX3997 to castration resulted in a significant delay in the onset of CRPC ( Fig 6M ) [20] . Also , Anti-EGFR leads to 0 . 77±0 . 128 and 0 . 86±0 . 157 fold tumor growth at 3 and 5 weeks after castration relative to castration only ( Fig 6N ) , which are close to the experimental observation of in vivo effects of EGFR inhibitors in 22RV1 xenografts mice model [28] . Above analyses indicate that the predictive capabilities of our HMSM model are high and the model-based predictions are reliable .
The focus of this work is to explore CRPC progression in the immune mE and to develop optimal treatment strategies in silico to improve therapeutic responses of CRPC . To systematically understand the role of the immune system in CRPC development , we generated RNA-seq data and integrated it with the GEO datasets . Through the analysis of these data , we found the potential factors/cytokines ( e . g . , WNT5A , and TRAIL ) associated with PC-immune interactions . Elevated levels of WNT5A have been reported in melanomas , lung cancer , breast cancer , and gastric cancer [69–73] . Lee , et al . investigated the Cap-bone stromal cells interaction , and reported that WNT5A secreted by bone stromal cells increases BMP-6 expression in Cap , thereby leading to Cap cell proliferation [33] . Our study demonstrated that WNT5A induced the activation of androgen-independent pathways and the elevated expression of TRAIL in CRPC cells after castration , indicating the enhancement of PC growth and immune suppression . As a type 2 membrane protein belonging to TNF superfamily , TRAIL is known to play a pivotal role in the immune regulation and antitumor immunity [74–76] . Early studies revealed that TRAIL has the potential to promote Treg proliferation in certain situations [77] . Ikeda and coworkers demonstrated that the proliferative effect of TRAIL on Tregs becomes apparent in autoimmunity [32] . The exploration of TRAIL function in prostate cancer may be of considerable significance for understanding CRPC mechanisms . We are the first to systemically model the CRPC development in the immune mE using an integrated 3D system ( S8 Fig ) . In our HMSM model , we simulated the PC growth before and after castration . The first stage covers a sequence of key biological events , including DC maturation , T cell activation , and division in lymph nodes triggered by DC , T cell migration and infiltration . The second stage denotes the initial castration therapy ( 5 weeks ) , in which the AR signaling reactivation appears around 2 weeks after castration ( S4 Fig ) . Therefore , the proposed model provides a new way to present the dynamic changes in tumor growth , immune response , and drug treatment effect . We also provide a novel computational platform to optimize the potential target therapy on the castrated PCs . ADT is a standard treatment for PC patients , including surgical castration , and AR disruption with pharmacologic interventions ( such as MDV3100 ( enzalutamide ) [78] ) . However , clinical studies indicate that AR antagonist can induce AR T878A mutation and result in AR reactivation [79–81] . Our analysis of a representative GEO dataset ( GSE67980 ) [82] also revealed that AR expression was increased when the patients with CRPC treated by enzalutamide ( S9 Fig ) . In recent years , active immunotherapy , such as therapeutic vaccines , provide new strategies for overcoming tumor-mediated immune suppression [83] . Multifaceted approaches that combine vaccine with targeted therapies may have the potential to improve the current therapeutic outcomes by targeting the suppressive immune microenvironment and tumor survival . In the present study , we evaluated several new therapeutic strategies in silico with our optimized HMSM model . The simulated results showed that the optimal prediction outcome was achieved from the treatment group with a combination of PLX , Anti-WNT5A , and Anti-IL-2 after castration , revealing the important role of Treg and TAM activation . Moreover , the proposed model includes a large number of parameters , and most of the parameters were tuned manually or determined based on the experimental results . In order to confirm the variability of the simulated results from the developed 3D hybrid multi-scale model , a parameter sensitivity analysis was performed by measuring the impact of a small perturbation ( 5% increase ) of individual 34 key parameters on the prostate tumor cell populations ( 5wk after castration ) . We found that 1st and 2nd parameters ( the basic proliferation rates caused by castration-dependent and castration-independent pathways in PCs ) were more sensitive than others ( S10 Fig ) . It indicates that ADT induced prostate cancer cells to progress and further express cytokines to promote CRPC occurence . The sensitivity analysis showed the changes in model outcomes were under 4% , indicating that the outcomes of the optimized model were stable . We also tested the effect of initial cell numbers and cut-off values in the ABM rules on the model variability . S11 Fig and S12 Fig show that simulated tumor growth is not sensitive to the perturbations on the initialization of cell number and the cut-off values in the ABM rules . We did additional analysis with the experimental time points overlaid as dots at the observed times , our results indicate that the optimized HMSM model is reliable ( S13 Fig ) . Although a number of mathematical approaches have been introduced to model the tumor growth and drug resistance in recent years , most of the well-defined 3D agent-based models not only neglect the stage-structured immune response during the tumor initialization and development , but also did not simulate the dynamics of intracellular pathways in the cell-cell communications [25 , 65] . Solovyev et al . was the first to put forward the concept “hybrid model” , which combined ODE model and agent-based model to mimic signal transduction processes at the intracellular scale , stochastic cell behaviors at the intercellular scale , and the dynamic distribution of growth factors at the tissue scale [84] . However , their model was only designed for two-dimensional space so that it cannot be used in 3D tumor study . Our 3D Hybrid model ( HMSM ) overcomes the limitations of existing models described above , and creates a new paradigm for systematically understanding the immunity leading to CRPC . There are several limitations of our HMSM model . We used some experimental data from in vitro 2D culture to model 3D microenvironment in this study . Ideally , experimental data obtained directly from 3D tissues can better reflect actual environmental status . However , such types of data are not easily available due to animal study settings . Moreover , using limited available animal data for ABM model training , validation , and prediction may not be enough for the validation of our large-scale-based ABM model . Incorporating much more observed data will increase the reliability of the model outcome . In the future , we will collect tumor tissue data from patients with PCa before and after castration to verify our 3D model . We will develop heterogeneity scoring approaches to evaluate cell-level heterogeneity ( receptor expression ) and tumor-level heterogeneity ( cytokine levels , and geometry ) . We will extend our model to simulate the effect of new blood vessels on the tumor growth , e . g . modeling increased cancer cell migration and invasion . To better address clinically relevant issues , we will further improve our model in terms of varying-degree inhibition with inhibitors , enabling it to predict dose-related treatment outcomes .
Androgen Receptor ( AR ) not only can be activated by extracellular growth factor through androgen-independent pathway network , but also by DHT from the tumor microenvironment through androgen dependent pathway . The activation or inhibition of these two ways depends on the stages of CRPC progression . In our study , we developed an ODE model to simulate the effect of WNT5A/EGF-triggered androgen-independent pathway on the proliferation of prostate cancer cells ( Fig 5D ) . The extracellular concentration of WNT5A and EGF are the input variables for the ODEs . In the training of ODEs , the input parameters ( WNT5A or EGF ) were restricted to the range [0 , 1] , and the maximal value “1” represents the maximal dose of ligands we used in our experiment . The output is the fold change of proliferation rates relative to no stimulation . The ODE system has the following form: d[ERK]dt=k1[WNT5A]H1+[WNT5A]−d1[ERK] ( 1 ) d[Skp2]dt=k2[EGF]H2+[EGF]−d2[Skp2] ( 2 ) d[AKT]dt=k3[Skp2]H3+[Skp2]−d3[AKT] ( 3 ) d[AR]dt= ( 1−D1 ) *k4[ERK]H4+[ERK]+ ( 1−D2 ) *k5[AKT]H5+[AKT]−d4[AR] ( 4 ) d[Prol]dt=k6[AR]H6+[AR]−d5[prol] ( 5 ) As mentioned above , our phosphor-proteomics data covered the key signaling proteins ( pERK , pAKT , AR , and Skp2 ) , which were involved in this androgen-independent signaling network ( Fig 5C ) . The effect of WNT5A and EGF on 22RV1 cell proliferation were also presented in Fig 5A and 5B . All above parameters involved in this ODE system were estimated by optimizing formula ( 6 ) via the GA algorithm [27]: θ*=argmin∑iϵI1 , tϵT1|Xit−X^it ( θ ) | ( 6 ) Where Xit and X^it ( θ ) denote the measurement from the experiments and the theoretical results obtained from the ODE model of protein i at the time point t . The parameter vector θ = {k1 , H1 , d1 , …… , k6 , H5 , d5} in above formulas ( 1–5 ) can be obtained by formula ( 6 ) . D1 and D2 represent the inhibitors of WNT5A and EGF pathways , respectively . The set I1 is the indexes of observed proteins in this signaling network , and time series set T1 = {0 , 30min , 60min , 420min} covers all the time points related with experimental data ( Fig 5C ) . S5 Table represents the estimated values of all parameters . The fitting accuracy of the predicted and measured values of key proteins is shown in Fig 5 . We defined five types of agents in the ABM model to represent PC , TAM , CTL , Treg , and EC , respectively ( Fig 3 ) . The ABM model simulates the effects of various cell-cell interactions on prostate tumor growth , angiogenesis , programmed immune response , and drug response in a simulated mE . We initialized the simulated microenvironment as a cuboid , which consists of two connected cubes . One is for the growth and proliferation of mixed PC , TAM , CTL , Treg , and EC compartments ( tumor space ) and the other is for the activation and division of T cells triggered by matured DC in lymph node and T cell infiltration from its lymph vessels . The proposed model simulated a series of key biological events involved in tumor growth , immune response , and CRPC development ( S4 Fig ) . The details of time points related with these events were described in the S1 Text . This multi-scale modeling includes intracellular , intercellular and tissue scales , which are illustrated in the S8 Fig , and described into details in the following sections . Detailed flowcharts of each agent were illustrated in the S1 Text . Individual cell behaviors were simulated by probability-based rule implementation [52 , 65] . A cell senses the hints in its neighborhood such as local cytokines and drugs and adjusts itself with the embedded signaling pathways , and outputs the corresponding changes on its cell behaviors , including proliferation , survival , differentiation , migration , and cytokine secretion rate . Cell fate decision is then determined by rolling a dice and compared with the probability threshold of cell behavior ( S14 Fig ) . In this study , the proliferation rate of PCs was determined by two pathways . Our hypothesis is that: the androgen concentration in blood and gland will be sharply decreased after castration , so that prostate tumor cell proliferation will be supported by WNT5A or EGF-mediated androgen-independent pathway until the occurrence of CRPC ( AR-reactivation ) . The ODE system for cell proliferation of PCs has been described in the above section . Except the ODE system was applied to model the intracellular signaling network in PC cells , Hill functions were used to simulate the signal transduction of other cells to calculate apoptosis and proliferation rates , and further determine the cell behaviors . In response to the changes of WNT5A , EGF or DHT in its local mE , each prostate cancer cell will proliferate , migrate , become quiescent , or undergo death process . PCs secrete CSF1 to promote macrophage infiltration . Macrophage-derived EGF enhances tumor cell invasion [38] . TAMs also suppress the immune response of T cells by releasing the immunosuppressive factor , such as IL10 [88] . Similarly , the WNT5A-TRAIL positive loop between PCs and Tregs and the associated molecules are also considered as important modulating components for CRPC development and immune suppression . In addition , prostate tumor cells can be killed by CD8+ T cells . Treg cells can migrate towards CD8+ T cells locally and suppress the proliferation of these cells in a manner of cell cycle arrest or apoptosis [89] . The framework of the ABM model was designed using the conception of “Object-Oriented Programming” and achieved with C++ . The ODE system of intracellular signaling pathways in PC was established with C and solved by the Fortran ODE Solver ( DLSODE [103] ) , and called in the ABM model of HMSM ( S1 . 10 Text ) . The ABM model was debugged and implemented under Linux environment on the cluster platform of Demon in Wake Forest Baptist Medical Center and Texas Advance Computing Center ( TACC ) . All of the parameters in the ABM model were tuned by running the system 100 times for each candidate solution . The model with optimal parameters should fit the training data well . For addressing the stochastic results from the ABM , we evaluated the model outcomes after replicating simulations ( repeat 100 times ) on a fixed model . Both average and standard deviation were used to present the results .
|
Prostate cancer ( PCa ) is the second leading cause of cancer-related deaths in American men . Androgen deprivation therapy ( ADT ) is the first-line therapy for advanced PCa , yet a significant number of primary PCa patients treated with ADT eventually develop incurable castration-resistant prostate cancer ( CRPC ) . Recent observations suggest that the immunosuppressive microenvironment of PCa might be responsible for the failures of various therapies . However , the role of immune system in CRPC progression is still unclear . To deeply understand the immunity leading to CRPC progression , we developed a unique systems biology approach ( HMSM ) . Based on our analysis , we identified the key molecules ( e . g . WNT5A , TRAIL , CSF1 , etc . ) mediating the communication of PCa and immune cells . Our HMSM system also revealed the optimal therapeutic strategy for PCa treatment . Collectively , our study provides a new insight to study tumor-related immune mechanisms and pave the way for the development of more effective treatments .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blood",
"cells",
"urology",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"cardiovascular",
"physiology",
"cancer",
"treatment",
"immunology",
"cancers",
"and",
"neoplasms",
"genitourinary",
"tract",
"tumors",
"basic",
"cancer",
"research",
"surgical",
"and",
"invasive",
"medical",
"procedures",
"castration",
"oncology",
"angiogenesis",
"developmental",
"biology",
"lymph",
"nodes",
"prostate",
"cancer",
"tumor",
"physiology",
"lymphatic",
"system",
"tumor",
"angiogenesis",
"reproductive",
"system",
"procedures",
"white",
"blood",
"cells",
"animal",
"cells",
"t",
"cells",
"exocrine",
"glands",
"prostate",
"diseases",
"cell",
"biology",
"anatomy",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"regulatory",
"t",
"cells",
"prostate",
"gland"
] |
2019
|
Systematically understanding the immunity leading to CRPC progression
|
It has been suggested that genetic susceptibility plays an important role in the pathogenesis of diabetic nephropathy . A large-scale genotyping analysis of gene-based single nucleotide polymorphisms ( SNPs ) in Japanese patients with type 2 diabetes identified the gene encoding acetyl-coenzyme A carboxylase beta ( ACACB ) as a candidate for a susceptibility to diabetic nephropathy; the landmark SNP was found in the intron 18 of ACACB ( rs2268388: intron 18 +4139 C > T , p = 1 . 4×10−6 , odds ratio = 1 . 61 , 95% confidence interval [CI]: 1 . 33–1 . 96 ) . The association of this SNP with diabetic nephropathy was examined in 9 independent studies ( 4 from Japan including the original study , one Singaporean , one Korean , and two European ) with type 2 diabetes . One case-control study involving European patients with type 1 diabetes was included . The frequency of the T allele for SNP rs2268388 was consistently higher among patients with type 2 diabetes and proteinuria . A meta-analysis revealed that rs2268388 was significantly associated with proteinuria in Japanese patients with type 2 diabetes ( p = 5 . 35×10−8 , odds ratio = 1 . 61 , 95% Cl: 1 . 35–1 . 91 ) . Rs2268388 was also associated with type 2 diabetes–associated end-stage renal disease ( ESRD ) in European Americans ( p = 6×10−4 , odds ratio = 1 . 61 , 95% Cl: 1 . 22–2 . 13 ) . Significant association was not detected between this SNP and nephropathy in those with type 1 diabetes . A subsequent in vitro functional analysis revealed that a 29-bp DNA fragment , including rs2268388 , had significant enhancer activity in cultured human renal proximal tubular epithelial cells . Fragments corresponding to the disease susceptibility allele ( T ) had higher enhancer activity than those of the major allele . These results suggest that ACACB is a strong candidate for conferring susceptibility for proteinuria in patients with type 2 diabetes .
Diabetic nephropathy is a leading cause of end-stage renal disease ( ESRD ) in Western countries [1] and in Japan [2] . The rising incidence of diabetic nephropathy , especially among patients with type 2 diabetes , is a serious worldwide concern in terms of both poor prognosis and medical costs . The pathogenesis of diabetic nephropathy has not been fully elucidated . However , susceptibility to diabetic nephropathy appears to be determined by multiple genetic and environmental risk factors , and genetic susceptibility plays an important role in its development and progression [3] , [4] . Both candidate gene approaches and genome-wide linkage analyses have suggested several candidate genes with potential impact on diabetic nephropathy . However , these findings have not been robustly replicated [5] , [6] , and many susceptibility genes for diabetic nephropathy remain to be identified . The recent development of single nucleotide polymorphism ( SNP ) typing technology and insights into patterns of linkage disequilibrium ( LD ) in the human genome have facilitated genome-wide association studies ( GWASs ) for investigating genes associated with disease susceptibility across the entire human genome . GWASs conducted by several independent research groups in Europe , United States [7] , [8] and Japan [9] , [10] have identified multiple loci associated with susceptibility to common complex traits , including type 2 diabetes . Recently conducted GWAS in a population of European descent identified 4 distinct loci associated with diabetic nephropathy in type 1 diabetes . Two of these loci were replicated in a population of the Diabetes Control and Complications Trial ( DCCT ) /Epidemiology of Diabetes Interventions and Complications ( EDIC ) cohorts [11] . With the aim of identifying loci involved in susceptibility to common diseases , we initiated a large-scale association study using SNPs from a Japanese SNP database ( JSNP: http://snp . ims . u-tokyo . ac . jp/ ) [12] , [13] , that was established before creation of the HapMap database . Through this project , we have previously identified genes encoding solute carrier family 12 ( sodium/chloride ) member 3 ( SLC12A3: MIM 600968 , Online Mendelian Inheritance in Man: http://www . ncbi . nlm . nih . gov/omim ) [14] , engulfment and cell motility 1 ( ELMO1: MIM 606420 ) [15] , and neurocalcin δ ( NCALD: MIM 606722 ) [16] as being associated with susceptibility to diabetic nephropathy . The ELMO1 association has been replicated in African Americans [17] and European Americans [18] . In the present study , we extended a previous large-scale association study for diabetic nephropathy , and provide evidence that a SNP within the acetyl-coenzyme A ( CoA ) carboxylase beta gene ( ACACB; MIM: 601557 ) contributes to an increased prevalence of proteinuria in patients with type 2 diabetes across different ethnic populations .
We extended our prior analysis to SNPs with p values between 0 . 01 and 0 . 05 , and examined the association of these SNPs with diabetic nephropathy in a larger study sample . In this analysis , a SNP within ACACB showed the strongest association with diabetic nephropathy in Japanese patients with type 2 diabetes ( rs2268388: intron 18 +4139 C > T , p = 1 . 4×10−6 , odds ratio [OR] = 1 . 61 , 95% confidence interval [Cl]: 1 . 33–1 . 96 , Table 1 ) . Subsequent LD mapping around this region with data for 264 SNPs with allele frequencies ≥0 . 1 from HapMap database ( HapMap: http://hapmap . ncbi . nlm . nih . gov/ ) for the Japanese , identified a 20-kb LD block that included an original marker SNP ( rs2268388 ) , which corresponded to a part of the ACACB gene ( Figure 1A and 1B ) . Therefore , we concluded that ACACB was likely a candidate for conferring susceptibility to diabetic nephropathy . We next analyzed 51 SNPs , including 31 tagging SNPs , within ACACB in our Japanese population ( Japanese1 ) . Several SNPs within the same LD block as rs2268388 were nominally associated with diabetic nephropathy ( Figure 1C , Table S1 ) . No single SNP or haplotype showed stronger association with diabetic nephropathy than the original marker SNP ( Figure S1 ) . To validate the association of this SNP with diabetic nephropathy , we examined the effects of the SNP on susceptibility to the disease in several independent populations from different ethnic groups ( Table 2 ) . The results indicated that the frequency of the T allele of rs2268388 was consistently higher among patients with type 2 diabetes with proteinuria ( combined meta-analysis gave a p value of 5 . 35×10−8 in the Japanese , 2 . 34×10−7 for all populations ) . Significant association with ESRD was detected in the relatively large European 2 samples ( 481 cases and 427 controls ) . The SNP was also modestly associated with ESRD in East Asian type 2 diabetes , but the direction of association differed . Overall , the distribution of the genotype for rs2268388 did not differ significantly between patients with ESRD and control patients having type 2 diabetes ( p = 0 . 47 ) . No significant association was detected in patients with type 1 diabetes having proteinuria . We next examined the expression profile of ACACB in various human tissues . Expression of ACACB was observed in adipose tissue , heart and skeletal muscle , and , to a lesser extent , in the kidney ( Figure 2A ) . The results of in situ hybridization with normal mouse kidney revealed that Acacb was localized to glomerular epithelial cells and tubular epithelial cells ( Figure 2B ) . We also observed the expression of ACACB in cultured human renal proximal tubular epithelial cells ( hRPTECs ) . To investigate the functional role of this SNP region , we examined the effects of a 29-bp DNA fragment containing the associated SNP ( rs2268388 ) on transcriptional activity in cultured hRPTECs . As shown in the Figure 3 , the 29-bp DNA fragments had significant enhancer activity ( promoter alone [P]: 39 . 4±13 . 1; susceptibility allele [T]: 384 . 3±104 . 1; major allele [C]: 238 . 5±81 . 9; relative luciferase activity , p = 0 . 0005 for P vs . T , p = 0 . 016 for P vs . C ) . Fragments corresponding to the disease susceptibility allele had stronger enhancer activity than those for the major allele ( [T] 10 . 5±3 . 4 vs . [C] 5 . 9±1 . 3 , fold increase over promoter alone , p = 0 . 045 , Figure 3B ) .
In the present study , we showed ACACB located at chromosome 12q24 . 1 to be a strong susceptibility gene for diabetic nephropathy in patients with type 2 diabetes . Our findings suggest that a SNP within ACACB ( rs2268388 , intron 18 + 4139 C > T ) contributes to the development of proteinuria in patients with type 2 diabetes . ACACB encodes acetyl-coenzyme A ( CoA ) carboxylase beta , which catalyzes the carboxylation of acetyl-CoA to malonyl-CoA , and controls fatty acid oxidation by means of the ability of malonyl-CoA to inhibit carnitine palmitoyl transferase I ( CPT1A; MIM 600528 ) , the rate-limiting step in fatty acid uptake and oxidation by mitochondria in non-lipogenic tissues . Mice lacking Acacb have a normal life span , a higher rate of fatty acid oxidation , lower amounts of fat , and increased insulin sensitivity [19]–[21]; therefore , ACACB might affect insulin sensitivity via modulation of fatty acid metabolism . However evidence suggesting a role for the ACACB in the pathogenesis of diabetic nephropathy was previously lacking . In this study , expression of ACACB was detected in heart , skeletal muscle and adipose tissues by real-time quantitative polymerase chain reaction ( PCR ) as previously reported [22] . We also showed that ACACB was expressed in human kidney , and in situ hybridization revealed that Acacb expression was localized to glomerular epithelial cells and tubular epithelial cells in normal mouse kidneys . Abnormalities in lipid metabolism [23] , [24] , including fatty acid metabolism have been shown to contribute to the development and/or progression of chronic kidney diseases , including diabetic nephropathy . Hence , genotype-based differences in expression and/or activity of this enzyme in the kidney might contribute to conferring susceptibility to diabetic nephropathy . Elucidating these functional differences will help us to understand how variation in this gene contributes to susceptibility to diabetic nephropathy . In this study , the 29-bp fragment that included the landmark SNP ( rs2268388 ) was shown to have significant enhancer activity . We also demonstrated that the DNA corresponding to the disease susceptibility allele had significantly higher enhancer activity than that for the major allele in cultured human RPTECs . Therefore , the intronic variation in the gene seems to be causal . We hypothesize that higher expression of ACACB in the kidneys of subjects having the disease susceptibility allele ( T ) may increase the susceptibility to diabetic nephropathy in type 2 diabetes . Interestingly , the association of the T allele of rs2268388 with type 2 diabetes-associated nephropathy was consistently observed in cases with proteinuria , whereas there was less consistent association between type 2 diabetic patients under chronic renal replacement therapy ( ESRD ) . Discrepancies in genomic loci underlying susceptibility to proteinuria versus ESRD were previously noted in a genome-wide linkage scan for diabetic nephropathy in type 2 diabetes [25] . Because most of the patients with ESRD were considered to have had proteinuria , and there is significant heterogeneity in the association with diabetic ESRD among the East Asian and European American populations ( heterogeneity p = 0 . 0006 , Table 2 ) , some selection bias , such as a survival effect , or ethnic differences might exist when patients with ESRD were used as cases . Since the presence of proteinuria is also recognized as a predictor of cardio-vascular diseases , the association of ACACB with proteinuria might reflect an association between the gene and cardio-vascular diseases or metabolic syndrome . However elucidation of a precise mechanism will require further investigation . Association of the ACACB with diabetic nephropathy could not be replicated in patients with type 1 diabetes , although these nephropathy cases had proteinuria . The clinical features or histological characteristics of diabetic nephropathy are similar in both type1 and type 2 diabetes mellitus; however , there are some differences in the background circumstances between both types of the disease . For example , patients with type 2 diabetes are generally older , and more obese than those with type 1 diabetes . Therefore , it is possible that genetic factors for nephropathy are different for type 1 and type 2 diabetes , although some overlap may exist . Because the statistical power of this study for patients with type 1 diabetes was probably not sufficient , the association of variations in ACACB with diabetic nephropathy in patients with type 1 diabetes should be re-evaluated in future studies . Recently , GWASs have been performed to identify susceptibility genes for common diseases . Convincing susceptibility genes for many diseases including type 2 diabetes have been successfully identified by GWASs . A GWAS for diabetic nephropathy was conducted in the Genetics of Kidneys in Diabetes ( GoKinD ) collection , and several novel loci were identified for diabetic nephropathy in type 1 diabetes [11] . Compared to our study , recent GWASs conducted in European and American groups had greater power to detect true associations . The power of the first and second test in the present study is estimated >40% and >90% respectively , for SNPs with minor allele frequency of 0 . 2 as in our sample , if we set a cut-off value at the p = 0 . 05 level , a genotypic relative risk ( γ ) of 1 . 5 , and the prevalence of diabetic nephropathy is assumed to be 10% ( CaTS power calculator , CaTS: http://www . sph . umich . edu/csg/abecasis/CaTS/ ) . Therefore , other important loci contributing to diabetic nephropathy in Japanese populations may not have been detected and should be searched for using a larger scale of GWAS . Limitations exist in increasing the numbers of subjects with diabetic nephropathy and diabetes lacking nephropathy in single center studies . Therefore , we examined several independently collected study populations , allowing for the presence of different selection criteria and the small sample size in some replication cohorts . Future analyses should attempt to avoid these limitations so as not to produce spurious results . In summary , based upon extensive genome-wide gene-centric SNP analyses , we identified ACACB as a candidate gene for conferring susceptibility to diabetic nephropathy . Subsequent replication studies in several different ethnic groups and a functional study suggest that the T-allele of a common intronic SNP ( rs2268338 ) within ACACB is a risk factor for the development and progression of proteinuria in patients with type 2 diabetes .
Japanese 1 ( genome-wide screening ) ; DNA samples were obtained from the peripheral blood of patients with type 2 diabetes who regularly attended outpatient clinics at Shiga University of Medical Science , Tokyo Women's Medical University , Juntendo University , Kawasaki Medical School , Iwate Medical University , Toride Kyodo Hospital , Kawai Clinic , Osaka City General Hospital , Chiba Tokusyukai Hospital or Osaka Rosai Hospital . All subjects provided informed consent before enrolling in this study . DNA extraction was performed by a standard phenol-chloroform method . Diabetic patients were divided into 2 groups according to the following diagnostic criteria: 1 ) nephropathy cases , i . e . , patients with diabetic retinopathy and overt nephropathy , indicated by a urinary albumin excretion rate ( AER ) ≥200 µg/min or a urinary albumin/creatinine ratio ( ACR ) ≥300 mg/g creatinine ( Cr ) , and 2 ) control patients who have diabetic retinopathy but no evidence of renal dysfunction ( i . e . AER less than 20 µg/min or ACR less than 30 mg/g Cr ) . Measurements of AER or ACR were performed at least twice for each patient . The SNPs for genotyping were randomly selected from our gene-based Japanese SNP ( JSNP ) database . The genotype of each SNP locus was analyzed with multiplex PCR-invader assays , as described previously [14]–[16] . Our first screening involved genotyping 94 nephropathy cases and 94 control patients for more than 100 , 000 SNP loci . In total , 76 , 767 SNP loci , which distributed to 13 , 707 gene-centric regions and covered approximately 35% of common SNPs ( MAF >0 . 15 ) in these regions , were successfully genotyped by the invader assay ( success rates >0 . 95 ) . We estimated identity by descent ( IBD ) sharing to assess relatedness among our initial GWAS populations . As shown in Figure S2 , the result indicates that there were no close relative pairs in this population . We also performed principal component analysis ( PCA ) using our initial GWAS populations to evaluate population structures . Since all subjects were in a single cluster in the PCA analysis ( Figure S3 ) , no evidence for population stratification between case and control groups appeared to exist in the present GWAS . After the first round of analysis that evaluated SNPs with p values less than 0 . 01 in the first screening ( previously published [14]–[16] ) , we extended our analysis to include SNPs with p values between 0 . 01 and 0 . 05 and examined them in a larger number of patients ( 754 nephropathy cases vs . 558 control patients ) . The study protocol was approved by the ethics committees of RIKEN Yokohama Institute and each participating institution . We obtained human cDNAs from multiple tissues from CLONTECH Inc . ( Palo Alto , CA , U . S . A . ) . The cDNAs were amplified by PCR with the following primers: human ACACB , sense 5′-CGG ATG CGT AAC TTC GAT CTG-3′ , antisense 5′-CTA TGG TCC GTC ACT TCC ACA C-3′; BACT , sense 5′- TCA CCC ACA CTG TGC CCA TCT ACG A -3′ , antisense 5′- CAG CGG AAC CGC TCA TTG CCA ATG G -3′ . Amplification was performed in a 22 µl reaction volume that contained 1× EX Taq Buffer , 200 nM dNTP , 1/20 , 000 SYBR Green , 0 . 2 µM Rox , 800 nM gene-specific primer , 0 . 05 U/µl EX Taq Hot Start Version ( Takara , Otsu , Japan ) , and 5 ng of template DNA . The thermal profile was 50°C for 2 min , at 95°C for 10 min , followed by 40 cycles at 95°C for 30 s and at 60°C for 60 s in thermal cycler ( Mx3000P Multiplex Quantitative PCR system; Stratagene , La Jolla , CA , U . S . A . ) . The results were normalized with human BACT . Under pentobarbital anesthesia , 20-week-old mice were flushed with PBS through the abdominal aorta followed by perfusion with 4% paraformaldehyde buffered with PBS ( pH 7 . 4 ) . The kidneys were quickly removed and cut into small pieces . The renal cortex tissue was immediately dissected and immersed into a fresh portion of the same fixative at 4°C overnight . All steps were carefully carried out to avoid contamination with RNase . Diethylpyrocarbonate-treated water was used at 0 . 1% to prepare each buffer . The fixed samples were thoroughly rinsed with PBS ( pH 7 . 4 ) and subsequently dehydrated by passage through an alcohol series and cleared in xylene . In situ hybridization was performed on paraffin-embedded sections using a previously described method [15] . Antisense and sense single-strand cRNAs were synthesized from cDNA fragments encoding Acacb using reverse-transcription PCR . The Acacb cDNA fragment was consisted of a 500 bp mouse sequence ( nucleotides 181–680 , GenBank accession number NM_133904 , GenBank: http://www . ncbi . nlm . nih . gov/Genbank/ ) . Three copies of the 29-bp DNA fragments including rs2268388 in ACACB were subcloned into a pGL3-promoter vector ( Promega , Madison , WI , U . S . A . ) at its multi-cloning site upstream of the SV-40 promoter . We introduced constructs corresponding to each allele into the human renal proximal tubular epithelial cells ( hRPTEC , Lonza , Basel , Switzerland ) along with a sea-pansy luciferase control vector , pRL-TK ( Promega ) , using the liposome transfection procedure ( Lipofectoamine 2000 , Life Technology Inc , Carlsbad , CA , U . S . A . ) . Twenty-four hours after transfection , luciferase activitiy was determined by means of the Dual Luciferase Reporter Assay System ( Promega ) . The luminescence of firefly luciferase was corrected by use of the sea-pansy luciferase , which reflected transfection efficiency . We tested the genotype and allele frequencies for Hardy-Weinberg equilibrium ( HWE ) proportions by use of the χ2 test [32] . We calculated the LD index , D' and r2 , as described elsewhere [33] . We analyzed the differences between the case and control groups with regard to the genotype distribution and allele frequency in the genome-wide screen by Fisher's exact test with dominant , recessive and allelic models with autosomal SNPs . The association of the ACACB locus with diabetic nephropathy in the replication study was evaluated with the Armitage test for trends using an additive model , as described previously [34] . Combined meta-analysis was performed by using the Mantel-Haenszel procedure with a fixed effect model after testing for heterogeneity . The data from the transfection experiments were analyzed by one-way analysis of variance , followed by Scheffe's test to evaluate statistical differences among 3 groups or by an un-paired t test to evaluate differences between 2 groups .
|
Although cumulative epidemiological findings have suggested that genetic susceptibility plays an important role in the pathogenesis of diabetic nephropathy , no gene conferring susceptibility to diabetic nephropathy has been definitively identified . In a large-scale association study of 1 , 312 Japanese subjects with type 2 diabetes using SNPs from a Japanese SNP database , we show that the T-allele of ACACB rs2268388 is associated with diabetic nephropathy . We also show that the association is consistently observed in patients with type 2 diabetes and proteinuria across different ethnic groups , including populations of European descent . Because a DNA fragment corresponding to the disease susceptibility allele is shown to have higher enhancer activity , we hypothesize that the increase in the expression and/or activity of the encoded acetyl-coenzyme A carboxylase beta contributes to the development and progression of diabetic nephropathy . Our present analysis provides novel insight into the pathogenesis of diabetic nephropathy . This finding is important because diabetic nephropathy is a leading cause of end-stage renal disease and affects life expectancy in subjects with type 2 diabetes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/complex",
"traits",
"nephrology",
"diabetes",
"and",
"endocrinology/type",
"2",
"diabetes"
] |
2010
|
A Single Nucleotide Polymorphism within the Acetyl-Coenzyme A Carboxylase Beta Gene Is Associated with Proteinuria in Patients with Type 2 Diabetes
|
The removal of introns from eukaryotic RNA transcripts requires the activities of five multi-component ribonucleoprotein complexes and numerous associated proteins . The lack of mutations affecting splicing factors essential for animal survival has limited the study of the in vivo regulation of splicing . From a screen for suppressors of the Caenorhabditis elegans unc-93 ( e1500 ) rubberband Unc phenotype , we identified mutations in genes that encode the C . elegans orthologs of two splicing factors , the U2AF large subunit ( UAF-1 ) and SF1/BBP ( SFA-1 ) . The uaf-1 ( n4588 ) mutation resulted in temperature-sensitive lethality and caused the unc-93 RNA transcript to be spliced using a cryptic 3′ splice site generated by the unc-93 ( e1500 ) missense mutation . The sfa-1 ( n4562 ) mutation did not cause the utilization of this cryptic 3′ splice site . We isolated four uaf-1 ( n4588 ) intragenic suppressors that restored the viability of uaf-1 mutants at 25°C . These suppressors differentially affected the recognition of the cryptic 3′ splice site and implicated a small region of UAF-1 between the U2AF small subunit-interaction domain and the first RNA recognition motif in affecting the choice of 3′ splice site . We constructed a reporter for unc-93 splicing and using site-directed mutagenesis found that the position of the cryptic splice site affects its recognition . We also identified nucleotides of the endogenous 3′ splice site important for recognition by wild-type UAF-1 . Our genetic and molecular analyses suggested that the phenotypic suppression of the unc-93 ( e1500 ) Unc phenotype by uaf-1 ( n4588 ) and sfa-1 ( n4562 ) was likely caused by altered splicing of an unknown gene . Our observations provide in vivo evidence that UAF-1 can act in regulating 3′ splice-site choice and establish a system that can be used to investigate the in vivo regulation of RNA splicing in C . elegans .
Eukaryotic genes contain intervening introns that are spliced from transcribed pre-mRNAs to generate functional coding mRNAs [1] , [2] . Alternative splicing results in distinct mRNAs that encode proteins with distinct functions , increases the proteome size and is believed to be important to the biological complexity of metazoans [1] , [3] , [4] . In C . elegans , mRNA transcripts of at least 13% of predicted genes are alternatively spliced [5] . In humans , most genes are alternatively spliced [6] , [7] . A dramatic example of alternative splicing is provided by the Drosophila gene Dscam ( Down syndrome cell adhesion molecule ) , which through alternative splicing could potentially generate over 30 , 000 isoforms [8] , some of which have been shown to play important roles in immune responses [9] and neuronal arborization [10]–[12] . Mutations affecting the splicing process or splicing machinery cause numerous human diseases [13] , [14] . Pre-mRNA splicing involves five small nuclear ribonucleoprotein particles ( snRNPs ) and numerous associated factors [1] , [2] , [15] . The U1 snRNP recognizes the 5′ splice donor site through base-pairing between the U1 snRNA and the 5′ splice site of the target intron [16] . The recognition of the 3′ splice acceptor site is achieved by SF1/BBP ( splicing factor one/branch-point binding protein ) and the large and small subunits of U2AF ( U2 auxiliary factor ) [17]–[23] . In mammals , SF1/BBP binds a weak consensus branch-point sequence , the U2AF large subunit binds a long polypyrimidine sequence and the U2AF small subunit binds the 3′ splice site YAG [19] , [21] , [24]–[26] . The yeast Saccharomyces cerevisiae lacks a U2AF small subunit and a polypyrimidine sequence in its introns , and the recognition of a 3′ splice site is achieved by binding of SF1/BBP to a highly conserved consensus branch-point sequence [17] , [24] , [27] , [28] . In the nematode Caenorhabditis elegans , there is no consensus branch-point sequence or long polypyrimidine sequence , and the recognition of a 3′ splice site is achieved by the binding of the U2AF large and small subunits to a consensus UUUUCAGR sequence in which “AG” is the 3′ splice site [23] , [29] . Splicing is also regulated by many Arginine-Serine-rich RNA-binding SR proteins [30]–[33] and hnRNP RNA-binding proteins [4] . These splicing factors recognize enhancer or silencer sequences in exons and introns to regulate the specificity and efficiency of splicing [4] . The genetic interactions among splicing factors and how signaling events regulate splicing efficiency and specificity are only partially understood . C . elegans is a genetically tractable organism and has been used to study a broad variety of biological problems . Our laboratory has analyzed a set of genes , unc-93 , sup-9 and sup-10 , that encode components of a presumptive C . elegans two-pore domain K+ channel complex and regulate muscle activity [34]–[37] . Rare gain-of-function ( gf ) mutations in any of these three genes cause abnormal body-muscle contraction and are thought to activate the SUP-9 K+ channel . The gf mutant animals are defective in egg laying , sluggish and exhibit a rubberband phenotype: when prodded on the head , the animal contracts and relaxes along its entire body without moving backwards . Complete loss-of-function ( lf ) mutations of unc-93 , sup-9 and sup-10 do not cause any obvious abnormalities [35] , [36] . The SUP-9 protein is similar to the mammalian Two-pore Acid Sensitive K+ channels TASK-1 and TASK-3 [34] . sup-10 encodes a novel single-transmembrane domain protein without identified mammalian homologs [34] . unc-93 encodes a multiple transmembrane-domain protein that defines a novel family of proteins conserved from C . elegans to mammals [34] , [37] . A mammalian UNC-93 homolog , UNC-93b , plays important roles in the innate immune response , probably by regulating signals mediated through Toll-like receptors [38]–[42] . Previous genetic screens for genes that affect the activities of unc-93 , sup-9 and sup-10 genes were not designed to identify genes essential for fertility or animal survival . To seek such essential genes , we performed a clonal genetic screen for suppressors of the locomotion defect caused by the unc-93 gf mutation e1500 . In this paper , we describe our studies of two suppressors identified from this screen and the establishment of a reporter system for in vivo analysis of RNA splicing in C . elegans . We suggest that the U2AF large subunit affects 3′ splice site recognition and that some aspect of the function of the putative UNC-93/SUP-9/SUP-10 two-pore domain potassium channel complex depends on an unidentified gene the processing of which requires the functions of the U2AF large subunit and SF1/BBP .
Rare gf mutations in the C . elegans genes unc-93 , sup-9 and sup-10 cause a rubberband Unc phenotype , while lf mutations in these genes result in a phenotypically wild-type phenotype [35] , [36] , [43] . Previous screens for suppressors of the rubberband Unc phenotype were not designed to identify genes essential for animal survival [35] , [36] , [43] , [44] . We performed a clonal genetic screen to seek new suppressors of the rubberband Unc phenotype caused by the unc-93 ( e1500 ) mutation , with the goal of identifying mutations that also cause sterility or lethality ( see Materials and Methods ) . We screened about 10 , 000 F1 progeny ( about 20 , 000 mutagenized haploid genomes ) of P0 animals mutagenized with EMS ( ethyl methanesulfonate ) and isolated the suppressors n4588 and n4562 . n4588 causes embryonic lethality at 25°C , and n4562 causes sterility at all temperatures . By mapping these mutations , we found that n4588 and n4562 are not alleles of any previously characterized suppressors of the rubberband Unc phenotype ( see Materials and Methods ) . n4588 is a strong recessive suppressor of the locomotion defect and rubberband phenotype of unc-93 ( e1500 ) animals ( Table 1 ) . n4588 is or is closely linked to a mutation that causes a recessive temperature-sensitive ( ts ) lethal phenotype and results in embryonic lethality at 25°C ( see Materials and Methods ) ( data not shown ) . At 20°C the lethal phenotype was incompletely penetrant . At 15°C n4588 animals appeared similar to wild-type animals . We mapped the ts-lethal phenotype of n4588 animals to an 80 kb region on the left arm of LG III ( see Materials and Methods ) . By determining the sequences of the coding exons of four of eight genes located within this 80 kb interval , we found a point mutation in the third coding exon of the major isoform of the gene uaf-1 ( uaf-1a in Figure 1A ) , changing codon 180 from ACT to ATT , a change predicted to replace a conserved threonine with an isoleucine . uaf-1a encodes the C . elegans ortholog of the highly conserved U2AF large subunit ( U2AF , U2 auxiliary factor ) [45] . In mammals , the U2AF large subunit binds a polypyrimidine sequence preceding the 3′ splice site [25] , [46] to regulate pre-mRNA splicing . In C . elegans , together with the U2AF small subunit ortholog UAF-2 , UAF-1 binds a consensus UUUUCAGR sequence , in which AG is the 3′ splice site [23] , [29] . The U2AF large subunit contains an RS-rich ( Arginine-Serine ) domain ( Figure 1B , RS ) , a U2AF small subunit-interacting domain ( Figure 1B , W ) , two RRM ( RNA recognition motif ) domains ( Figure 1B , RRM ) [26] , [47] and a C-terminal UHM ( U2AF homology motif ) domain that binds the splicing factor SF1/BBP [48] . The T180I change caused by the n4588 mutation lies between the U2AF small subunit-interacting domain and the first RRM domain of UAF-1a ( Figure 1B ) . To test whether the point mutation found in the uaf-1a isoform caused the suppressor activity of n4588 , we generated transgenic animals expressing a UAF-1a::GFP fusion protein under the control of a myo-3 myosin promoter , which drives transgene expression in body-wall muscle cells [49] . This uaf-1a cDNA , which encodes a predicted full-length UAF-1 protein , restored the Unc phenotype when expressed in uaf-1 ( n4588 ) unc-93 ( e1500 ) animals ( Figure 1B and Table 2 ) . A predicted short uaf-1 isoform , uaf-1b , which contains only part of the second RRM domain and the C-terminal UHM domain , failed to restore the Unc phenotype ( Figure 1B and Table 2 ) . Expression of these myo-3-driven transgenes ( uaf-1a and uaf-1b ) in wild-type animals did not cause a rubberband Unc phenotype or any other visible abnormality ( data not shown ) . Introducing stop codons or the n4588 T180I mutation into the full-length uaf-1a cDNA abrogated its rescuing activity ( Table 2 ) . Heat-shock-driven expression of a transgene expressing the full-length uaf-1a cDNA under control of a heat-shock promoter [50] partially rescued both the suppression of unc-93 ( e1500 ) by uaf-1 ( n4588 ) and the ts-lethality caused by uaf-1 ( n4588 ) ( Table 2 and data not shown ) , suggesting that the T180I mutation also caused the ts-lethal phenotype . Feeding unc-93 ( e1500 ) animals with uaf-1 RNAi-expressing bacteria ( Figure 1B and Materials and Methods ) also partially suppressed the Unc phenotype ( Table 2 ) , suggesting that normal expression of uaf-1 is required for the rubberband Unc phenotype caused by unc-93 ( e1500 ) . We isolated a uaf-1 deletion mutation , n5222Δ , which removes the fourth exon ( encoding part of the first RRM and part of the second RRM of UAF-1a ) of the uaf-1a isoform ( Figure 1A ) and is predicted to cause a frameshift after amino acid 229 if the third and fifth exons of the uaf-1a isoform are spliced together . uaf-1 ( n5222Δ ) /+ animals grew and moved like the wild type , and uaf-1 ( n5222Δ ) /+ did not suppress the rubberband Unc phenotype of unc-93 ( e1500 ) animals ( data not shown ) . uaf-1 ( n5222Δ ) homozygous mutants arrested and died at the late L1 to early L2 larval stages ( based on body size ) , which precluded examination of the rubberband Unc behavior of n5222Δ homozygous animals ( see Materials and Methods ) . uaf-1 ( n4588 ) /uaf-1 ( n5222Δ ) suppressed the rubberband Unc phenotype of unc-93 ( e1500 ) animals as strongly as did homozygous uaf-1 ( n4588 ) ( Table 1 ) . Similar to uaf-1 ( n4588 ) homozygotes , uaf-1 ( n4588 ) /uaf-1 ( n5222Δ ) animals died embryonically at 25°C ( data not shown ) . These results establish that n4588 is an allele of uaf-1 and that reducing the dosage of the uaf-1 ( n4588 ) allele by 50% does not affect the suppression of the rubberband Unc phenotype of unc-93 ( e1500 ) animals . These data suggest that uaf-1 ( n4588 ) causes either a reduction/loss of uaf-1 activity or an altered uaf-1 activity that is antagonized by the wild-type uaf-1 gene ( see Discussion ) . Locomotion defects similar to those caused by unc-93 ( e1500 ) are also caused by the unc-93 ( n200 ) mutation [36] and by gf mutations in the genes sup-9 and sup-10 [35] , [43] . We tested whether the uaf-1 ( n4588 ) mutation could suppress the Unc phenotype caused by these other mutations ( Table 1 ) . Neither the weak locomotion defect nor the weak rubberband Unc defect caused by unc-93 ( n200 ) was suppressed by uaf-1 ( n4588 ) ( Table 1 ) . sup-10 ( n983 ) , which causes a rubberband Unc phenotype that is more severe than that of unc-93 ( n200 ) animals but less severe than that of unc-93 ( e1500 ) animals , was completely suppressed by uaf-1 ( n4588 ) ( Table 1 ) . The strongest rubberband mutant , sup-9 ( n1550 ) [43] , was not suppressed by uaf-1 ( n4588 ) ( Table 1 ) . These data suggest that uaf-1 ( n4588 ) is an allele-specific suppressor of unc-93 but not a gene-specific suppressor of the rubberband Unc mutants and is distinct in its suppression pattern from other known suppressors of unc-93 , sup-9 and sup-10 ( see Discussion ) . Null mutations of unc-93 , sup-10 and sup-9 do not cause visible abnormalities in a wild-type background [35] , [36] . We tested whether these genes might function redundantly with uaf-1 , by generating double mutants containing uaf-1 ( n4588 ) and null mutations of unc-93 , sup-10 or sup-9 . We found that such double mutant animals grew and behaved indistinguishably from uaf-1 ( n4588 ) single mutant animals ( Table S1 ) , suggesting that unc-93 , sup-10 and sup-9 are not functionally redundant with uaf-1 . To examine whether uaf-1 ( n4588 ) can suppress gf mutations affecting other two-pore domain potassium channels , we generated double mutant animals containing the uaf-1 ( n4588 ) mutation and the unc-58 ( e665sd ) [51] ( J . Thomas , personal communication ) , egl-23 ( n601sd ) [52] ( J . Thomas , personal communication ) or twk-18 ( e1913sd ) [53] mutations . The behavioral defects of these mutants were not suppressed by uaf-1 ( n4588 ) ( Table S1 ) . To determine if the expression of unc-93 , sup-9 , sup-10 or any of the other genes known to interact with these genes is reduced in uaf-1 ( n4588 ) animals , we examined unc-93 , sup-10 , sup-9 , sup-18 [35] ( I . Perez de la Cruz and H . R . H . , unpublished results ) and sup-11 [54] ( E . Alvarez-Saavedra and H . R . H , unpublished results ) mRNA levels . Like lf mutations in unc-93 , sup-10 and sup-9 , lf mutations in sup-18 and gf mutations in sup-11 can suppress the rubberband Unc phenotype caused by gf mutations in unc-93 , sup-9 and sup-10 . Using real-time qRT–PCR , we found no obvious reduction of the mRNA levels of these five genes ( Figure S1 ) . We also examined the expression of UAF-1 protein using western blotting [45] and found no apparent difference in UAF-1 protein levels between wild-type and uaf-1 ( n4588 ) animals ( data not shown ) , suggesting that the suppression of unc-93 ( e1500 ) by uaf-1 ( n4588 ) is not caused by a reduction of the level of the UAF-1 protein . We tested whether the splicing of unc-93 is altered by uaf-1 ( n4588 ) . We examined the splicing of each exon of unc-93 in wild-type , uaf-1 ( n4588 ) , unc-93 ( e1500 ) , uaf-1 ( n4588 ) unc-93 ( e1500 ) , unc-93 ( n200 ) and uaf-1 ( n4588 ) unc-93 ( n200 ) animals by RT–PCR ( Figure 2 ) . Every exon other than exon 9 of the unc-93 gene was spliced similarly in all genotypes examined ( Figure 2A and 2B ) . However , we had difficulty in consistently amplifying a cDNA band from uaf-1 ( n4588 ) unc-93 ( e1500 ) animals ( data not shown ) using the PCR primer pairs at the 3′ end of exon 8 and the 5′ end of exon 9 ( indicated in black in Figure 2A ) . We therefore used a new pair of PCR primers that should amplify a larger region between exons 8 and 9 ( Figure 2A , red arrows ) . With the new pair of PCR primers , we found that in unc-93 ( e1500 ) animals the region between exon 8 and 9 corresponded to a weak but consistent RT–PCR product of a reduced length ( Figure 2C , lane 3 , lower arrow ) , and this RT–PCR product was seen only in samples from unc-93 ( e1500 ) mutant animals ( Figure 2C , lower arrow ) . In uaf-1 ( n4588 ) unc-93 ( e1500 ) animals , the RT–PCR product of reduced length was the most prominent product ( Figure 2C , lane 4 , lower arrow ) . We determined the sequence of this RT–PCR product and found that it was a consequence of an alternative splicing event that utilized a cryptic 3′ splice site in exon 9 . This cryptic 3′ splice site was generated by the unc-93 ( e1500 ) missense mutation , which has a G-to-A transition that changes amino acid 388 from Gly to Arg [37] ( Figure 2E ) . Quantification using Taqman RT–PCR ( see Figure 4A for probe designs ) indicated that the alternatively spliced exon 9 was about 1 . 3% of all spliced unc-93 exon 9 in unc-93 ( e1500 ) animals and 68% in uaf-1 ( n4588 ) unc-93 ( e1500 ) animals ( Figure 4B ) . Both non-quantitative ( Figure 2C ) and quantitative RT–PCR ( Figure 4B ) analyses failed to detect alternatively spliced exon 9 from wild-type , uaf-1 ( n4588 ) , unc-93 ( n200 ) or uaf-1 ( n4588 ) unc-93 ( n200 ) animals , all of which lack the cryptic 3′ splice site caused by the unc-93 ( e1500 ) mutation . The alternatively spliced unc-93 transcript is predicted to encode a truncated protein lacking 12 amino acids in one of the predicted transmembrane domains [37] ( data not shown ) . To test whether the alternatively spliced unc-93 transcript in uaf-1 ( n4588 ) unc-93 ( e1500 ) animals encoded a functional UNC-93 protein , we expressed the cDNA in the body-wall muscles of sup-9 ( n1550 ) ; unc-93 ( lr12Δ ) animals [34] , [44] and found that this transgene did not restore the rubberband Unc phenotype ( Table S2 ) . By contrast , expression of the wild-type unc-93 cDNA in these animals restored the severe rubberband Unc phenotype . These results suggested that the alternatively spliced unc-93 transcript encoded a lf UNC-93 protein or possibly a dominant-negative UNC-93 protein . To test the latter possibility , we expressed either unc-93 wild-type cDNA or the alternatively spliced unc-93 cDNA in the body-wall muscles of unc-93 ( e1500 ) animals ( Table S3 ) . Consistent with previous observations that unc-93 ( e1500 ) /+ animals have better locomotion than unc-93 ( e1500 ) animals [36] , [37] , overexpression of wild-type unc-93 cDNA dramatically improved the locomotion of unc-93 ( e1500 ) animals ( Table S3 ) . If the alternatively spliced unc-93 transcript encoded an UNC-93 protein that could interfere with the endogenous UNC-93 function and cause the suppression of the rubberband Unc phenotype by uaf-1 ( n4588 ) ( 68% alternatively spliced unc-93 transcript ) , the transgene should also suppress the Unc phenotype of unc-93 ( e1500 ) animals . However , expression of the alternatively spliced unc-93 transcript in the body-wall muscles did not suppress the Unc phenotype of unc-93 ( e1500 ) animals ( Table S3 ) , suggesting that the alternatively spliced unc-93 transcript caused a loss of unc-93 function and did not interfere with endogenous unc-93 function . To examine whether reducing UAF-1 expression , like the uaf-1 ( n4588 ) mutation , would alter the splicing of unc-93 ( e1500 ) exon 9 , we fed animals with bacteria expressing dsRNA targeting uaf-1 and assessed unc-93 exon 9 splicing . As shown in Figure 2D and Figure 4B , reducing UAF-1 did not increase the relative level of alternatively spliced unc-93 ( e1500 ) exon 9 . The RNAi treatment did significantly reduce the level of UAF-1 protein ( Figure S2 ) . That reducing uaf-1 expression with RNAi did not cause altered splicing of unc-93 ( e1500 ) exon 9 similarly to that by the uaf-1 ( n4588 ) mutation is consistent with the hypothesis that uaf-1 ( n4588 ) does not reduce the function of UAF-1a but rather alters the function of UAF-1a , which leads to the recognition of the cryptic 3′ splice site of unc-93 ( e1500 ) exon 9 ( see Discussion ) . However , it is possible that uaf-1 ( n4588 ) reduces uaf-1 function and that uaf-1 ( RNAi ) does not reduce uaf-1 function as much . uaf-1 ( n4588 ) suppressed the rubberband Unc phenotype of sup-10 ( n983 ) animals but did not suppress the rubberband Unc phenotype of unc-93 ( n200 ) and sup-9 ( n1550 ) animals ( Table 1 ) . Quantitative RT–PCR did not indicate reduction of sup-10 mRNA in uaf-1 ( n4588 ) animals ( Figure S1 ) . We examined whether the sup-10 ( n983 ) transcript was alternatively spliced in uaf-1 ( n4588 ) mutants . uaf-1 ( n4588 ) did not cause the appearance of a sup-10 cDNA band different in size from the full-length sup-10 cDNA ( Figure S3A ) . We determined the sequences of the sup-10 cDNA RT–PCR products from wild-type , sup-10 ( n983 ) , uaf-1 ( n4588 ) and uaf-1 ( n4588 ) ; sup-10 ( n983 ) animals and failed to identify an alternatively spliced sup-10 transcript ( data not shown ) . To test whether uaf-1 ( n4588 ) can affect the splicing of all genes known to be alternatively spliced , we tested for genetic interactions between uaf-1 ( n4588 ) and unc-52 ( e669 ) . unc-52 encodes the C . elegans ortholog of human basement membrane-specific heparan sulfate proteoglycan core protein , and mutations affecting unc-52 cause adult paralysis [55] , [56] . The Unc phenotype of unc-52 ( e669 ) can be suppressed by lf mutations of smu-1 and smu-2 , genes that encode C . elegans homologs of mammalian splicing factors [57]–[59] . The unc-52 ( e669 ) mutation causes a pre-mature stop in unc-52 exon 17 [60] , and smu-1 and smu-2 lf mutations suppress unc-52 ( e669 ) by removing exon 17 and generating an alternatively spliced and functional transcript [58] . The unc-52 ( e444 ) mutation causes a pre-mature stop in unc-52 exon 18 , which is not removed in smu-1 and smu-2 mutant animals , leading to a transcript with a premature stop codon . Double mutants containing the unc-52 ( e444 ) mutation and the smu-1 or smu-2 mutations display an Unc phenotype [58] . We examined the Unc phenotypes of unc-52 ( e669 ) ; uaf-1 ( n4588 ) and unc-52 ( e444 ) ; uaf-1 ( n4588 ) animals and found that uaf-1 ( n4588 ) did not suppress either unc-52 ( e669 ) or unc-52 ( e444 ) ( Table S1 ) , implying that the uaf-1 ( n4588 ) mutation did not affect the alternative splicing of unc-52 ( e669 ) exon 17 and thus does not affect all cases of alternative splicing non-specifically . The mutation n4562 was also isolated from our clonal screen as a suppressor of the rubberband Unc phenotype of unc-93 ( e1500 ) animals . The suppressed phenotype was recessive , and n4562 caused a completely penetrant recessive sterility that was temperature independent and was tightly linked to its suppressor activity ( see Materials and Methods ) . Like uaf-1 ( n4588 ) , n4562 suppressed unc-93 ( e1500 ) and sup-10 ( n983 ) but did not suppress unc-93 ( n200 ) or sup-9 ( n1550 ) ( Table 1 ) . Therefore , n4562 is also an allele-specific suppressor of unc-93 gf mutations but not a gene-specific suppressor for the rubberband Unc genes . We mapped n4562 to the right of LG IV ( see Materials and Methods ) . No known suppressors of unc-93 ( e1500 ) are located in this region . The genes uaf-2 , encoding the C . elegans U2AF small subunit ortholog [61] and Y116A8C . 32 , encoding the SF1/BBP ( splicing factor 1/branch-point binding protein ) ortholog [62] , are located in this genomic region and are expressed from the same operon together with three other genes ( Wormbase WS189 ) [61] , [62] . Orthologs of UAF-2 and SF1/BBP function with the ortholog of UAF-1 to regulate pre-mRNA splicing [2] , leading us to consider these two genes as candidates for being mutated by n4562 . We determined the DNA sequences of coding regions of uaf-2 and Y116A8C . 32 from n4562 animals and identified a nonsense mutation in Y116A8C . 32 , which we named sfa-1 ( sfa , splicing factor ) ( Figure 3A ) . n4562 changed amino acid 458 from a Cys ( TGT ) to an opal stop ( TGA ) codon in a conserved C2HC-type zinc finger domain of the predicted SFA-1 protein ( Figure 3A and 3B ) . This mutation is predicted to cause the expression of a truncated SFA-1 protein . We rescued the suppression of unc-93 ( e1500 ) by sfa-1 ( n4562 ) by expressing in body-wall muscles an SFA-1::GFP fusion protein driven by the myo-3 promoter [49] ( Table 2 ) . Feeding unc-93 ( e1500 ) animals with bacteria expressing dsRNA targeting sfa-1 partially suppressed the rubberband Unc phenotype ( Table 2 ) . We isolated an sfa-1 deletion mutation , n5223Δ , which removes the third and fourth exons and a majority of the fifth exon ( Figure 3A ) . Together these regions are predicted to encode most ( 101 aa ) of the U2AF large subunit-interacting domain ( 118 aa ) of SFA-1 [62] . n5223Δ is predicted to cause a frameshift after amino acid 188 if the second exon and the residual fifth exon are spliced together . sfa-1 ( n5223Δ ) caused recessive embryonic lethality , and sfa-1 ( n5223Δ ) /+ did not suppress the rubberband Unc phenotype of unc-93 ( e1500 ) animals ( data not shown ) . sfa-1 ( n4562 ) / ( n5223Δ ) similarly caused embryonic lethalilty ( data not shown ) , suggesting that the lethal phenotype of sfa-1 ( n5223Δ ) homozygotes is caused by the sfa-1 ( n5223Δ ) mutation . The embryonic lethality caused by sfa-1 ( n5223Δ ) and sfa-1 ( n4562 ) /sfa-1 ( n5223Δ ) precluded the use of n5223Δ for an analysis of the rubberband Unc phenotype , because our behavioral assay is performed with young adults ( see Materials and Methods ) . To test whether , like uaf-1 ( n4588 ) , sfa-1 ( n4562 ) caused alternative splicing of unc-93 ( e1500 ) exon 9 , we used RT–PCR to examine the splicing of unc-93 ( e1500 ) exon 9 . As shown in Figure 3C and Figure 4B , sfa-1 ( n4562 ) did not cause increased alternative splicing of unc-93 ( e1500 ) exon 9 . We tested the effect of sfa-1 on unc-93 ( e1500 ) exon 9 splicing by reducing sfa-1 expression using RNAi ( Figure 3D ) . sfa-1 ( RNAi ) did not increase exon 9 alternative splicing ( Figure 3D and Figure 4B ) . Because ( 1 ) the sfa-1 ( n4562 ) mutation causes a recessive sterile phenotype , which is less severe than the recessive embryonic lethality caused by sfa-1 ( n5223Δ ) ( likely a null allele ) or by sfa-1 ( n4562 ) /sfa-1 ( n5223Δ ) , ( 2 ) sfa-1 ( RNAi ) phenocopies sfa-1 ( n4562 ) in the suppression of the rubberband Unc phenotype of unc-93 ( e1500 ) animals , and ( 3 ) sfa-1 ( RNAi ) phenocopies sfa-1 ( n4562 ) in affecting the splicing of unc-93 ( e1500 ) exon 9 , we propose that sfa-1 ( n4562 ) is a partial lf allele of sfa-1 and that the suppression of the rubberband Unc phenotype of unc-93 ( e1500 ) animals by sfa-1 ( n4562 ) is likely caused by reduced sfa-1 function . To examine whether sfa-1 ( n4562 ) affects the alternative splicing of unc-93 ( e1500 ) exon 9 caused by the uaf-1 ( n4588 ) mutation , we generated uaf-1 ( n4588 ) ; sfa-1 ( n4562 ) animals with or without unc-93 ( e1500 ) . Most uaf-1 ( n4588 ) ; sfa-1 ( n4562 ) animals ( with or without unc-93 ( e1500 ) ) died embryonically , and the few that hatched arrested at the L2 larval stage ( based on body size ) ( Figure S4 ) . We failed to obtain a sufficient number of animals for RT–PCR analysis . We also examined sup-10 splicing in sfa-1 ( n4562 ) animals and failed to detect alternative splicing of the sup-10 transcript ( Figure S3B ) . The ts-lethality of uaf-1 ( n4588 ) offered a genetic approach to seek new regulators of RNA splicing by screening for suppressors of the ts-lethal phenotype . We performed a genetic screen for suppressors of uaf-1 ( n4588 ) ts-lethality at 25°C . From this screen , we isolated four intragenic suppressors , n5120 , n5123 , n5125 and n5127 ( Table 3 ) and seven extragenic suppressors ( see Materials and Methods ) . To date we have characterized only the intragenic suppressors . uaf-1 ( n5123 ) caused an I180F ( ATT-to-TTT ) change at the same residue mutated by n4588 ( T180I ) ( ACT-to-ATT ) ( Table 3 and Figure 5 ) and eliminated the suppression of the locomotion defect of unc-93 ( e1500 ) ( Table 4 ) . uaf-1 ( n4588 n5120 ) is predicted to cause a V179M ( GTG-to-ATG ) change in addition to the n4588 T180I mutation ( Table 3 and Figure 5 ) , and uaf-1 ( n4588 n5120 ) weakly suppressed unc-93 ( e1500 ) ( Table 4 ) . uaf-1 ( n4588 n5125 ) is predicted to cause a P177L ( CCA-to-CTA ) change in addition to the n4588 T180I mutation ( Table 3 and Figure 5 ) , and uaf-1 ( n4588 n5125 ) also weakly suppressed unc-93 ( e1500 ) ( Table 4 ) . uaf-1 ( n4588 n5127 ) is predicted to cause an M157I ( ATG-to-ATA ) change in addition to the n4588 T180I mutation ( Table 3 and Figure 5 ) and was still a strong suppressor of unc-93 ( e1500 ) ( Table 4 ) . That the unc-93 ( e1500 ) suppressor activities of uaf-1 ( n4588 n5120 ) , uaf-1 ( n4588 n5125 ) and uaf-1 ( n4588 n5127 ) were caused by uaf-1 mutations was confirmed by the observation that transgenes expressing uaf-1a ( Figure 1A ) in the body-wall muscles rescued the unc-93 ( e1500 ) suppressor phenotype of these mutants ( Table 4 ) . We quantified the splicing of unc-93 ( e1500 ) exon 9 in animals containing these uaf-1 mutations using Taqman RT–PCR . As shown in Figure 4B , these uaf-1 mutations exhibited differential effects on the splicing of unc-93 ( e1500 ) exon 9 . uaf-1 ( n5123 ) had no apparent effect ( 1 . 6% alternative splicing vs . 1 . 3% for unc-93 ( e1500 ) alone ) . uaf-1 ( n4588 n5120 ) and uaf-1 ( n4588 n5125 ) weakly ( 4% and 2 . 8% , respectively ) and uaf-1 ( n4588 n5127 ) moderately ( 25% ) increased the alternative splicing of unc-93 ( e1500 ) exon 9 ( Figure 4B ) . None of these mutations affected this splicing event as much as did uaf-1 ( n4588 ) ( 68% ) ( Figure 4B ) . To test whether uaf-1 ( n4588 ) could suppress the Unc phenotype of unc-93 ( e1500 ) independently of unc-93 splicing , we generated transgenic animals that overexpressed in the body-wall muscles a full-length unc-93 cDNA containing the e1500 gf mutation . Since this cDNA generates a fully spliced form of unc-93 mRNA , if alternative splicing of unc-93 ( e1500 ) accounted for the suppression of the rubberband Unc phenotype , overexpressed unc-93 ( e1500 ) cDNA would not generate an alternatively spliced isoform and the animals would be as Unc in an uaf-1 ( n4588 ) background as in a wild-type background . As shown in Table 5 , in wild-type animals , overexpression of the unc-93 ( e1500 ) cDNA caused a strong rubberband Unc phenotype . The presence of the uaf-1 ( n4588 ) mutation reduced the severity of the rubberband Unc phenotype caused by the same transgenes . Similarly , overexpression of the unc-93 ( e1500 ) cDNA also caused a weaker rubberband Unc phenotype in sfa-1 ( n4562 ) animals than in wild-type animals ( Table 5 ) . This result implied that uaf-1 ( n4588 ) can suppress the unc-93 ( e1500 ) rubberband Unc phenotype through mechanism ( s ) other than by affecting the splicing of unc-93 . This finding was consistent with our results showing that although sfa-1 ( n4562 ) suppressed the rubberband Unc phenotype of unc-93 ( e1500 ) animals , sfa-1 ( n4562 ) did not affect the alternative splicing of unc-93 ( e1500 ) exon 9 , which also suggested that the suppression of unc-93 ( e1500 ) by sfa-1 ( n4562 ) was mediated by a mechanism other than by affecting the alternative splicing of unc-93 ( see Discussion ) . The alternative splicing between the intron 8 endogenous 3′ splice site ( I8 ) and the exon 9 cryptic 3′ splice site ( E9 ) in wild-type and uaf-1 mutant animals allows an analysis of the effects of different nucleotides on the in vivo recognition of these alternatively spliced sites . We constructed a transgene that fuses the genomic sequence between exon 8 and exon 10 of unc-93 ( e1500 ) and the GFP reporter gene ( Figure 6A ) and placed the fusion transgene under the control of 2 kb of the promoter region of unc-93 . We used a pair of PCR primers ( Figure 6A , red arrows ) that recognize unc-93 exon 8 and the GFP sequences to specifically amplify transgene cDNAs in RT–PCR experiments . The Taqman probes shown in Figure 4A were used to quantify the wild-type and alternatively spliced isoforms ( Figure 6A ) . Because I8 and E9 have the same nucleotides at positions −3 to −1 ( CAG ) , our mutagenesis analysis focused on nucleotides −7 to −4 ( Figure 6B ) , which are variable and are known to be critical for recognition and binding by the U2AF complex [29] , [63]–[65] . We named each of 16 transgene constructs 1–16 ( Figure 6B–6E ) . We examined the splicing of a transgene ( Figure 6B and 6E , No . 1 ) containing the same I8 and E9 as unc-93 ( e1500 ) . In wild-type animals the splicing mimics that of the endogenous unc-93 ( e1500 ) , with very little splicing at E9 ( Figure 6B , 1 . 8% , compare to 1 . 3% in Figure 4B ) . Splicing of the same transgene ( No . 1 ) in uaf-1 ( n4588 ) mutants occurred almost exclusively at E9 ( >99% ) ( Figure 6B ) ; endogenous splicing of unc-93 ( e1500 ) was qualitatively but not quantitatively similar , occurring mostly at E9 ( 68% ) ( Figure 4B ) . These results suggest that the splicing of mutated unc-93 transgenes could provide important information concerning the in vivo recognition of 3′ splice sites . We replaced E9 with the sequence of I8 ( Figure 6B , No . 2 ) . In both wild-type and uaf-1 ( n4588 ) animals , splicing occurred mostly at the new E9 ( 97% and 98% , respectively ) ( Figure 6B and 6E ) . When I8 was replaced with the sequence of E9 ( Figure 6B , No . 3 ) , splicing again occurred mostly at E9 in both wild-type and uaf-1 ( n4588 ) animals ( 97% and >99% , respectively ) ( Figure 6B and 6E ) . These results suggest that the sequence that surrounds the original E9 is preferred by the splicing machinery in both wild-type and uaf-1 ( n4588 ) animals when two identical 3′ splice sites are present . We next switched the positions of I8 and E9 ( Figure 6B , No . 4 ) . In the wild type most splicing ( >99% ) occurred at the new E9 ( Figure 6B and 6E ) . Similarly , in uaf-1 ( n4588 ) animals , most splicing ( 80% ) occurred at the new E9 ( Figure 6B and 6E , No . 4 ) . However , that a significant amount of splicing ( 20% ) occurred at the new I8 in uaf-1 ( n4588 ) animals ( Figure 6B and 6E , No . 4 ) suggested that the mutant UAF-1 can efficiently recognize the original E9 sequence even at the I8 position , which is normally a less favorable position . The pattern of alternative splicing in cell culture can depend on the promoter used [66] . unc-93 is expressed in body-wall muscles [34] , [37] . We tested whether a different muscle-specific promoter would alter the splicing pattern of transgene No . 1 by expressing the transgene under the control of a myo-3 promoter [49] ( Figure 6B and 6E , No . 5 ) . We found almost identical splicing patterns of the transgene driven by the myo-3 promoter and the unc-93 promoter ( Figure 6B and 6E , compare No . 1 and No . 5 ) , suggesting that the alternative splicing of unc-93 ( e1500 ) involves a mechanism that is not promoter-specific . We examined the effects of base substitutions at I8 . Replacing I8 with the C . elegans consensus 3′ splice site TTTTcag [29] , [63]–[65] caused splicing to occur exclusively ( 100% ) at the new I8 in both wild-type and uaf-1 ( n4588 ) animals ( Figure 6C and 6E , No . 6 ) . To identify the nucleotides required for the recognition of I8 in wild-type animals , we substituted each base from −7 to −4 of I8 with a G ( Figure 6C , No . 7 to No . 10 ) . G is the least used nucleotide from −7 to −4 of identified 3′ splice sites [29] , [64] and in previous studies substituting T with G at any of the four bases from −7 to −4 of the highly consensus TTTTcag site significantly compromised binding of the U2AF complex to this site [29] . A G substitution at −7 ( No . 7 ) , −5 ( No . 9 ) and −4 ( No . 10 ) of I8 all dramatically reduced splicing at the new I8 ( to the level of 15% , 0% , 0% , respectively; Figure 6C and 6E ) in wild-type animals , suggesting that these nucleotides are critical for the recognition by wild-type UAF-1 . However , a G substitution at −6 ( Figure 6C and 6E , No . 8 ) did not cause a significant change of splicing at the new I8 ( which is 96% compared to 98% of No . 1 ) , suggesting this nucleotide is not essential for recognition by wild-type UAF-1 . We also substituted the A at −6 with a C to generate an I8 more similar to E9 ( Figure 6C and 6E , No . 11 ) . Splicing at this I8 ( No . 11 ) was similar to that of transgenes No . 1 and No . 8 in wild-type animals , consistent with the notion that this base is not essential for the recognition by wild-type UAF-1 . For all the transgenes with single-base substitutions of I8 , splicing ( Figure 6C and 6E , No . 7 , 8 , 9 , 10 and 11 ) in uaf-1 ( n4588 ) animals is similar to that of transgene No . 1 ( Figure 6B and 6E ) , suggesting that none of the substitutions significantly increased the affinity of I8 for mutant UAF-1 . To test whether we could increase the recognition of E9 , we replaced E9 with the highly conserved consensus 3′ splice site TTTTcag sequence [29] , [63]–[65] ( Figure 6D and 6E , No . 12 ) . As expected , in both wild-type and uaf-1 ( n4588 ) animals , splicing occurred exclusively at the new E9 ( 100% and 100% , respectively ) . We next changed each of the non-T bases to T from −7 to −4 of E9 ( Figure 6D and 6E , No . 13 , 14 and 15 ) . A T at −7 , −6 or −4 increased splicing at E9 in wild-type animals ( Figure 6D and 6E , No . 13 , 14 and 15 ) ( 53% , 100% and 91% for positions −7 , −6 and −4 , respectively ) , suggesting these substitutions increased recognition of E9 by wild-type UAF-1 . For all three of these transgenes , splicing in uaf-1 ( n4588 ) animals occurred exclusively at E9 ( 100% for all three ) ( Figure 6D and 6E , No . 13 , 14 , and 15 ) , suggesting none of the T substitutions significantly reduced the recognition of E9 by mutant UAF-1 . We mutated the C at −6 of E9 to an A ( Figure 6D and 6E , No . 16 ) , generating a 3′ splice site with a T-to-G substitution at −4 of transgene No . 2 ( Figure 6B , 6D , and 6E ) . Splicing of this transgene occurred exclusively at I8 ( 100% ) in wild-type animals ( Figure 6D and 6E , No . 16 ) . In uaf-1 ( n4588 ) animals , splicing at the new E9 was reduced ( to 62% , Figure 6D and 6E , No . 16 ) compared to that at E9 of transgene No . 2 ( 100% , Figure 6B and 6E , No . 2 ) .
The mechanism ( s ) of the suppression of the Unc phenotype caused by the unc-93 ( e1500 ) mutation by uaf-1 ( n4588 ) and sfa-1 ( n4562 ) remains to be determined . Four observations indicate that although uaf-1 ( n4588 ) causes alternative splicing of unc-93 ( e1500 ) exon 9 , this alternative splicing is not the basis of the suppression . First , unc-93 ( e1500 ) /unc-93 ( lf ) heterozygous animals are as Unc as unc-93 ( e1500 ) homozygous animals [36] , indicating that reducing unc-93 expression by 50% does not reduce the rubberband Unc phenotype . By contrast , uaf-1 ( n4588 n5127 ) reduced unc-93 ( e1500 ) expression by 25% ( since there was 25% alternative splicing ) , and these animals were strongly suppressed . Also , in uaf-1 ( n4588 ) unc-93 ( e1500 ) animals , the unc-93 ( e1500 ) transcript was reduced by 68% ( there was 68% alternative splicing ) , and these animals might have been expected to be slightly less Unc than unc-93 ( e1500 ) /unc-93 ( lf ) animals but instead were strongly suppressed . Thus , the level of reduction of the unc-93 ( e1500 ) transcript does not correlate with the level of the suppression of the unc-93 ( e1500 ) Unc phenotype by uaf-1 ( n4588 n5127 ) and uaf-1 ( n4588 ) . Second , the strong rubberband Unc phenotype caused by overexpression of the unc-93 ( e1500 ) -specific cDNA in body-wall muscles was partially suppressed by uaf-1 ( n4588 ) , suggesting that unc-93 ( e1500 ) splicing is not needed for uaf-1 ( n4588 ) -mediated suppression . Third , sfa-1 ( n4562 ) suppressed unc-93 ( e1500 ) without affecting the splicing of unc-93 ( e1500 ) exon 9 , and sfa-1 ( n4562 ) partially suppressed the rubberband Unc phenotype caused by overexpression of the unc-93 ( e1500 ) cDNA in the body-wall muscles . Again , suppression can occur without affecting unc-93 ( e1500 ) mRNA splicing . Similarly , we did not identify an alternatively spliced sup-10 transcript in either uaf-1 ( n4588 ) ; sup-10 ( n983 ) or sfa-1 ( n4562 ) ; sup-10 ( n983 ) animals , suggesting that sup-10 ( n983 ) was suppressed by uaf-1 ( n4588 ) and sfa-1 ( n4562 ) by a mechanism other than alternative splicing of the sup-10 transcript . Fourth , reducing the expression of uaf-1 and sfa-1 by RNAi suppressed the rubberband Unc phenotype of unc-93 ( e1500 ) but did not cause altered splicing of unc-93 ( e1500 ) exon 9 . Based on these arguments , we propose that uaf-1 and sfa-1 mutations suppress unc-93 ( e1500 ) and sup-10 ( n983 ) by affecting the splicing of one or more unidentified genes required for the expression of the unc-93 ( e1500 ) and sup-10 ( n983 ) rubberband Unc phenotype . We cannot exclude the possibility that the alternative splicing of unc-93 ( e1500 ) contributed to the suppression of unc-93 ( e1500 ) by uaf-1 mutations . The known suppressors of gf mutations of unc-93 , sup-9 and sup-10 are of three classes . First , lf mutations in any of these three genes are recessive suppressors of the rubberband Unc phenotypes caused by gf mutations in any of these three genes , because the functions of all three genes are necessary for expression of the Unc phenotype [35] , [36] , [43] . Second , rare gf mutations of sup-11 are strong dominant suppressors of unc-93 ( e1500 ) and unc-93 ( n200 ) and partial recessive suppressors of sup-9 ( n1550 ) and sup-10 ( n983 ) [54] . The mechanism of sup-11 suppression is unknown . Third , lf mutations of sup-18 are strong recessive suppressors of sup-10 ( n983 ) and weak recessive suppressors of unc-93 ( e1500 ) , unc-93 ( n200 ) and sup-9 ( n1550 ) [35] . The mechanism of sup-18 suppression is also unknown . uaf-1 ( n4588 ) and sfa-1 ( n4562 ) define a new class of suppressors: they are recessive and allele-specific for unc-93 gf mutations ( unc-93 ( e1500 ) was suppressed , but unc-93 ( n200 ) was not ) but not gene-specific ( sup-10 ( n983 ) was also suppressed ) . Previous genetic and molecular studies from our laboratory led to the hypothesis that UNC-93 , SUP-9 and SUP-10 form a protein complex in the body-wall muscles [34]–[37] . The identification of multiple suppressors of the rubberband Unc phenotype with distinct suppression patterns suggests that the presumptive UNC-93/SUP-9/SUP-10 complex could have multiple in vivo functions regulated in different ways . As mentioned above , we propose that mutations in uaf-1 and sfa-1 affect the splicing of one or more unknown genes required for unc-93 and sup-10 activity . This unknown gene might be required specifically for the expression of the rubberband Unc phenotype caused by unc-93 ( e1500 ) or sup-10 ( n983 ) but have a negligible role in the expression of the rubberband Unc phenotypes caused by unc-93 ( n200 ) or sup-9 ( n1550 ) . Similarly , sup-11 and sup-18 could affect functions of the UNC-93/SUP-9/SUP-10 complex distinct from that affected by uaf-1 and sfa-1 . The SF1/BBP and U2AF proteins are critical splicing factors that regulate splicing by binding the branch-point sequence and the 3′ splice sites [1] , [2] , respectively . Mutations that affect the U2AF subunits and SF1/BBP in the yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe and the fruit fly Drosophila melanogaster have significantly facilitated the understanding of the in vivo function and regulation of these splicing factors [17] , [27] , [67]–[70] . Studies of S . cerevisiae identified genetic and biochemical interactions between the U2AF large subunit and SF1/BBP [17] , [27] , and studies of S . pombe provided in vivo evidence that the U2AF subunits are required for splicing [68] , [70] . In Drosophila null mutations of the U2AF large or small subunits cause lethality [67] , [69] , hindering genetic analysis of these splicing factors . Similarly , in C . elegans , reducing the expression of uaf-1 or sfa-1 by RNAi causes lethality [61] , [62] , suggesting that these genes are essential for animal survival . We identified mutations that affect uaf-1 and sfa-1 and allow the survival of animals in permissive conditions , such as at lower temperatures or when derived from heterozygous mothers . These mutations provide a valuable resource for analyzing the function and regulation of the U2AF large subunit and SF1/BBP genes in vivo in animals . The recognition of 3′ splice sites is achieved by interactions between SF1/BBP and the U2AF large and small subunits , which together bind specific intronic sequences [1] , [2] . However , it is not clear how these factors regulate the choice of the correct splice site when two or more potential 3′ splice sites are proximal in vivo . Distinguishing different 3′ splice sites is a critical aspect of alternative splicing . The unc-93 ( e1500 ) missense mutation generates a new cryptic 3′ splice site ( AG ) within exon 9 ( ACTGcag ) . This site differs from the consensus 3′ splice site for C . elegans ( TTTTcag ) [29] and is more rarely used by C . elegans than is TTTTcag or the intron 8 endogenous 3′ splice site ( AATTcag ) ( Table S4 ) . Based on in vitro studies , this cryptic site should not be or be more weakly recognized by UAF-1 compared to TTTTcag and probably the intron 8 site AATTcag [29] . In a wild-type background , the choice between the wild-type 3′ splice site of unc-93 ( e1500 ) intron 8 and the cryptic non-consensus site in unc-93 ( e1500 ) exon 9 followed this prediction , as only 1 . 3% of the splicing events utilized this cryptic 3′ splice site . Strikingly , however , the n4588 missense mutation in uaf-1 shifted this specificity , causing splicing to occur mostly at the cryptic site , generating 68% of aberrantly spliced transcripts . This result suggests that UAF-1 might play an important role in determining the choice among alternative 3′ splice sites in vivo ( see discussion below ) . The n4588 mutation did not cause an apparent change of UAF-1 protein level , and reducing UAF-1 using RNAi did not increase the relative amount of alternatively spliced exon 9 , suggesting that uaf-1 ( n4588 ) might alter UAF-1 function . However , RNAi-treatment did not abolish the expression of UAF-1 ( Figure S2 ) , and we might have failed to detect an effect of UAF-1 on the splicing of unc-93 ( e1500 ) exon 9 because of residual UAF-1 protein in RNAi-treated animals . Thus , the altered splicing of unc-93 ( e1500 ) exon 9 in uaf-1 ( n4588 ) unc-93 ( e1500 ) mutants might reflect the consequence of the absence of UAF-1 activity . It is also possible that uaf-1 ( n4588 ) causes both a loss of function and an altered function of UAF-1 , which cause the suppression of the rubberband Unc phenotype of the unc-93 ( e1500 ) animals and the altered splicing of unc-93 ( e1500 ) transcript , respectively . The other 14 introns of the unc-93 transcript appeared to be spliced similarly in wild-type and uaf-1 ( n4588 ) animals , suggesting that uaf-1 ( n4588 ) did not alter the recognition of most wild-type 3′ splice sites . We also found that uaf-1 ( n4588 ) did not suppress the Unc phenotype caused by the unc-52 ( e669 ) mutation , which can be suppressed by mutations in the splicing factor genes smu-1 and smu-2 . We conclude that the uaf-1 ( n4588 ) mutation does not affect all cases in which alternative splicing is possible . We isolated four intragenic suppressors of the temperature-sensitive lethality caused by uaf-1 ( n4588 ) . Three of the suppressors ( n4588 n5120 , n4588 n5125 and n4588 n5127 ) carried both the original n4588 mutation and a second site mutation in uaf-1 . These three new uaf-1 mutations reduced the alternative splicing of unc-93 ( e1500 ) exon 9 to levels intermediate between those of uaf-1 ( n4588 ) and wild-type animals . This finding supports the hypothesis that UAF-1 is important in 3′ splice-site choice . The fourth intragenic suppressor , n5123 , affected the same site as the original n4588 mutation by generating a phenylalanine codon different from both the wild-type codon ( threonine ) and the codon generated by the n4588 mutation ( isoleucine ) . The uaf-1 ( n5123 ) allele behaves like the uaf-1 ( + ) allele , suggesting that this mutation restored the normal specificity of UAF-1 . The amino acids affected by these uaf-1 mutations ( n4588 , n5120 , n5123 , n5125 and n5127 ) are confined to a region between the U2AF small subunit-interacting domain [47] and the first RRM domain [26] , [45] ( Figure 5 ) . We postulate that this region of UAF-1 defines a domain of UAF-1 important for 3′ splice-site selection . In C . elegans , the first two nucleotides ( −2 to −1 ) of 3′ splice sites are more highly conserved than nucleotides −7 to −3 [29] , [64] , which affect the binding of the U2AF factors [29] . We sought to identify the nucleotides that affect the recognition of I8 and E9 by UAF-1 in vivo . First , we conclude that the location of a 3′ splice site affects its recognition . We found that the location of E9 was preferred to that of I8 by both wild-type and mutant UAF-1 when identical 3′ splice sites were present at the two locations ( Figure 6B , No . 2 and 3 ) . However , this positional effect was not absolute . When the high-affinity 3′ splice site TTTTcag was placed at either of the two locations , the site with the TTTTcag was preferred by both wild-type and mutant UAF-1 ( Figure 6 , No . 6 and No . 12 ) . That splicing using I8 and E9 ( both are likely weak 3′ splice sites , since there are fewer such sites in C . elegans introns than there are copies of the strong 3′ splice site sequence TTTTcag ( Table S4 ) ) was more affected by position than was splicing using the sequence TTTTcag suggests that weak 3′ splice sites might be preferably used for alternative splicing , and , strong 3′ splice sites such as TTTTcag might be generally used for constitutive splicing . If so , we might identify alternatively spliced genes by searching apparently weak 3′ splice sites and then performing RT–PCR analyses . That TTTTcag is strongly recognized by mutant UAF-1 is consistent with our finding that the uaf-1 ( n4588 ) mutation does not appear to affect the splicing of most other introns of unc-93 ( Figure 2B ) , which have a 3′ splice site identical or highly similar to TTTTcag ( data not shown ) . Second , we conclude that the nucleotides at −7 , −5 and −4 were more important than the nucleotide at −6 for wild-type UAF-1 to recognize the sequence of I8 ( Figure 6 , No . 7 to No . 10 ) . The nucleotides at −4 and −5 appear to be more important than that at −7 . That the nucleotide at −4 is more important than the nucleotide at −6 also appears to be the case for splicing at E9 by wild-type UAF-1 ( No . 15 and No . 16 , compared to No . 2 , Figure 6 ) , which indicates that nucleotide substitution at −4 ( No . 16 ) dramatically reduced splicing and nucleotide substitution at −6 ( No . 15 ) had a minimal effect at E9 in wild-type animals . Third , substituting individual non-T nucleotides with T in E9 improved its recognition by wild-type UAF-1 . The original I8 ( AATTcag ) and E9 ( ACTGcag ) are both rare 3′ splice sites compared to TTTTcag , which is found in about 26% of the approximate 40000 introns analyzed , and is the most commonly used 3′ splice site in C . elegans ( Table S4 ) [29] , [64] . I8 appears more frequently in introns than does E9 ( Table S4 ) , suggesting that E9 has a lower affinity for the wild-type UAF-1 than does I8 . That the wild-type UAF-1 rarely recognized E9 even in the more favored position ( Figure 6 , No . 1 ) is consistent with this notion . We found that substituting any E9 non-T nucleotide with T could increase the recognition of E9 in wild-type animals ( Figure 6 , No . 13 to 15 ) , and a T substitution at −6 and −4 had a much stronger effect than one at −7 . Fourth , we conclude that the T180I ( n4588 ) mutation caused UAF-1 to be more tolerant of a G nucleotide at −4 of E9 . In transgenes No . 11 and No . 16 , a G substitution at −4 of E9 dramatically reduced splicing at E9 in wild-type animals but did not or only moderately affected splicing at E9 in uaf-1 ( n4588 ) mutants ( Figure 6 ) . The splicing of transgenes No . 1 , No . 4 and No . 5 is consistent with this observation , implying that a G at −4 is more tolerated in uaf-1 ( n4588 ) animals than in wild-type animals . Based on these observations , we propose that the G nucleotide at position −4 of E9 is critical for its recognition by the mutant UAF-1 . In vivo studies have suggested functions for the U2AF large subunit beyond regulating pre-mRNA splicing . For example , Drosophila mutants with a temperature-sensitive U2AF large subunit are defective in the nucleus-to-cytoplasm export of intronless mRNAs at elevated temperatures [71] , suggesting that lack of U2AF large subunit function can affect mRNA export in addition to pre-mRNA splicing . Studies of SF1/BBP suggest that this splicing factor might not be essential for splicing in vitro or in vivo . Biochemical depletion of SF1/BBP in extracts from HeLa cells [72] and S . cerevisiae [73] or genetic depletion of SF1/BBP in extracts from S . cerevisiae [73] did not significantly affect splicing in vitro . Reducing SF1/BBP expression by RNAi in HeLa cells does not affect the splicing of several endogenous genes and a reporter gene [74] . That uaf-1 ( n4588 ) and sfa-1 ( n4562 ) suppressed the Unc phenotype of unc-93 ( e1500 ) but had different effects on the splicing of unc-93 ( e1500 ) mRNA at the cryptic 3′ splice site suggests that uaf-1 and sfa-1 could have both distinct and shared in vivo functions in C . elegans . Specifically , the splicing of some genes might be affected similarly by uaf-1 and sfa-1 , with other genes differentially affected . Alternatively , it is possible that uaf-1 ( n4588 ) has a stronger effect on the splicing of unc-93 ( e1500 ) exon 9 , while sfa-1 ( n4562 ) has a weaker effect not detected in the experiments we performed . The lack of conditionally viable mutants of the U2AF large subunit and SF1/BBP has impeded the analysis of the in vivo functions of these splicing factors in animals . The mutations we isolated affecting these two splicing factors should allow novel approaches for in vivo analyses of RNA splicing and of the functions of the U2AF large subunit and SF1/BBP in C . elegans . The transgene splicing system we developed provides an in vivo reporter assay for understanding the role of UAF-1 and possibly other splicing factors in regulating alternative 3′ splice site recognition .
C . elegans strains were grown at 20°C as described [55] , except where otherwise specified . N2 ( Bristol ) was the reference wild-type strain . CB4856 ( Hawaii ) was used for mapping mutations using single-nucleotide polymorphisms [75] . Mutations used in this study include: LGI: sup-11 ( n403 ) [54] . LGII: sup-9 ( n1550 , n2287 ) [34] , [43] , unc-52 ( e444 , e669 ) [55] , [60] . LGIII: vab-6 ( e697 ) and dpy-1 ( e1 ) [55] , uaf-1 ( n4588 , n5120 , n5123 , n5125 , n5127 , n5222Δ ) ( this study ) , unc-93 ( lr12 , n200 , n1912 , e1500 ) [36] , [37] , [44] , sup-18 ( n1014 ) [35] . LGIV: egl-23 ( n601sd ) [52] , dpy-4 ( e1166 ) [55] , sfa-1 ( n4562 , n5223Δ ) ( this study ) . LGX: twk-18 ( e1913sd ) [53] , unc-58 ( e665sd ) [51] and sup-10 ( n983 , n3564 ) [34] , [35] . The translocation nT1 IV;V with the dominant gfp marker qIs51 [76] was used to balance the sfa-1 locus , and the translocation sC1 ( s2023 ) [dpy-1 ( s2170 ) ] ( A . Rose , D . Baillie and D . Riddle , the Genetic Toolkit project ) was used to balance the uaf-1 ( n5222Δ ) locus . Synchronized L4 unc-93 ( e1500 ) animals ( P0 ) were mutagenized with EMS ( ethyl methanesulfonate ) as described [55] . F1 progeny from these animals were picked to single wells of 24-well culture plates with OP50 bacteria grown on NGM agar . F2 progeny were observed using a dissecting microscope to identify animals with improved locomotion . From ∼10 , 000 F1 clones screened , 100 independent suppressed strains were isolated . 97 of the isolates , including two weak recessive sterile suppressors and 95 recessive fertile suppressors , were kept as frozen stocks for possible later study . Three stronger suppressors that caused or were closely linked to mutations that caused sterility ( n4562 ) or ts-lethality ( n4588 and n4564 ) were chosen for further analysis . The analysis of n4564 is ongoing . We mapped n4588 to the left of dpy-1 on LGIII based on the suppression of unc-93 ( e1500 ) using standard methods . As the suppressor activity and ts-lethality were very closely linked , e . g . , more than 500 n4588 unc-93 ( e1500 ) /+ unc-93 ( e1500 ) individuals failed to segregate Sup non-Let progeny , we then followed the phenotype of ts-lethality to further map n4588 . We mapped n4588 to the right of nucleotide 186577 on BE0003N10 ( cosmid BE0003N10 sequences refer to nucleotides of accession no . AC092690 ) using 10 Vab recombinants recovered after crossing vab-6 ( e697 ) n4588 hemaphrodites with males of the Hawaiian strain CB4856 [75] and to the left of nucleotide 13164 on Y92C3A ( accession no . AC024874 ) using 37 Dpy recombinants recovered after crossing n4588 dpy-1 ( e1 ) hemaphrodites with males from the Hawaiian strain CB4856 . We determined the coding sequences of four genes in this interval , uaf-1 , rab-18 , kbp-4 and par-2 , and identified a missense mutation in the third exon of the uaf-1a isoform . As the suppressor activity and sterility of n4562 were very closely linked , e . g . , over 500 unc-93 ( e1500 ) ; n4562/+ individuals failed to segregate Sup non-Ste progeny , we followed the sterility phenotype to map n4562 to the right of dpy-4 on LG IV using standard methods . We next mapped n4562 to the right of nucleotide 37163 on Y43D4A ( accession no . AL132846 ) using 234 Dpy recombinants recovered after crossing dpy-4 ( e1166 ) n4562 with males of the Hawaiian strain CB4856 . The sequences of coding exons of sfa-1 and uaf-2 , both located in this region , were determined , and a Cys458Opal ( TGT-to-TGA ) mutation was identified in sfa-1 . Genomic DNA pools from EMS-mutagenized animals were screened for deletions using PCR as described [77] . Deletion mutant animals were isolated from frozen stocks and backcrossed to the wild type at least three times . uaf-1 ( n5222Δ ) removes nucleotides 9786 to 11082 of YAC Y92C3B . sfa-1 ( n5223Δ ) removes nucleotides 207818 to 208925 of YAC Y116A8C . Total RNA was prepared using Trizol according to the manufacture's instructions ( Invitrogen ) , treated with RNase-Free DNase I ( New England Biolabs ) and followed by incubation at 75°C for 10 minutes to inactivate DNase I . First-strand cDNA was synthesized with random hexamer primers ( New England Biolabs ) using the Superscript II or III First-Strand Synthesis Kit ( Invitrogen ) . Quantitative RT–PCR was performed using either a DNA Engine Opticon System ( MJ Research ) or a Mastercycler realplex system ( Eppendorf ) . For the SYBR green-based assay ( DNA Engine Opticon System ) , each 30 µl PCR reaction contained 1 to 10 ng RT template , 0 . 5 mM PCR primers and 15 µl 2× SYBR Green PCR Master Mix ( Applied Biosystems ) . Three independent samples of synchronized wild-type ( N2 ) and uaf-1 ( n4588 ) L1 animals were prepared , and levels of control genes ( rpl-26 , gpd-2 , act-1 ) and tested genes ( myo-3 , unc-93 , sup-9 , sup-10 , sup-11 , sup-18 ) were quantified from each biological replicate . For the Taqman probe-based assay ( Mastercycler realplex system ) , the probes ( Figure 4A ) were labeled at their 5′-ends with 6-carboxyfluorescein ( FAM ) and at their 3′-ends with Black Hole Quencher ( BHQ-1 ) ( Integrated DNA Technologies ) . Two independent samples of each genotype of animals of mixed stages were prepared , and levels of rpl-26 and unc-93 wild-type and alternatively spliced transcripts were quantified from each biological replicate . For RNAi-treated animals ( see below ) , one sample for each assay was quantified . PCR primers and Taqman probes are listed in Table S5 . Synchronized uaf-1 ( n4588 ) animals ( P0 ) at the L4 larval stage grown at 15°C were mutagenized with EMS as described [55] . These P0 animals were allowed to grow to young adults at 15°C in a mixed population and bleached , and F1 progeny were synchronized at the early L1 stage by starvation in S medium [78] . The F1 animals were placed on 50 Petri plates ( ∼1000 animals/plate ) with NGM agar seeded with OP50 and permitted to grow to young adults at 15°C and then moved to 25°C . After six days at 25°C , the animals were grown at 20°C for six days and examined each day for the presence of living F2 animals . About 50 , 000 F1 progeny from a mixture of more than 10 , 000 P0 animals were screened . From the screen , we recovered 13 surviving F2 animals from 13 different F1 plates . Six of the 13 suppressors were intragenic suppressors representing four different mutations: one was n5120 , one was n5123 , three were identical to n5125 and one was n5127 . It is possible that the three isolates containing the n5125 mutation were derived from the same P0 animal , because all of the P0 animals were in a mixed population when bleached to release eggs . The other seven suppressors were extragenic mutations , i . e . , n4588/+; sup/+ animals segregated n4588-like progeny . Some or all of the extragenic isolates could have been derived from the same P0 animal . Young adult animals ( wild-type or unc-93 ( e1500 ) ) were fed HT115 ( DE3 ) bacteria containing plasmids directing the expression of dsRNAs targeting either uaf-1 or sfa-1 on NGM plates with 1 mM IPTG and 0 . 1 mg/ml Ampicillin [79] . Surviving F1 progeny of the unc-93 ( e1500 ) animals ( escapers ) were examined for suppression of locomotion defects . Animals were washed from plates , rinsed three times with H2O , and resuspended in Trizol ( Invitrogen ) for preparation of total RNA or in 2× protein loading buffer ( see Western blots , below ) for SDS-PAGE analysis . We generated the DNA construct expressing dsRNA targeting uaf-1 ( see below ) . The bacterial strain expressing dsRNA targeting sfa-1 was obtained from a whole-genome RNAi library [80] , and the sequences of plasmids from single colonies of the strain were determined to confirm the presence of sfa-1 coding sequences . L4 animals were picked 16–24 hrs before assaying and were grown at 20°C . Young adults were then individually picked to Petri plates containing NGM agar seeded with OP50 , and bodybends were counted for 30 seconds using a dissecting microscope as described [81] . The rubberband phenotype was scored as described [43] . Animals were washed from plates with H2O , rinsed three times with H2O , resuspended in one volume of 2× SDS loading buffer ( 100 mM Tris . Cl ( pH 6 . 8 ) , 200 mM DTT , 4% SDS , 0 . 2% bromophenol blue , 20% glycerol ) , boiled for 5 minutes , and samples were then loaded onto 8% polyacrylamide gels containing SDS . Protein samples were transferred from polyacrylamide gels to Immobilon-P Transfer Membranes ( Millipore ) . Primary and secondary antibody incubations were performed with 5% non-fat milk in TBST ( 25 mM Tris-HCl , PH 8 . 0; 125 mM NaCl; 0 . 1% Tween-20 ) at room temperature for one hour each . Signals were visualized using Chemiluminescence Reagent Plus ( PerkinElmer Life Sciences ) and X-ray film ( BioMax XAR film , Kodak ) . Primary antibody was rabbit anti-UAF-1 ( 1∶20000 ) [45] . Secondary antibody was HRP-conjugated goat anti-rabbit ( 1∶3000 ) ( BioRad ) . To rescue the suppression of the Unc phenotype of unc-93 ( e1500 ) by uaf-1 ( n4588 ) or sfa-1 ( n4562 ) , uaf-1 and sfa-1 cDNAs were subcloned to vector pPD93 . 97 using BamHI and AgeI restrictions sites . uaf-1b cDNA was amplified with PCR using uaf-1a cDNA as template and subcloned to pPD93 . 97 using BamHI and AgeI restrictions sites . uaf-1a cDNA was subcloned to pPD49 . 83 ( for heat-shock induced expression of uaf-1a cDNA ) using BamHI and SacI restriction sites . An XhoI/SpeI fragment of uaf-1a cDNA subcloned in a pGEM-TA easy vector ( Promega ) was subcloned to pPD129 . 36 ( for the uaf-1 RNAi construct ) using XhoI and NheI sites . To test whether the truncated unc-93 ( Δ ) cDNA caused by altered splicing of the unc-93 ( e1500 ) transcript in uaf-1 ( n4588 ) mutants encodes a functional UNC-93 protein , unc-93 cDNA and unc-93 ( Δ ) cDNA were subcloned to pPD93 . 97 using BamHI ( blunt ) and AgeI sites . To test whether uaf-1 ( n4588 ) or sfa-1 ( n4562 ) mutations could suppress the Unc phenotype caused by ectopic expression of the unc-93 ( e1500 ) cDNA , unc-93 ( e1500 ) cDNA was subcloned to pPD93 . 97 using BamHI ( blunt ) and AgeI sites . To examine the effect of nucleotide substitutions on the recognition of the intron 8 endogenous 3′ splice site and the exon 9 cryptic 3′ splice site , we fused the genomic sequence between exon 8 and exon 10 of unc-93 ( e1500 ) in-frame with the GFP gene of pPD93 . 97 using BamHI and AgeI sites . We replaced the myo-3 promoter of pPD93 . 97 with a 2 kb promoter of unc-93 using PmlI and BamHI sites . Point mutations in uaf-1 ( stop codons ) , the unc-93 ( e1500 mutation ) or mutated transgenes were introduced using QuickChange II or III Site-Directed Mutagenesis Kit ( Stratagene ) with primers containing corresponding mutations . PCR was performed using Eppendorf Cyclers , and DNA products were resolved using agarose gels . DNA sequence determination was performed with an ABI Prism 3100 Genetic Analyzer . PCR primers are listed in Table S5 . Germline transgene experiments were performed as described [82] . Transgene mixtures generally contained 20 µg/ml 1 kb DNA ladder ( Invitrogen ) , 20 µg/ml Arabidopsis genomic DNA and 10 µg/ml of the transgene of interest . When the transgene did not cause the expression of a GFP fusion protein , 10 µg/ml pPD95 . 86-GFP plasmid ( expressing GFP in body-wall muscles ) or 5 µg/ml pmyo-2dsRED ( expressing RFP in pharynx ) was added to the injection mixture as a visible fluorescence marker to identify animals carrying the transgene . We downloaded approximate 40 , 000 unique intronic sequences from WormMart ( WormBase Release 195 ) and processed the sequences using BBEdit and MS Excel softwares . Identical 3′ splice sites ( positions −7 to −1 ) were grouped and counted .
|
Eukaryotic genes contain intervening intronic sequences that must be removed from pre-mRNA transcripts by RNA splicing to generate functional messenger RNAs . While studying genes that encode and control a presumptive muscle potassium channel complex in the nematode Caenorhabditis elegans , we found that mutations in two splicing factors , the U2AF large subunit and SF1/BBP suppress the rubberband Unc phenotype caused by a rare missense mutation in the gene unc-93 . Mutations affecting the U2AF large subunit caused the recognition of a cryptic 3′ splice site generated by the unc-93 mutation , providing in vivo evidence that the U2AF large subunit can affect splice-site selection . By contrast , an SF1/BBP mutation that suppressed the rubberband Unc phenotype did not cause splicing using this cryptic 3′ splice site . Our genetic studies identified a region of the U2AF large subunit important for its effect on 3′ splice-site choice . Our mutagenesis analysis of in vivo transgene splicing identified a positional effect on weak 3′ splice site selection and nucleotides of the endogenous 3′ splice site important for recognition . The system we have defined should facilitate future in vivo analyses of pre–mRNA splicing .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/animal",
"genetics",
"genetics",
"and",
"genomics/gene",
"discovery",
"genetics",
"and",
"genomics/gene",
"function",
"genetics",
"and",
"genomics/gene",
"expression"
] |
2009
|
Mutations in the Caenorhabditis elegans U2AF Large Subunit UAF-1 Alter the Choice of a 3′ Splice Site In Vivo
|
The elderly are particularly susceptible to influenza A virus infections , with increased occurrence , disease severity and reduced vaccine efficacy attributed to declining immunity . Experimentally , the age-dependent decline in influenza-specific CD8+ T cell responsiveness reflects both functional compromise and the emergence of ‘repertoire holes’ arising from the loss of low frequency clonotypes . In this study , we asked whether early priming limits the time-related attrition of immune competence . Though primary responses in aged mice were compromised , animals vaccinated at 6 weeks then challenged >20 months later had T-cell responses that were normal in magnitude . Both functional quality and the persistence of ‘preferred’ TCR clonotypes that expand in a characteristic immunodominance hierarchy were maintained following early priming . Similar to the early priming , vaccination at 22 months followed by challenge retained a response magnitude equivalent to young mice . However , late priming resulted in reduced TCRβ diversity in comparison with vaccination earlier in life . Thus , early priming was critical to maintaining individual and population-wide TCRβ diversity . In summary , early exposure leads to the long-term maintenance of memory T cells and thus preserves optimal , influenza-specific CD8+ T-cell responsiveness and protects against the age-related attrition of naïve T-cell precursors . Our study supports development of vaccines that prime CD8+ T-cells early in life to elicit the broadest possible spectrum of CD8+ T-cell memory and preserve the magnitude , functionality and TCR usage of responding populations . In addition , our study provides the most comprehensive analysis of the aged ( primary , secondary primed-early and secondary primed-late ) TCR repertoires published to date .
The elderly population is particularly susceptible to novel infections , especially the annual , seasonal epidemics caused by influenza A viruses [1] , [2] , with increased occurrence , severity of infection and reduced vaccine efficacy being attributed to age-related decline in immune capacity [3]–[6] . The ageing effect on the immune system is considered to be multifactorial , arising from the diminished thymic export of naïve precursors due to thymic involution [7] , [8] , the impaired recruitment [9] , [10] of naïve CD8+ T cell precursors and the replicative senescence of memory cells [11]–[14] . Ageing can also be associated with abnormal cellular functions such as distorted cytokine secretion ( IL-2 , IL-4 and IFN-γ ) profiles [15]–[17] , decreased granzyme B production [18] , [19] and reduced proliferative capacity due to the loss of CD28 expression [20] . Perturbations in the naïve TCR repertoire have also been reported , with abnormal TCR spectratype ( CDR3β length ) patterns in aged mice reflecting the massive , antigen-independent expansion , of a few clonotypes [21] . Naïve T cell attrition has also been inferred from observed reductions in the diversity of antigen-specific TCR repertoires in aged mice [5] , [22] . Previous mouse studies have established that ageing can be associated with diminished CD8+ T cell efficacy and delayed influenza virus clearance [23]–[25] . Recent evidence has further shown that the selective loss of primary , influenza-specific CD8+ T cell responsiveness in older mice is characterized by a narrowing in the spectrum of TCR usage and is seen predominantly for low frequency populations , with this effect being best characterized for the prominent DbNP366+CD8+ T cell set [5] , [26] . Overall , the findings so far suggest that the capacity to respond effectively to new influenza infections in aged mice requires the maintenance of a diverse pool of functional peripheral T cells . As CD8+ T cells tend to be specific for peptides derived from more conserved proteins that are internal to the virus , priming effective CD8+ T cell memory has obvious potential for countering newly emerged seasonal or pandemic influenza strains . The importance of long-lived , antigen-specific memory CD8+ T cells capable of rapid recall following the secondary infection has been well documented for the respiratory viruses in mice [27] , [28] and humans [29] , [30] . Such long-term maintenance of memory T cells leading to enhanced secondary response forms the basis for vaccination strategies based on priming CD8+ T cell memory to promote early virus clearance and decreased morbidity . The question is though , whether such CD8+ T cell memory can be effectively recalled in the elderly . A recent study [6] suggested that infecting mice with LCMV or influenza at an extreme age ( 18–20 months ) leads to defective CD8+ T cell memory and diminished recall responses following virus challenge . What happens , though , if CD8+ T cell memory is established when the mice are young ? The analysis reported here compares the CD8+ T cell response profiles for young ( <3 months ) and aged ( 22 month ) mice , with the latter cohort being first exposed to immunogenic influenza epitopes early or late in life . The results suggest that designing influenza vaccines which promote as broad as possible spectrum of CD8+ T cell memory in adolescence could be beneficial , even if such benefit emerges long after the subject has first been given the protective immunogen .
The comparison of the HK-induced CD8+ T cell responses utilized young or old mice that were either immunologically naïve ( primary , 10; Figure 1A ) or had been primed at 2 months of age with the PR8 virus and challenged 20 months later ( secondary , 20; Figure 1B ) . Immunodominant and subdominant CD8+ T cell responses were measured in the spleen ( Figure 1CD ) by the ex vivo IFN-γ ICS assay . Following 10 challenge , the size of the low precursor frequency DbNP366+CD8+ set in the spleen ( Figure 1C ) was markedly diminished in the aged animals relative to the young controls as previously observed [5] , [31] . Conversely , any age-related effects on CD8+ T cell numbers were not significant for DbPA224 ( Figure 1C ) . The unaffected DbPA224+CD8+ T cell responses are intriguing , as the naïve CD8+ T cell frequencies [36] found for DbPA224–specific T cells in young mice are significantly higher than those detected for DbNP366 ( ≥72 versus <40 per individual , respectively ) , suggesting that a larger naïve CD8+ T cell pool size minimizes the extent of age-related attrition and , as a consequence , the effect on primary CD8+ T cell response magnitude ( Figure 1C ) . Reduced magnitude of the immunodominant DbNP366-specific CD8+ T cell response that was detected for the primary , influenza-specific CD8+ T cell response in older mice ( Figure 1CD ) , was not sustained following secondary HK challenge of mice that had been primed early with the PR8 virus ( Figure 1D ) . The numbers of DbPA224CD8+ T cells were significantly diminished across combined experiments but , otherwise , the recall responses for memory T cell pools in young or old mice primed at <2 months ( at least 20 months previously ) were not obviously different , emphasizing the durability of virus-specific CD8+ T cell memory [37] . In particular , the overdominance of the DbNP366-specific set that is characteristic of the secondary response to these viruses [38] was still apparent in the aged mice ( Figure 1D ) . The beneficial effect of the early CD8+ priming on the immunodominant low-precursor responses like the DbNP366-specific population following influenza infection at the extreme age was most striking when the relative contributions of particular antigen-specific CD8+ T cells were analysed based on total cell numbers ( Figure 2 , calculations based on Figure 1 for immunodominant DbNP366+CD8+ and DbPA224+CD8+ pools , and data not shown for subdominant DbPB1703+CD8+ and KbPB1-F262+CD8+ populations ) . In the aged mice , the primary CD8+ T cell responses showed a shift in the typical immunodominance hierarchy ( Figure 2B ) , with the contribution of the immunodominant DbNP366+CD8+ population being significantly lower in the aged mice ( 9 . 4±3 . 6% ) in comparison to young animals ( 43 . 4±15%; p<0 . 01; Figure 2A ) . The differential immunodominance hierarchy resulted mainly from significantly increased contribution of KbPB1703+CD8+ T cells ( Figure 2 ) . This led to major modifications in response hierarchy following primary influenza virus infection of aged mice KbPB1703>DbPA224>DbPB1-F262>DbNP366 , with the comparable profile for young mice being DbNP366>DbPA224 = KbPB1703≫DbPB1-F262 . Conversely , recall of CD8+ T cells primed at a young age preserved the overall contribution of T cell specificities and retained the immunodominance hierarchy in aged mice primed early at 6 weeks ( Figure 2D ) , reflecting the characteristic immunodominance hierarchy in young controls ( Figure 2C ) . These findings show clearly that priming the CD8+ T cell compartment at an early age leads to subsequent preservation of CD8+ T cell numbers and immunodominance hierarchies for influenza infection in the elderly . One measure of CD8+ T cell function is the capacity to produce multiple cytokines simultaneously [39] following in vitro stimulation with peptide in the standard , 5 h ICS assay . For the primary DbPA224+CD8+ T cell population that remained relatively constant in numbers with age ( Figure 1C ) , the frequencies of double ( IFN-γ/TNF-α ) and triple-producers ( IFN-γ/TNF-α/IL-2 ) were significantly lower in comparison with the young mice ( Figure 3AB ) . Furthermore , taking mean fluorescence intensity ( MFI ) , which represents the intensity and therefore amount of cytokine production , it also seems that the DbPA224+CD8+ population tended to produce less TNF-α , though this diminution effect was not apparent for either IFN-γ or IL-2 ( Figure 4A ) . Taking the prevalence and MFI data together ( Figure 3 and Figure 4 ) , there appears to be a general decrease in cytokine polyfunctionality for the primary DbPA224+CD8+ response . Conversely , analysis of aged mice primed early showed that functional characteristics appear to be locked-in early and maintained in the long-term for memory T cell populations ( Figure 3CDEF ) . Can we detect other evidence of enduring functional change ? Given that the influenza-specific CD8+ T cells generated following primary infection of aged mice were either of suboptimal functional quality ( DbPA224+CD8+; Figure 3 , Figure 4 ) or reduced in number ( DbNP366+CD8+; Figure 1 ) , the further question was whether there was any effect on cell surface activation phenotype [34] , [40]–[42] . Comparison of phenotypic markers associated with activation , trafficking and memory potential: CD62L vs . IL-7Rα ( CD127 ) , CD27 vs . CD43 , and IL-7Rα vs . KLRG-1 for the DbPA224+CD8+ and DbNP366+CD8+ sets ( Figure S2 ) showed that , with the exception of a decrease in the relative prevalence of the less activated CD27loCD43loCD8+ DbPA224+ cells in the older mice ( Figure S2AD ) , there were no significant differences in phenotype with age . Previous studies have found a significant skewing in TCR Vβ usage ( mAb staining ) and CDR3β length ( spectratyping ) for CD8+ T cell responses developed from naïve and memory populations by the infection of aged versus young mice [13] , [21] , [43] . Thus , we looked more closely at the expansion and maintenance of responding T cell clonotypes [44] , [45] . As our earlier analysis of influenza-specific CD8+ TCR clonotype diversity has focused on the prominent Vβ8 . 3+DbNP366+ [44] , [46] and Vβ7+DbPA224+ sets [47] , we first assessed the Vβ mAb-staining profiles to determine whether these characteristic TCRs were also selected following primary or secondary challenge of aged mice . Indeed for both DbNP366+ CD8+ and DbPA224+CD8+ T cell responses , the characteristic Vβ8 . 3 and Vβ7 usage was observed ( Figure 5 ) , though additional Vβ6 , Vβ7 and Vβ9 biases were variously detected in individual , older mice for the primary DbNP366+CD8+ population ( Figure 5C ) , possibly due to the recruitment of low frequency alternate DbNP366-specific CD8+ T cells . Despite the presence of a prominent Vβ8 . 1/8 . 2+DbNP366+ set in one of the early-primed , secondarily-challenged at 22 month mice , the bias was generally to Vβ8 . 3 suggesting that the characteristic DbNP366+CD8+ TCRβ usage profile is retained in the persistent memory population . The DbPA224+ set was characterised across groups by Vβ7 TCR usage ( Figure 5BDF ) , which was more consistent than the DbNP366+Vβ8 . 3 usage , possibly reflecting the higher number of precursors with Vβ7 surviving within the 22 month old mice . Since priming at a young age led to the typical magnitude and quality of influenza-specific CD8+ T cell responses following viral infection in the aged mice , we asked whether priming the mice via a non-replicative route ( i . p . priming with 1 . 5×107 pfu of PR8 ) at extreme age ( 22 months ) would be also beneficial for the subsequent influenza virus infection . Since the reduced primary DbNP366+CD8+ T cell responses in aged mice has been attributed to the lower naïve precursors in young mice [5] , this experiment would determine whether old naive mice could be primed at an extreme age ( at 22 months ) and subsequently challenged i . n . with 1×104 pfu of the HK influenza strain ( at ∼24 months; Figure 6A ) to mount an effective recall response after the attrition had occurred . Surprisingly , despite the reduced primary DbNP366+CD8+ T cell responses ( Figure 1C ) and lower magnitude of secondary DbPA224+CD8+ sets ( Figure 1E ) in the spleens of aged animals , the recall of influenza-specific CD8+ T cells was robust and equivalent in magnitude to the young controls ( Figure 6 ) . The numbers of both immunodominant DbNP366+CD8+ and DbPA224+CD8+ populations were normal ( Figure 6B ) . This resulted in the maintained contribution of each of the T cell specificities to influenza-specific responses ( Figure 6E ) . Conversely , the polyfunctionality of those secondary CD8+ T cell populations in mice primed at the extreme age did not always resemble effectiveness of influenza-specific CD8+ T cells recruited in young individuals ( Figure 6C ) . Perturbations in the TCR usage with extreme age were evident macroscopically in the TCR Vβ usage for DbPA224+CD8+ ( Figure 6G ) and especially the DbNP366+CD8+ ( Figure 6F ) responses , with the usage of alternate Vβ8 . 1/8 . 2 for DbPA224+CD8+ , and Vβ7 and Vβ8 . 1/8 . 2 for DbNP366+CD8+ populations . The characteristic Vβ8 . 3 usage for DbNP366+CD8+ was only dominant in 1 of 4 mice ( Figure 6F ) , reflecting narrowing of the naïve DbNP366+CD8+ set with extreme age that initially limited the primary response ( Figure 1C ) and/or the clonal expansions characteristic for the aged animals as previously reported [13] , [14] . A substantial body of work from previous studies has defined the young B6 CDR3β TCR usage at high resolution [44] , [47] , therefore using these data sets from young mice we were able to compare the spectrum of clonotype prevalence in aged mice using single-cell RT-PCR and sequencing of the CDR3β region to determine the spectrum of TCRβ diversity . Analysis of 1489 CDR3β sequences for primary and secondary ( primed-young and primed-old ) responses from the 22 month old mice ( Tables 1 and 2 ) showed that the dominant Jβ regions and CDR3β loop lengths in the aged animals ( Tables S1 , S2 , S3 , S4 , S5 , S6 ) were comparable to those found early in life ( Figures 7 and 8 for comparison with young animals ) . However , more inter-individual variation in the primary responses was observed in the older group ( Figure S3 ) . While >83% of each of the TCRβ repertoires involved in the primary responses to DbNP366 in young mice utilized Jβ2 . 2 and a CDR3 length of 9 amino acids ( aa ) , this profile was substantially diminished to <57% of the TCRβ repertoire for 2/7 aged mice . Similarly , Jβ1 . 1 , Jβ1 . 5 , and Jβ2 . 6 collectively dominated the primary DbPA224+CD8+ responses for 7/7 young mice , while Jβ2 . 1 and Jβ2 . 3 emerged strongly ( >55% each ) for 2 of the older mice . While the primary DbPA224+CD8+ repertoires in individual young mice mostly featured diverse CDR3 lengths of 5 , 6 , and 7 aa , >94% of the primary DbPA224+CD8+ T cell repertoires in two of the aged mice could be attributed to one particular CDR3 length ( i . e . 6 aa in one mouse and 7 aa in the other mouse ) . Age-associated changes in TCRβ repertoire usage were investigated for the DbNP366+CD8+ and DbPA224+CD8+ populations by sequencing individual CDR3β TCR signatures ( Tables 1 and 2 , Tables S1 , S2 , S3 , S4 , S5 , S6 ) and the extent of TCRβ repertoire diversity was then assessed using both the number of different aa-defined clonotypes and Simpson's diversity index , which accounts for the clonal dominance hierarchy . These measures of diversity were estimated for a standard 22 TCRβ sequences per epitope per mouse to adjust for differences in total number of sequences obtained per mouse [48] . The primary DbPA224+CD8+ TCRâ repertoires were found to be significantly less diverse in aged versus young mice , with a lower number of clonotypes per individual ( median: 8 vs . 14 , p = 0 . 005; Figure 7C ) and a decreased Simpson's diversity index ( median: 0 . 72 vs 0 . 94 , p = 0 . 007; Figure 7G ) , despite there being no significant change in the DbPA224-specific CD8+ T cell response magnitude ( Figure 1A ) . Some age-related contraction in the number of different DbPA224+CD8+ TCRâ clonotypes was also found following secondary infection ( early priming ) ( median: 10 vs . 12 , p = 0 . 007; Figure 7D ) , though the difference was not as large as in the primary response , largely due to the increased median diversity for the recall response in older mice . Interestingly , when mice were primed at 22 months of age and then challenged ( primed old ) , similar results were obtained as early priming , however there appeared to be substantial increase in the similarity between some pairs of mice ( Figure 7P ) . Surprisingly , the reduced diversity seen in the DbPA224+ CD8+ primary response ( Figure 7CG ) , from which the late priming response is derived , was not carried over to the primed-old recall TCRβ repertoire ( Figure 7DH ) . This suggests that priming plays a positive role in preserving a broader spectrum of clonotype availability within the inherently diverse DbPA224+CD8+ T cell repertoire , due to enhanced response magnitude . In contrast , despite the greatly diminished magnitude of the primary DbNP366+CD8+ T cell response in older mice ( Figure 1A ) , the extent of TCRβ repertoire diversity analysed at the aa level was not significantly different for young and old mice ( Figure 7AE , Table 1 , Table S1 ) . The public DbNP366+Vβ8 . 3 clonotypes can be encoded by up to 10 different nucleotide ( n . t . ) sequences each , with as many as 4 distinct n . t . -defined variants being present in an individual young mouse [44] . Following primary exposure of aged animals or when mice were primed late , the three main public Vβ8 . 3+ DbNP366+CD8+ clonotypes: SGGANTGQL , SGGGNTGQL , SGGSNTGQL [44] were encoded by a total of 10 and 9 distinct n . t . sequences respectively ( Tables S1 and S5 ) , in contrast to the 16 different clonotypes detected in the secondary-infected ( primed early ) , aged mice ( Table S3 ) . As a consequence , priming early or late prior to challenge preserved a mean of 2 . 9±1 . 1 and 3 . 0±0 . 7 n . t . -defined public clonotypes in comparison to the 1 . 7±1 . 1 public n . t . sequences detected following infection of old , naïve mice . Such reduced availability of n . t . -defined public sequences in the primary aged mice resulted in a loss of one of the major public clonotypes SGGGNTGQL ( Figure 8A ) in all 7 animals tested following primary virus challenge ( Table S1 ) . This was associated in turn with a markedly greater contribution of the SGGANTGQL clonotypes ( 57% versus 23% ) in primarily-infected aged animals in comparison to those that were secondary challenged ( Figure 8B ) . It is interesting to note that previously the SGGANTGQL clonotype has been associated with low pMHC avidity [49] . Thus , although the DbNP366+CD8+ repertoire is dominated by public TCRs encoded by multiple distinct n . t . sequences , due to codon redundancy the selective , age-related exclusion of one n . t . -defined clonotype does not necessarily equate to the disappearance of any given aa clonotype from the naïve pool . However , it is still possible that the prominent TCR signatures ( like SGGGNTGQL ) can be lost or significantly decreased with ageing . Significantly higher inter-individual similarity of DbNP366TCRβ repertoires was seen in the recall response of aged mice that were primed old compared with aged mice primed young ( Figure 7N ) . The proportion of individual mouse TCRβ repertoires comprised of shared clonotypes was consistently high across age and priming groups ( Figure 7IJ ) . Furthermore , there was higher inter-individual similarity during the secondary DbNP366+CD8+ responses in aged mice primed old ( Figure 7N ) was largely due to the SGGSNTGQL clonotype that was dominant in 4/5 mice , and therefore dominated the primed aged secondary response ( Figure 8B , Table S5 ) . The lesser prevalence and dominance of this SGGSNTGQL clonotype in the aged primary response ( Figure 8A , Table S1 ) could be related to the avidity of individual clonotypes recruited during recall and preferential homeostatic proliferation , which is reminiscent of the lower avidity SGGANTGQL clonotype dominating the primary aged response above . Overall , there was a trend towards lower TCR diversity in the DbNP366+CD8+ response to secondary infection in aged mice , regardless of age of priming , compared with young mice . However , due to the extreme dominance of SGGSNTGQL ( Figure 8B ) , and the significantly greater inter-individual similarity ( Figure 7N ) in aged mice primed late versus early , the timing of priming has a narrowing effect on the population-wide Vβ8 . 3+ DbNP366+CD8+ TCRβ repertoire . Thus , encountering an immunogenic epitope leads to a relative preservation of TCRβ diversity at the n . t . level ( the ‘actual’ clonotypes ) , even if repertoire diversity at the aa level appears unchanged . Priming also prevents the attrition of dominant public TCRs with age and mediates their recruitment into the CD8+ T cell effector pool in the elderly . The results of the present study also confirm our previous longitudinal analysis of DbNP366+CD8+ responses [44] and differential clonotype hierarchy usage in the primary young and secondary young mice ( Figures 8A and 8B ) . While SGGGNTGQL is a preferential clonotype after the i . p . priming ( as well as after the primary i . n . infection ) , the hierarchy changes after re-challenge , with SGGANTGQL and SGGSNTGQL clonotypes dominating the secondary response .
The present analysis establishes the importance of priming the CD8+ T cell compartment early in life in order to preserve CD8+ T cell numbers , functional quality and preferential profiles of TCR usage for influenza-specific CD8+ effector T cell responses in the elderly . In contrast , primary CD8+ T cell responses in aged animals tended to show alterations in the typical CD8+ T cell immunodominance hierarchy , with T cell responses to some epitopes being reduced in magnitude , a decrease in the capacity to make multiple cytokines , and changes in the extent of TCRβ repertoire diversity as a consequence of the diminished availability of naïve clonotypes . These effects were minimal for the recall responses generated from memory T cell populations that were generated early , and then recalled by virus challenge more than 18 months later . Overall , the results emphasize both the durability and constancy of immune memory . The response hierarchy following primary influenza virus infection of aged mice was KbPB1703>DbPA224>DbPB1-F262>DbNP366 , with the comparable profile for young mice being DbNP366>DbPA224 = KbPB1703≫PB1-F262 . Typically subdominant epitopes accounted for 59% of the response in aged naïve mice challenged with virus compared with a 34% ( Figure 2A ) contribution in the young . Thus , immunodominance hierarchies may be relative to age , an idea that is clearly more relevant to the situation in long-lived humans than in mice . In contrast , the typical hierarchy [36] was maintained for both young and old mice that were primed early , with a relative contribution by subdominant epitopes of 10% and 12% ( Figure 2D ) respectively . Whereas when mice were primed at an extreme age subdominant epitopes contributed 26% of the anti-influenza CD8+ T cell response and , therefore , the immundodominance hierarchy was perturbed ( Figure 6E ) , to a lesser extent than the primary response in aged mice . The difference in naïve precursor frequency for the DbNP366+CD8+ and DbPA224+CD8+ T cell sets is only two-fold ( 36 vs 79 naïve precursors , respectively ) [36] , yet any age-related diminution in magnitude for the primary response to DbPA224 was less apparent , suggesting that expanding CD8+ T cell precursors prevalence by an estimate of 2–4 fold may protect immune capacity in the long term . As all the naïve , endogenous and non-transgenic DbNP366+CD8+ , and DbPA224+CD8+ T cells are recruited into the primary immune response [36] , there would be no naïve precursors left to mount a primary CD8+ T cell responses after re-challenge for these three sets of influenza-specific CD8+ T cell populations , unless new precursors had emerged subsequently from the thymus . With age , the relative loss in magnitude for the normally prominent DbNP366-specific response can be most likely attributed to the loss of naïve precursors with time as previously suggested [5] . Despite multiple attempts to repeat the naïve CD8+ T cell analysis for aged ( 22 mo ) B6 mice , we were unable to recover viable tetramer+CD8+ populations ( data not shown ) following the application of the rigorous magnetic separation procedure that is required to recover very small numbers of antigen-specific cells from the total , peripheral CD8+ T cell pool [36] , [50] in the aged mice comparing to normal precursor frequencies in the young controls . This could reflect diminished structural integrity due , for instance , to senescence-associated changes in membrane lipids [51] . Thus , at this time we were unable to compare naïve influenza-specific CD8+ T cell precursor frequencies of aged mice to established precursor frequencies in the young controls , but rather infer results from the immunodominance hierarchy of the aged primary responses . The comparable sizes and immunodominance hierarchies of influenza-specific CD8+ T cell responses in young and elderly following recall reflects the stability of long term-memory pools , which has also been evidenced by earlier data showing stable memory numbers for both DbNP366+CD8+ and DbPA224+CD8+ T cells until at least d575 after primary infection [32] . Together with the present analysis , evidence for the preservation of Vaccinia virus-specific memory populations in humans primed more than 20 years previously [52] reinforces the view that early antigen encounter minimizes the attrition of CD8+ T cell responses in the elderly . Furthermore , analysis of the 2009 H1N1 ( swine-origin influenza ) response in human populations showed that this newly emerged pandemic virus shared immunogenic peptides with the catastrophic 1918 H1N1 strain [53] , emphasizing the likely value of establishing effective CD8+ T cell memory to all known influenza epitopes . Early priming of the CD8+ T cell compartment also preserves CD8+ T cell functionality in the very long term . In contrast to the suboptimal peptide-induced , polyfunctional cytokine profiles expressed by CD8+ T cells generated from naïve CD8+ T cells in aged animals , the recall of influenza-specific CD8+ T cell memory in the elderly is associated with functional profiles comparable to those found in the young . Since polyfunctionality ( simultaneous IFN-γ , TNF-α and IL-2 production ) of CD8+ T cells is thought to correlate with protective efficacy [54] , [55] , [56] , establishing optimal cytokine profiles early may provide a clear advantage for virus-specific CD8+ T cell responses in the elderly . Ageing is often associated with the attrition of the peripheral TCR repertoire , reflecting the loss of some T cell clonotypes and the large expansion of others [5] , [14] , [21] . Our study provides the most comprehensive analysis of the aged ( primary , secondary primed-early and secondary primed-late ) TCR repertoire published to date . The present , unbiased single-cell RT-PCR analysis of CDR3β usage in the elderly showed a diminished number of clonotypes during the aged primary DbPA224+CD8+ responses when compared with the normal profiles for young individuals [44] , [47] , [57] . As naïve DbNP366+CD8+ and DbPA224+CD8+ T precursors are efficiently recruited into the primary immune response [36] , this primary repertoire analysis can be considered to reflect the loss of a substantial proportion of naïve TCRs with ageing . Whilst a previous study [42] suggested that age-related clonal TCR attrition is more prevalent for the low precursor frequency DbNP366+CD8+ repertoire , we found a greater reduction in the numbers of DbPA224+CD8+ ( down 60 . 8% ) versus DbNP366+CD8+ ( down 34 . 9% ) -specific nucleotide clonotypes per mouse recovered following primary infection of older mice ( Tables 1 and 2 ) . This is likely to reflect that there are a greater variety of n . t . types encoding public DbNP366-specific aa clonotypes across all mice than for DbPA224+-specific aa clonotypes , which potentially makes DbPA224+CD8+ aa-defined clonotypes more vulnerable to total clonotype loss and thus reduced diversity . The public , aa-defined DbNP366+CD8+ CDR3β clonotypes can be encoded by up to 10 different n . t . sequences [44] , meaning that the loss of one n . t . -defined public TCR may not necessary result in the elimination of that particular CDR3β aa sequence . Thus , it is not surprising that the DbNP366+CD8+ CDR3β clonotypes in the aged mice following primary infection are encoded by a limited number of n . t . sequences ( 1 . 7±1 . 1 per mouse ) inferring a loss of DbNP366-specific CD8+ T cells . This was associated with the decreased contribution of two main public clonotypes ( SGGGNTGQL and SGGSNTGQL ) and the increased prominence of one public clonotype ( SGGANTGQL ) in aged mice following primary influenza virus challenge . Similar epitope-specific TCRβ repertoire homogenisation across a population of aged mice has been recently observed for CD8+ T cell responses to HSV-1 [22] . As SGGANTGQL is of lower pMHC avidity [49] , the dominance of this clonotype in the aged repertoire may be one reason for the lower functional quality of DbNP366+CD8+ T cell responses in the elderly . The real advantage of priming CD8+ T cell responses early in life is reinforced by the demonstration that n . t . -defined clonotype diversity is preserved for the public DbNP366+ CD8+ T cell response , resulting in more equal contribution of the 3 main public clonotypes ( SGGANTGQL , SGGGNTGQL and SGGSNTGQL ) , which was not seen when mice were primed later in life ( where SGGSNTGQL alone dominated ) . Similarly , the secondary DbPA224+CD8+ response in aged mice is slightly more diverse than that generated following primary virus challenge . Thus , early priming of the CD8+ T cell compartment induces a more diverse , aged repertoire by promoting the survival of public DbNP366+CD8+ clonotypes . This may in turn reflect the selection of “best-fit” TCRs . Maintaining TCR repertoire diversity can enhance the efficacy of CD8+ T cell-mediated immunity [58] , diminish the likelihood that mutated pathogens ‘escape’ immune recognition [59] and lead to more cross-reactive CD8+ T cell responses [53] , [60] . Preserving a greater breadth of responding TCRs is thus likely to be favorable for the elderly population . Taken together , our study supports the evolution of vaccine strategies to prime CD8+ T cells early in life in order to preserve the magnitude , functionality , TCR repertoire diversity and preferential TCR usage of responding populations .
All animal experimentation was conducted following the Australian National Health and Medical Research Council Code of Practice for the Care and Use of Animals for Scientific Purposes guidelines for housing and care of laboratory animals and performed in accordance with Institutional regulations after pertinent review and approval by the University of Melbourne Animal Ethics Experimentation Committee in Melbourne . Female C57BL/6J ( B6 , H2b ) mice were bred and housed under specific pathogen free ( SPF ) conditions at the Department of Microbiology and Immunology , University of Melbourne . Primary responses: For generation of acute primary influenza CD8+ T cell responses , mice were lightly anaesthetised by inhalation of methoxyflurane and infected intranasally ( i . n . ) with 1×104 plaque forming units ( pfu ) of H3N2 ( HK ) influenza A viruses in 30 µµl of PBS . Young mice were infected at 6–8 weeks , while aged mice were infected at 22 months of age . Secondary responses: To study the effects of early priming on aged CD8+ T cell responses , mice were first primed intraperitoneally ( i . p . ) at 6 weeks of age with 1 . 5×107 pfu of H1N1 PR8 influenza A virus and subsequently challenged with the serologically distinct H3N2 HK virus at extreme age of 22 months ( 6 weeks->22 months; primed young->challenged old ) . To study the effects of late priming on aged CD8+ T cell responses , mice were first primed i . p . with PR8 at 22 months and challenged 6 weeks later with HK ( 22 months ->23 . 5 months; primed old->challenged old ) . Control young animals were primed at 6 weeks , then challenged at 12 weeks of age ( 6 weeks->12 weeks; primed young->challenged young ) . The aged cohort of mice were held for up to 24 months in SPF conditions , monitored for signs of infection , weight loss and spontaneous tumor growth . Spleens were recovered from mice at acute phases of the primary and secondary infections ( day ( d ) 10 and d8 , respectively ) . Spleens were depleted of B cells by incubation on αIgG/IgM coated plates ( Jackson ImmunoResearch Labs ) for 45 mins at 37°C , and unbound cells harvested . Enriched lymphocytes from the spleen were stained with DbNP366 and DbPA224 tetramers conjugated to Strepavidin-APC or -PE ( Invitrogen ) at optimal staining concentrations for 1 hr at room temperature . Cells were then washed twice in FACS buffer ( PBS with 1%BSA/0 . 02% sodium azide ) and stained with 1 µg/ml CD8-PerCP Cy5 . 5 ( all BD Biosciences unless stated ) plus either: 5 µg/ml CD27-PE and 5 µg/ml CD43-FITC ( activation associated glycoform: clone 1B11 , eBiosciences ) or 5 µg/ml CD62L-FITC and 5 µg/ml CD127-PE ( IL-7Rα chain ) , or 5 µg/ml KLRG1-FITC ( Abcam ) and 5 µg/ml CD127-PE . For Vβ usage analysis , tetramer-stained cells were incubated a panel of FITC conjugated anti-Vβ mAbs ( 2 , 3 , 4 , 5 . 1/5 . 2 , 6 , 7 , 8 . 1/8 . 2 , 8 . 3 , 9 , 10 , 12 , 13 , 14 and 17 ) [61] at 5 µg/ml , and 1 µg/ml anti-CD8-PerCPCy5 . 5 . Cells were stained for 30 mins on ice , washed twice and analyzed by flow cytometry using a FACS Calibur ( BD Biosciences ) and Flowjo software ( Treestar ) . Splenocytes were stimulated with 1 µM NP366 or PA224 peptides ( AusPep ) for 5 hrs at 37°C , 5% CO2 in the presence of 1 µg/ml Golgi-Plug ( BD Biosciences ) and 10 U/ml recombinant human IL-2 ( Roche ) . Cells were washed twice with FACS buffer , stained with 1 µg/ml anti-CD8-PerCP Cy5 . 5 mAb for 30 mins on ice , fixed , permeabilised using the BD Cytofix/Cytoperm kit and stained with 5 µg/ml anti-IFN-γ-FITC , 2 µg/ml anti-TNF-α-APC , and 2 µg/ml anti-IL-2-PE mAbs . Samples were acquired by flow cytometry using a FACS Calibur and analysed by Flowjo . The total cytokine production was calculated by subtracting background fluorescence using no peptide controls . Splenocytes were stained with DbNP366-PE or DbPA224-PE tetramers in sort buffer ( PBS with 0 . 1% BSA ) for 1 hr at room temperature , washed and stained with 1 µg/ml anti-CD8-APC and 5 µg/ml of either anti-Vβ8 . 3 or anti-Vβ7-FITC for 30 mins on ice , washed twice with sort buffer . Single lymphocytes were isolated by sorting with a FACS Aria ( BD Immunocytometry ) into 80 wells of an empty 96 well twin-tec plate ( Eppendorf ) . mRNA transcripts were reversed transcribed to cDNA , using a Sensiscript kit ( Qiagen ) according to manufacturer's instructions , and the CDR3β region amplified by a nested PCR using Vβprimers [44] , [47] , [57] . Positive PCR products were purified using QIAGEN PCR purification kit and sequenced . Magnitude , phenotype and function were compared between experimental aged and young groups by an unpaired Student's t test . Clonotypic diversity was quantified using both the number of different clonotypes and Simpson's diversity index . The overlap of TCRβ repertoires between mice was quantified using both the proportion of the TCRβ repertoires per mouse comprised of shared clonotypes and the Morisita-Horn similarity index . The Simpson's diversity and Morisita-Horn similarity indices account for both the variety of distinct clonotypes ( defined either at the level of the amino acid or nucleotide sequence ) and the clone size ( number of copies ) of each clonotype involved in the epitope-specific response within each mouse [48] , [62] . The Simpson's diversity and Morisita-Horn similarity indices vary between 0 ( minimum diversity/similarity ) and 1 ( maximum diversity/similarity ) . The diversity and similarity measures were calculated in conjunction with a randomization procedure to correct for differences in sample sizes between mice [48] , [62] , and were estimated for a subsample of 22 TCRβ sequences . To estimate the proportion of the TCRβ repertoires per mouse comprised of shared clonotypes , clonotypes were pre-defined as shared based on their presence in more than one mouse prior to the random subsampling of 22 sequences . A Mann-Whitney test was used to compare , between pairs of groups , the diversity ( and similarity ) between the aged and young groups of mice in primary responses and between young and aged ( primed-young ) and aged ( primed-old ) in secondary responses . Bonferroni correction for multiple pairwise comparisons was applied for the comparisons between the three secondary response groups ( i . e . each pairwise test was assessed at the significance level of α = 0 . 05/3 = 0 . 0167 ) . All statistical analyses were performed using GraphPad Prism version 5 . 04 ( GraphPad Software Inc , San Diego , CA ) .
|
The elderly population is particularly susceptible to novel infections , especially the annual , seasonal epidemics caused by influenza viruses . Established T cell immunity directed at conserved viral regions provides some protection against influenza infection and promotes more rapid recovery , thus leading to better clinical outcomes . We asked whether priming early in life limits the time-related attrition of immune competence . We found that although influenza-specific T cell responses are compromised in the aged mice , vaccination with influenza early ( but not late ) in life ‘locks’ optimal T-cell responsiveness , maintains functional quality , persistence of preferred clones and a characteristic T cell hierarchy . Overall , our study supports development of vaccines that prime T cells early in life to elicit the broadest possible spectrum of pre-existing T cell memory and preserve the magnitude , functionality and clonal usage of responding populations for life-long immunity against influenza viruses .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"infectious",
"diseases"
] |
2012
|
Early Priming Minimizes the Age-Related Immune Compromise of CD8+ T Cell Diversity and Function
|
Schistosomiasis japonica still remains of public health and economic significance in China , especially in the lake and marshland areas along the Yangtze River Basin , where the control of transmission has proven difficult . In the study , we investigated spatio-temporal variations of S . japonicum infection risk in Anhui Province and assessed the associations of the disease with key environmental factors with the aim of understanding the mechanism of the disease and seeking clues to effective and sustainable schistosomiasis control . Infection data of schistosomiasis from annual conventional surveys were obtained at the village level in Anhui Province , China , from 2000 to 2010 and used in combination with environmental data . The spatio-temporal kriging model was used to assess how these environmental factors affected the spatio-temporal pattern of schistosomiasis risk . Our results suggested that seasonal variation of the normalized difference vegetation index ( NDVI ) , seasonal variation of land surface temperature at daytime ( LSTD ) , and distance to the Yangtze River were negatively significantly associated with risk of schistosomiasis . Predictive maps showed that schistosomiasis prevalence remained at a low level and schistosomiasis risk mainly evolved along the Yangtze River . Schistosomiasis risk also followed a focal spatial pattern , fluctuating temporally with a peak ( the largest spatial extent ) in 2005 and then contracting gradually but with a scattered distribution until 2010 . The fitted spatio-temporal kriging model can capture variations of schistosomiasis risk over space and time . Combined with techniques of geographic information system ( GIS ) and remote sensing ( RS ) , this approach facilitates and enriches risk modeling of schistosomiasis , which in turn helps to identify prior areas for effective and sustainable control of schistosomiasis in Anhui Province and perhaps elsewhere in China .
Schistosomiasis , caused by trematode worms belonging to the Schistosoma genus [1] , remains a serious public health problem worldwide [2] . More than 200 million people in approximately 76 countries are affected by the disease with at least a loss of 1 . 7 to 4 . 5 million disability-adjusted life years ( DALYs ) [3] and probably considerably more [4] . The true global burden of schistosomiasis japonica alone has been shown to be between four to 30 times greater than previously expected [5] . Of the three main schistosome species , Schistosoma japonicum is responsible for human and animal infections in southern China , large parts of The Philippines , and limited foci in Indonesia [6] . According to geographical patterns of the endemic areas and ecological characteristics of the vector snail , schistosomiasis endemic regions in China have been classified into three types: lake and marshland regions , plain regions with water-way networks , and hilly and mountainous regions [7] . Compared to the other two regions , control of the disease in the lake and marshland regions has proved to be difficult due to vast areas of Oncomelania hupensis habitats [8] , and over 80% of schistosomiasis cases occurred in these regions [9] . With the Yangtze River passing across the province and the presence of large amounts of wet land , Anhui presents suitable environmental conditions for the formation of lake-and-marsh endemic regions . The transmission of schistosomiasis is closely associated with the distribution of the intermediate host snail , which largely depends on environmental conditions such as vegetation coverage , temperature of freshwater , and quality and humidity of the soil [10 , 11] . In the lower reaches of the Yangtze River Basin , snails are distributed along the shore of rivers or lakes . The regular tide , together with large amount of marshland , provides an ideal environment for snail growth and reproduction , which can be characterized by “land in winter , water in summer” [12] . Under such favorable physical conditions , there is a concern that snails might explode and spread again , possibly giving rise to extensive re-emergence of infections among humans and domestic animals in the basin . Techniques of geographic information system ( GIS ) and remote sensing ( RS ) , combined with geostatistics , have been widely used in modeling the burden of schistosomiasis over the past decades [12 , 13 , 14 , 15 , 16] . However , most of the previous work only considered spatial correlations of prevalence data , and only a few attempts have been made to investigate spatio-temporal correlations and to assess how environmental factors affect these correlations . In this study , we aim to investigate changes in the spatio-temporal pattern of schistosomiasis in Anhui Province of China to better understand how environmental factors affect the changes of the disease , using a spatio-temporal kriging model [17] . This method offers a variety of techniques to make optimal use of measurement information for interpolating attributes in space and time , and has been applied over the last decade in such diverse scientific disciplines as environmental science [18] , meteorology [19] , and soil science [20] .
In the present study , we integrate data sourced by application of GIS and RS with spatio-temporal kriging modeling to assess the schistosomiasis risk . The analysis is conducted at the village level with the study area located in Anhui in eastern China ( Fig . 1 ) . Anhui is a province spanning approximately 139 , 600 square kilometers and with a population of 59 . 9 million ( 2012 ) . Most of the province is very flat , with a series of hills and ranges covering southwestern and southeastern Anhui . Major rivers include the Huaihe River in the north and the Yangtze River in the south . The province enjoys a subtropical humid monsoon climate . Plum rains occur in June and July and may cause flooding . The S . japonicum infection prevalence data during 2000–2010 were obtained from cross-sectional surveys carried out by health professionals of the Anhui Institute of Parasitic Diseases ( AIPD ) . The data were collected annually through village-based field surveys using a two-pronged diagnostic approach: screening by a serological test of all residents 5 to 65 years old and confirmed by Kato-Katz stool examination [21] . The results were reported to AIPD via county offices . The database used in this study consisted of 161 sample villages located in 24 schistosome-endemic counties , which were selected from the database of annual schistosomiasis surveys with the criteria that the village should be surveyed every year and the examined people should be over 100 . Fig . 1 shows the locations of the sample villages in the endemic area . Approval for oral consent and other aspects of this survey was granted by the Ethics Committee of Fudan University ( ID: IRB#2011–03–0295 ) . Written informed consent was also obtained from all participants . Climatic data . The climatic data include normalized difference vegetation index ( NDVI ) and land surface temperature at daytime ( LSTD ) and night ( LSTN ) . All 8-day global 1 km products for LST and monthly global 1 km products for NDVI during the period 2000–2010 that covered Anhui Province were downloaded from the Level 1 and Atmosphere Archive and Distribution System ( http://ladsweb . nascom . nasa . gov/data/search . html ) . These images were georeferenced and subsetted in ENVI software ( version 5 . 0 , Research System Inc . ; Boulder , CO , USA ) . AcrGIS software ( version 10 . 0 , ESRI Inc . ; Redlands , CA , USA ) was used to extract monthly-average NDVI , LSTD , and LSTN , for each pixel of the image . Four indices ( minimum , maximum , mean , and standard deviation ( SD ) ) of these monthly-average variable for each year were obtained for each village to capture , albeit crudely , the effects of overall climatic condition and seasonal variation in local climate . Distance to the Yangtze River . Data on the Yangtze River were downloaded from Conservation Science Data Sets of World Wildlife Foundation at http://worldwildlife . org . For each sample village , the Euclidian distance to the Yangtze River was calculated using AcrGIS software . Ordinary least squares ( OLS ) regression models were fitted to schistosomiasis data to identify significant environmental covariates . Initially , univariate analyses were conducted and variables with P>0 . 2 were excluded . With the remaining variables , backwards-stepwise regression was conducted using P >0 . 1 as the exit criterion and P ≤0 . 05 as the entry criterion . In the final multivariate model , SD of NDVI , SD of LSTD , and distance to the Yangtze River remained . A universal spatio-temporal kriging model [22] was fitted to investigate the spatio-temporal pattern of schistosomiasis as well as the effects of environmental factors on the pattern . Let Y ( s , t ) denote prevalence of schistosomiasis in village s at year t . The model assumes that the spatio-temporal process of the prevalence variable is composed of the sum of a trend and a stochastic residual as follows: Y ( s , t ) =m ( s , t ) +ε ( s , t ) ( 1 ) where m is the trend ( i . e . , a linear function of the covariates ) , which can be determined by the result of the OLS regression model above , and εis the spatio-temporal correlated stochastic component with zero mean . To ease statistical inference , it is commonly assumed that the zero-mean stochastic part is multivariate normally distributed . To estimate the spatio-temporal covariance structure of ε , we assume that the variance of ε is constant and that the covariance at sample villages ( s , t ) and ( s + h , t + u ) only depends on the separation distance ( h , u ) , where h is the Euclidean spatial distance and u is the distance in time . Or , simply , we assume ε to be stationary and spatially isotropic . The spatio-temporal covariances are usually described using a spatio-temporal variogram ( γ ) , which measures the average dissimilarity between data separated in the spatio-temporal domain using the distance vector ( h , u ) defined as follows: γ ( h , u ) =12E[ε ( s , t ) −ε ( s+h , t+u ) ]2 ( 2 ) where γ ( h , u ) denotes the semivariance of ε and E denotes the mathematical expectation . In practice , when dealing with real-world data , spatio-temporal variograms are fitted by introducing simplifying statistical assumptions . In this study , we use an inseparable model called “product-Sum” model . This assumes that the spatio-temporal variogram consists of three stationary and independent components: γ ( h , u ) =γs ( h ) +γt ( u ) −kγs ( h ) γt ( u ) ( 3 ) where γs ( h ) and γt ( u ) are purely spatial and temporal variograms respectively , and k is a real coefficient . The product-Sum model can be seen as a surface with six parameters: two parameters for each variogram ( sill and range ) and a joint spatio-temporal sill and nugget . In turn , these parameters can be used in spatio-temporal kriging to compute the best linear unbiased predictor ( i . e . , with minimum expected mean-squared error ) for any space-time point where ε ( and Y ) was not observed . The formulas of kriging in the spatio-temporal domain do not differ fundamentally in a mathematical or statistical sense from those of spatial kriging: ε⌢ ( s0 , t0 ) =c0Tc−1ε¯ ( 4 ) where c is the n×n variance-covariance matrix of the residuals at the n observation space-time points , as derived from the spatio-temporal variogram , c0 is a vector of covariances between the residuals at the observation and prediction points , T denotes matrix transpose , and ε¯ is a vector of residuals at the n observation points . The final prediction of prevalence Y at a village ( s0 , t0 ) is defined as Y^ ( s0 , t0 ) =m^ ( s0 , t0 ) +ε^ ( s0 , t0 ) ( 5 ) where z^ ( s0 , t0 ) is the estimated multivariate linear regression trend . For prediction , the endemic area is divided at the 2-km resolution level . A Box-Cox transformation [23] of the crude prevalence is performed to apply the Gaussian model before implementing spatio-temporal kriging . A purely spatial universal kriging was also fitted by year separately for comparative purpose . The performance of the kriging interpolation would be substantially affected if the spatial stratification is strong and spatial autocorrelation is weak [24] . To ensure this precision , we investigated the spatial heterogeneity of the yearly prevalence of schistosomiasis by employing an indicator of power of determinant ( PD ) [25] . The PD value , ranging from 0 to 1 , quantifies how similar is the spatial distribution of a disease with that of a risk factor . If the PD value is closer to 1 , the disease has more similar spatial stratification with that of the factor; if it is closer to 0 , the spatial stratifications of the two are quite different . Spatial stratifications of environmental factors were zoned as follows: SD of NDVI and SD of LSTD over the study area were both classified by four equal intervals; distance to the Yangtze River was divided into four buffers: 0~5km , 5~10km , 10~20km , and over 20km . Cross-validation is applied for assessing the accuracy of the predictions made for prevalence in sample villages as obtained with the spatio-temporal and purely spatial kriging . Specifically , we use leave-one-out cross validation . The method proceeds as follows: one sample village is retained as the validation data for testing the spatio-temporal kriging model , and the remaining sample villages are used as training data to construct the model . This is repeated such that each sample village is used once as the validation data . An indicator of root-mean-square error ( RMSE ) defined as follows is used to assess the final accuracy of the model: RMSE=111n∑t=20002010∑i=1n[Y^ ( si , t ) −Y ( si , t ) ]2 ( 6 ) where Y^ ( si , t ) and Y ( si , t ) are the predictive and the observed prevalence at the sample village ( si , t ) respectively , and n is the number of sample villages . PD values were calculated using the GeoDetector software freely available at http://www . sssampling . org/Excel-GeoDetector/ . Spatio-temporal kriging were implemented in the R package gstat [17] , and mapping of predicted prevalence of schistosomiasis and its corresponding variance were performed using the same package as well .
Fig . 2 gives some statistics about the annual prevalence of schistosomiasis during the study period . The mean observed prevalence generally decreased from 0 . 76% in 2000 to 0 . 17% in 2011 and the Kruskal-Wallis test revealed that the mean prevalence significantly differed by year ( χ2 ( 10 , N=1771 ) =30 . 47 , p<0 . 01 ) . This downward trend was accompanied by a decreasing variation in prevalence across villages with the interquartile range ( IQR ) contracting from 0–0 . 8/100 in 2000 to 0–0 . 1/100 in 2010 , indicating a decreasing disease burden . Table 1 shows parameter estimates from the OLS regression model . The results indicate that SD of NDVI , SD of LSTD , and distance to the Yangtze River are significantly associated with S . japonicum risk . In particular , the infection prevalence increases with decreasing SD of NDVI ( coef = -1 . 21e-04 , p<0 . 01 ) , with decreasing SD of LSTD ( coef = -5 . 61e-05 , p<0 . 01 ) , and with shorter distance to the Yangtze River ( coef = -1 . 43e-03 , p = 0 . 01 ) . The left-hand side of Fig . 3 shows the sample residual variogram of the infection prevalence of schistosomiasis , while the right-hand side presents the fitted residual variogram . The rising trend at both spatial and temporal dimension in sample variogram indicates that spatio-temporal correlation is present although the correlation seems not very strong , and therefore , spatio-temporal kriging of residuals is applicable . Table 2 summaries the parameter estimate of the product-Sum variogram model . Note that all variogram components were modeled as exponential function . The range of spatial dependency and temporal dependency is 18 km and 5 . 48 years ( i . e . , 2000 days ) , respectively . The cross-validation results on spatio-temporal kriging and purely spatial kriging yield RMSE of 0 . 61 and 0 . 84 , respectively , which indicate that the spatio-temporal kriging model has a better predictive ability . Table 3 presents PD values of the environmental factors for each year , which help to investigate spatial heterogeneity of the prevalence of schistosomiasis . The PD values of SD of NDVI range from 0 . 15 to 0 . 20 with the mean of 0 . 17 , those of SD of LSTD vary from 0 . 13 to 0 . 18 with the mean of 0 . 15 , and those of distance to Yangtze River change from 0 . 26 to 0 . 33 with the mean of 0 . 28 . Fig . 4 displays the annual map of predicted prevalence for S . japonicum infection and it can be seen that the infection prevalence is generally low , namely , most areas with prevalence below 0 . 1% and very limited areas with prevalence over 1% . The infection risk showed a focal spatial pattern and the pattern fluctuated temporally with a peak ( the largest spatial extent with prevalence over 0 . 1% ) in 2005 and then contracted gradually but with scattered distribution until 2010 . Note that clusters of schistosomiasis risk mostly occurred along the Yangtze River . Fig . 5 represents corresponding estimates of the variance of the predictions . The maps present similar patterns across the study period: a lower level of uncertainty is apparent in locations close to sampled villages while a higher level of uncertainty is present in locations distant from sampled villages . The prediction uncertainty is generally low over the study area .
This study demonstrates the use of a spatio-temporal kriging model in assessing how environmental factors affect the outcome of human schistosomiasis based on spatio-temporally correlated disease data . Our results confirmed spatio-temporal differences in the infection risk and the important role of environmental factors in explaining the variations . The predicted risk maps , in return , provide an empirical basis for identifying priority areas when implementing schistosomiasis controls locally . Schistosomiasis is a water-borne disease and its transmission is strongly associated with environmental factors . We , therefore , considered three key elements that characterize schistosomiasis transmission [26 , 27 , 28] , namely , temperature , wetness , and access to infected water , using LSTD , LSTN , NDVI , and distance to the Yangtze River . Many studies [29 , 30 , 31 , 32 , 33] had explored effects of these elements , but seldom did those use the four indices ( minimum , maximum , mean , and SD ) of climatic factors to account for effects of overall climatic condition and seasonal variation in local climate . Our risk analysis showed that seasonal variation of LSTD and NDVI , and the Yangtze River were significantly , negatively correlated with the risk of S . japonicum . There is a plausible , biological explanation for these associations as discussed below . The explanation for the negative association between seasonal variation of LSTD and schistosomiasis risk derives from the parasite’s life cycle , of which several stages ( i . e . , the excreted egg , miracidium , sporocyst , and cercaria ) require fresh water environment . Previous studies have shown that the development of the parasite residing in the intermediate host snail is closely related to the environmental temperature , e . g . , large seasonal temperature differences ( characterized by higher SD ) would hamper the development of miracidia into cercariae and the situation worsens if the thermal limits are exceeded [34 , 35] . The parasite cannot complete its life cycle optimally in areas with larger temperature differences , and hence less cercariae are released into freshwater environments . As the cercaria is the infective stage for both humans and mammalian reservoir hosts , the disease transmission intensity decreases . NDVI maps , indicating the amount of vegetation present at each location , have been widely and successfully used for prediction of intermediate host snails of schistosomiasis [31] and are often used as a proxy for suitable O . hupensis habitats . An area with higher SD of NDVI indicates the vegetation coverage is not constant and hence it is not ideal for formation of snail habitat . The distance to the Yangtze River can be seen as a proxy of exposure ( due to increased water contact ) . Individuals living near the shore are more likely to risk contact with water containing infected snails as a result of their professional work ( e . g . agricultural activities , fishing , etc . ) and life style ( e . g . , cleaning and swimming ) . In addition , frequent seasonal flooding events might also increase their exposure to cercariae . Parameters of fitted residual variogram of the infection prevalence of schistosomiasis can characterize the spatial and temporal variation of the disease . The spatial range of 18km suggested that the spatial correlation become negligible after 18km , and such distance implied that transmission occurred between villages rather than within and around them . On the other hand , the temporal range of about 6 years indicated that temporal correlation become negligible after this range , and such a long period reflected that the burden of schistosomiasis should not vary greatly for each year and was probably related to the low infection rate as a result of implementation of schistosomiasis control strategies . The rising trend of the sample residual variogram ( as shown by the 3-D plot in Fig . 3 ) indicated that spatio-temporal correlation was present in schistosomiasis , though it is not very strong , after adjusting the current environmental factors . This mild spatio-temporal correlation suggested that the spatio-temporal pattern in schistosomiasis risk is probably already captured by the environmental factors . Kriging interpolation is principally based on spatial/spatio-temporal autocorrelation , the precision of which would be poor if the spatial stratification is strong and spatial autocorrelation is weak . The mean PD value of 0 . 17 for distance to snail habitat indicated that SD of NDVI explained 17% of the variation of schistosomiasis; similarly , SD of LSTD and distance to snail habitat explained 15% and 28% of the variation of the disease , respectively . The findings implied that prevalence of schistosomiasis showed week spatial heterogeneity within the buffers . Combined with ST correlation , the week spatial heterogeneity of the disease over the study area during the study period justified the employment of spatio-temporal kriging . The spatio-temporal variations of schistosomiasis risk shown in Fig . 4 can be explained by the schistosomiasis control strategies implemented in the study period . In the early 1990s , the Chinese government launched a 10-year World Bank Loan Project ( WBLP ) on schistosomiasis control [36] , strongly based on large-scale chemotherapy but with additional intervention activities such as health education , chemical control of snails , and other environmental exposure modifications . The disease , however , rebounded shortly after the conclusion of WBLP in 2001 [37 , 38] . This rebound can be seen from maps in 2002–2005 . In order to deal with the rebounding trend , a revised strategy , aimed at reducing the role of bovines and humans as infection sources and based on integrated measures , was implemented from 2005 [39] . In addition to chemotherapy and health education , water buffaloes and cows were replaced by tractors and the integrated program also included such strategies as treatment of night-soil and provision of piped , safe water [40] , keeping domestic animals in barns [41] , and reduction of snail habitats through the construction of water conservancy projects [41] . However , the scattered distribution of schistosomiasis risk shown in maps of 2007 to 2009 suggested that the integrated strategy could not effectively compress the spatial extent of the disease , indicating there are still large populations at risk . As a reflection of rebound trend of schistosomiasis , infected O . hupensis snails are still found in certain locations along the Yangtze River [42] , and there is concern that they might spread further , possibly resulting in extensive re-emergence of infections among people and domestic animals . Since control measures are limited to bovines and humans and more than 40 species of mammalians can serve as potential zoonotic reservoirs , the infectious-source measures can’t block the life cycle of the parasite completely [43] . Furthermore , the integrated strategy is expensive as it involves many diverse activities while budgets on schistosomiasis control may probably be reduced in the foreseeable future . A less costly , but still effective and sustainable control strategy is urgently needed . Targeting the snail habitats within areas of high schistosomiasis risk can be a way out as the amphibious O . hupensis is the only intermediate snail host and may , therefore , be the weak link in the parasite’s life cycle . Our analysis provides an empirical basis for identifying priority areas . As seen in Fig . 4 , areas with relatively high risk ( i . e . , prevalence > 0 . 1% ) , especially those areas with constant clusters of risk , would definitely be a priority for targeting schistosomiasis control in local regions . Some limitations in our study deserve further discussions . Firstly , the slightly rising trend of the sample residual variogram in Fig . 3 is probably due to lack of other risk factors , which indicates that more risk factors should be considered . Therefore , in addition to the current environmental factors ( i . e . , LST , NDVI , and the Yangtze River ) , many other factors , such as landscape metrics , socio-economic impacts , interventions , etc . , should be warranted in further studies . Second , spatio-temporal kriging model relies on an assumption of stationarity . This assumption is appropriate when mapping disease over small areas , but might be questionable over wide areas , such as a country [44] . To investigate the effects of non-stationarity is interesting and challenging topic and should be considered in further studies . Finally , the specificity of serological assays and the sensitivity of stool examination tests are not perfect [45] . The generally low levels of infection with S . japonicum in recent years result in uncertainty both with regard to sensitivity and specificity of infection [46] . Modeling with diagnostic errors should be considered in future studies . In summary , this study investigated a region where schistosomiasis endemic remains of public health and economic significance and might spread as infected snails are still found . Combined with techniques of GIS and RS , the spatio-temporal kriging model disclosed the spatio-temporal patterns of S . japonicum infection , which , we believe , helps to facilitate and enrich risk modeling of the disease . The results can be used to identify priority areas where control efforts should be taken . Compared to the ongoing costly integrated strategy with infectious-source controlling as an emphasis , targeting snails ( the only intermediate host in the parasite’s life cycle ) and taking corresponding effective actions ( e . g . , mollusciciding and environmental modification ) in these priority areas would be sustainable in schistosomiasis control in the long term .
|
Schistosomiasis japonica is one of the most serious parasitic diseases in China . It is estimated that more than 50 million people are still at risk , especially those living in the lake and marshland areas along the Yangtze River Basin . The Chinese government has made great efforts to implement schistosomiasis control programs since 1950s . The latest , major two programs are the 10-year World Bank Loan Project ( WBLP ) terminated in 2001 , which was based on large-scale chemotherapy , and the national integrated control strategy implemented since 2005 , which was aimed at reducing the roles of bovines and humans as infection sources . Based on spatio-temporal analyses of the S . japonicum infection prevalence data during 2000–2010 in Anhui Province , we found schistosomiasis prevalence remained at a low level but the spatial distribution of the disease became widely scattered at the later stage of the study period , suggesting that the integrated program could not fully effectively reduce the spatial extent of schistosomiasis risk . To achieve an effective and sustainable control strategy , we emphasize the need to control snail habitats within areas of high schistosomiasis risk .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Spatio-temporal Transmission and Environmental Determinants of Schistosomiasis Japonica in Anhui Province, China
|
Multifunctionality is a common trait of many natural proteins and peptides , yet the rules to generate such multifunctionality remain unclear . We propose that the rules defining some protein/peptide functions are compatible . To explore this hypothesis , we trained a computational method to predict cell-penetrating peptides at the sequence level and learned that antimicrobial peptides and DNA-binding proteins are compatible with the rules of our predictor . Based on this finding , we expected that designing peptides for CPP activity may render AMP and DNA-binding activities . To test this prediction , we designed peptides that embedded two independent functional domains ( nuclear localization and yeast pheromone activity ) , linked by optimizing their composition to fit the rules characterizing cell-penetrating peptides . These peptides presented effective cell penetration , DNA-binding , pheromone and antimicrobial activities , thus confirming the effectiveness of our computational approach to design multifunctional peptides with potential therapeutic uses . Our computational implementation is available at http://bis . ifc . unam . mx/en/software/dcf .
The combination of multiple functions is an ubiquitous feature of naturally occurring proteins . More than 60% of the proteins in archaea and bacteria and more than 80% of eukaryotic proteins include more than one functional domain [1] . These numbers are further increased by a group of at least 200 moonlighting proteins , incorporating several functions in a single domain [2–4] . A similar situation has been described for peptides . For instance , some antibacterial peptides have the ability to either bind DNA , to penetrate cells or both [5 , 6] . Thus , multifunctionality in proteins and peptides seems to be a common feature in nature rather than an exception and it is then relevant to understand the basis for this diversity . In this sense , understanding how proteins acquire multiple functions is important to understand the structure-function relationship of proteins and aid in the design of polypharmacological peptides , an area of great interest in recent years to both academia and industry [7 , 8] . In the present work we report a novel computational method to design peptides with multiple functions . Two kinds of relevant peptides with therapeutic application are cell-penetrating and antimicrobial peptides . Cell-penetrating peptides ( CPPs ) have the intrinsic ability to cross a variety of cellular membranes , which has been used for medical applications , particularly for cargo delivery [9–12] . Antimicrobial peptides ( AMPs ) on the other hand comprise a large class of naturally occurring peptides used to fight bacterial or fungal infections [13–15] . Fusing CPPs and AMPs has been shown to render multifunctional peptides useful for treating cancer , obesity and potentially many other diseases [16–19] . As such , it comes of no surprise that there have been ongoing efforts to design new cell penetrating or antimicrobial peptides [20–24] . However , fusing two or more activities into a peptide increases the peptide length and consequently its cost and immunogenicity , or may create an inactive peptide . Alternatively , the inclusion of several functions into a single antimicrobial peptide has medical potential as it allows to include mechanisms for specificity , organelle targeting or cargo delivery . As such it remains a challenge to design multifunctional antimicrobial peptides . We have previously reported one computational strategy to create multifunctional peptides . Our designed peptides , referred to as Iztli peptides , embed the α-factor pheromone of Saccharomyces cerevisiae within an AMP sequence [25 , 26] . The characterization of these peptides leads us to propose that cell-penetrating and antimicrobial activities are closely related , that is , the design of an AMP may be compatible with a CPP activity and vice versa . We show here that this principle of compatibility can be used to design a set of multifunctional peptides , including two different functional domains and having an additional antimicrobial moonlighting activity . That is , we have designed CPPs with antimicrobial activity that embeds both pheromone and nuclear localization activities . This new design avoids some of the common limitations in the design of antibacterial peptides and produces multifunctional peptides . Our results may provide an alternative way to explore further the structure-function relationship of peptides and proteins .
The design of our peptides was based on two major assumptions . First , obtaining a probabilistic predictor for CPPs and optimizing peptide sequences to be consistent with this predictor may yield a compatible antimicrobial activity . Second , protein domains with different functions and high predicted CPP probability ( or any other experimentally validated compatible function ) should be particularly easy to embed in a peptide sequence that retains multifunctionality . To this aim , we designed multifunctional peptides that would fit the rules for CPPs at the sequence level . We distinguished two different sets of CPPs: low and high efficient CPPs . Consequently , we tested several machine learning algorithms for their accuracy in identifying peptide sequences with CPP activity or efficient CPP activity . The accuracy ( fraction of correctly predicted CPP sequences ) was estimated from a 4-fold cross validation test ( 16 samples in total; the training set was randomly divided into 4 groups and in each turn , one group was retained as validation set and the rest was used for training ) . In agreement with previous results [27 , 28] , we found that support vector machines ( 91% and 66% accuracies for CPPs and efficient CPPs , respectively ) and random forests ( 90% and 67% accuracies for CPPs and efficient CPPs , respectively ) yielded the best results . Here , random forests were much faster in training and rendering predictions and directly gave a probability estimate . Thus , here we used random forests for our study . The trained random forest predictor could now be used to assign a probability of being a CPP , P ( CPP ) or highly efficient CPP P ( eff ) , to arbitrary peptide sequences . The described classification model was used to find compatible protein activities . According to our hypothesis , our model should predict peptide sequences with intrinsic AMP activity . To test this , we first predicted the mean probability to constitute a CPP and also be an efficient CPP , P ( CPP , eff ) = P ( CPP ) ·P ( eff ) , for all conserved domains sequences in the PFAM database ( version 27 . 0 , 14 . 831 protein families , >10 Mio . sequences , see Fig 1 ) . This model yielded 82 protein families with P ( CPP ) >0 . 5 and only a single family , “Protamine P1” , with P ( CPP , eff ) >0 . 5 ( for the complete distribution of P ( CPP ) values see S1 Fig and S1 Table ) . The 82 protein families were classified by their biological function and biological process GO terms as well as their descriptions in the PFAM database . Only 33 families had functional characterization in GO and 24 of these constituted DNA binding domains or structural components of the ribosome , thus , indicating that interaction with nucleic acids is compatible with CPP function . Families with antimicrobial activity formed the second largest functional group . 4 of the 9 families annotated as “antimicrobial” in PFAM were contained in the 82 CPP compatible families , as well as 3 additional antimicrobial families ( Myotoxins , Ponericin and Mellitin ) annotated in GO database . This was consistent with our aim and previous findings that CPP and antimicrobial peptides may be functionally related [26] . In summary , our results suggested that not only AMP but also DNA-binding activities are compatible with CPP rules . To validate these results , we designed several peptides that embed a Nuclear Localization Sequence ( SV40 large T-antigen nuclear localization factor , NLS ) and a ligand that is endocytosed ( α-factor pheromone of S . cerevisiae [29] ) within sequences that optimally matched CPP rules ( see Table 1 ) . Those optimal peptide sequences embedding other compatible activities were obtained by coupling the previously obtained random forest model with a Simulated Annealing optimization procedure . In this procedure , every peptide sequence with a compatible activity ( here also referred to as peptide templates ) were linked and flanked by short amino acid fragments ( see Fig 1A ) whose sequences are obtained from the optimization algorithm ( see Materials and Methods ) . This method was used to create three different peptides using three different designs . In our first design we placed α-factor before NLS and optimized only the probability to constitute a CPP-like sequence ( named α-NLS-C ) . Here , three short fragments of two to three amino acids each were sufficient to obtain a P ( CPP ) >95% . In the remaining two designs , we switched the order of the templates and did not optimize the C-terminal end of the peptide since C-terminal modification of α-factor has been reported to inhibit its activity [30] . In the first design , the entire sequence was optimized to obtain a maximal joint probability of constituting a CPP-like sequence with high efficiency ( named NLS-α-CE ) . In the second design , we generated a chimera peptide where the NLS template was flanked by a fragment constituting a high joint probability of being a CPP-like sequence with high CPP efficiency; the other fragment harboring the α-factor template was flanked by a fragment yielding a high joint probability of being a CPP-like sequence with similarity to membrane-binding sequences . As can be observed in Table 1 , chimera peptide presents the largest P ( membind ) score in comparison with the other two designed peptides in this study ( α-NLS-C and NLS-α-CE ) . The rules to bind membranes were included to test if these may improve CPP activity . This chimera peptide has a hydrophobic N-terminal NLS tail , a hydrophilic cationic C-terminal α-factor tail and a globally high probability of being CPP . During optimization we also chose the sequences with the largest predicted fraction of α-helices from the top 20 sequences . Table 1 shows the sequences of all designed peptides ( listed in bold ) and controls , together with their global predicted probabilities to match CPP rules ( P ( CPP ) ) , or to match rules of highly efficient CPP ( P ( eff ) ) or a membrane-binding peptide rules ( P ( membind ) ) . We first evaluated whether the designed peptides retained the activity of the templates . The presence of the NLS sequence may locate peptides in the nuclei of cells; for that to happen first the peptides have to be internalized . Hence , we tested the ability of these peptides to be internalized into two different cell types: yeast and mammalian cells . For yeast cells , confocal microscopy was used to evaluate CPP activity in our peptides ( see Fig 2 ) . Here , we used TAMRA-labeled versions of the peptides and as controls , variants of these peptides in which Lys and Arg residues were replaced by Glu residues ( woRK peptides in Fig 2 ) . α-NLS-C , NLS-α-CE and chimera did all quickly accumulate in the cell membrane and inside the yeast cells , whereas none of the controls were detected inside the yeast cells within the first 10 min . Next , we tested the ability of these peptides to be localized also in mammalian cells , and whether they would reach the nuclei . Here , we performed internalization assays in mouse embryonic fibroblasts ( primary cultures ) in which nuclear DNA was stained with DAPI after 20 min of cells being exposed to the designed peptides or the corresponding “woRK” controls ( see Methods ) . By fluorescence microscopy we observed indeed both internalization and nuclear localization ( co-localization with DAPI ) only for the designed peptides , but not for the control “woRK” peptides ( see Fig 3 ) . The co-localization of TAMRA ( labeled peptides ) and DAPI ( nuclei ) fluorescence indicated that the designed peptides are also internalized into MEF cells and localize at the nuclei , possibly interacting with DNA . To test the ability of the designed peptides to bind nucleic acids , we evaluated by electrophoretic mobility shift assays ( EMSAs ) the retention of DNA by the designed peptides . Each EMSA was executed with peptide quantities that were multiples of the DNA to peptide charge ratio . A DNA to peptide charge ratio of 1:1 is represented by the number of positive charges of the peptide required to neutralize the negative charges of the DNA backbone . Retention in the gel is an indirect measure of the fraction of positive charges within the peptide participating in binding the DNA molecules . All EMSA assays were performed using the yeast plasmid pGREG546 . As negative control , pGREG546 without any peptide was used , and as positive control the known DNA-binding cell penetrating peptide MPG was used [31 , 32] . We observed that NLS alone was not capable of binding DNA in low concentrations . However , retention in the gel was observed when NLS was joined to any larger peptide fragment . Furthermore , all designs including the NLS sequence showed almost complete DNA retention at the minimal DNA to peptide charge ratio of 1:1 and this behavior remained the same when increasing the peptide concentrations to yield a DNA to peptide charge ratio of 1:4 ( see Fig 4A–4C ) . α-factor alone showed a weak ability to bind DNA . DNA retention was around 60% in the 1:4 DNA to peptide charge ratio suggesting that only a small fraction of the positive charges in α-factor participate in DNA-binding . Adding positive charges to the N-terminus of α-factor increased its affinity for DNA , such as in the case of the Iztli peptide 1 [25] . These results indicate that increasing the number of positive charges in peptides , as it often occurs in CPP design , promotes the peptides' affinity for DNA and supports that these activities are compatible . Nevertheless , DNA binding may be governed by diverse structural factors , but our results show that positively charged residues contribute to it . It should be noted that the optimized peptides showed the same in vitro DNA affinity as MPG , a peptide known to bind DNA in vivo [24] . Finally , the activity of the α-factor embedded in the peptides was evaluated by activation of the pheromone-signaling pathway in MATa cells of S . cerevisiae . We monitored the pheromone signaling activity by following the expression of the GFP-labeled Fus1 protein . Fus1 is among the proteins with the highest pheromone-induced expression and is required for cell-cell fusion during mating [33 , 34] . NLS-α and α-NLS-C showed significant induction of expression of Fus1-GFP , however , higher concentration of both peptides is required compared to the free α-factor ( see Fig 5A–5G ) . The observed phenotype was consistent with the activation of signaling showing a mating protrusion , often called “shmoo” as well as the absence of budding due to inhibition of the Cdc28 cyclin [35 , 36] . NLS-α-CE and chimera showed partial shmoo-like phenotypes , a diffuse GFP , strong vacuolar fragmentation and a less refractive membrane in differential interference contrast bright field . As the same phenotype could also be observed for Iztli-1 , a known fungicide , it is likely that this phenotype was the consequence of cellular damage . In conclusion , optimizing the designed peptides by our algorithm enabled the DNA-binding activity of the NLS subunit both in vitro and in vivo , presented the CPP activity and retained the pheromone activity of the α-factor subunit , although this last activity was reduced probably caused by cellular damage . If CPP , DNA binding and AMP activities are compatible , the designed peptides should inhibit bacterial growth . Here we exposed E . coli DH5α cells to varying low concentrations of the designed peptides as well as several controls ( see Fig 4 ) . All designed peptides inhibit E . coli growth , with NLS-α-CE being the most potent inhibitor abolishing E . coli growth completely at concentrations as low as 2 μM . This peptide also induced flocculation of E . coli cultures , as noted by the higher initial OD ( see S2 Fig ) . Both α-NLS-C and the chimera peptide showed partial growth inhibition at 2 μM and 50% growth inhibition at 16 μM . However , α-NLS-C activity showed large variation , particularly between different plates , an indication of high sensibility to initial conditions ( see S2 Fig ) . Surprisingly , coupling the NLS sequence to α-factor without optimization was sufficient to yield an antibacterial activity providing partial growth inhibition at 4 μM and 50% inhibition at 16 μM . To test whether α-factor was important for AMP activity we also tested pure α-factor , pure NLS , the optimized NLS fitting CPP rules ( NLS-CE ) and the Iztli-1 peptide , a previously reported peptide with an embedded α-factor specifically designed to have antibacterial activity [25] . We observed that pure α-factor and NLS had no antibacterial activity , whereas NLS-CE yielded a partial growth inhibition of E . coli . The Iztli-1 peptide showed a very similar E . coli growth inhibition as NLS-α , probably due to the slightly positive charge of its N-terminal tail , resembling NLS . This indicates that the antibacterial activity observed in NLS-α-CE , α-NLS-C and the chimera peptide is a synergistic effect between the aggregation of NLS , α-factor and the CPP optimization . A linear trend could be observed between concentration and growth inhibition ( regression lines with standard deviations shown in Fig 6 ) . We used this linear adjustment to predict and test the required concentrations to completely inhibit growth in E . coli . The minimal inhibitory concentrations for E . coli are reported in Table 2 ( growth curves are shown in S4 Fig ) . It is noteworthy to observe that the NLS-α-CE peptide is a more potent antibiotic than Ampicillin . We studied the structural properties of the peptides in vitro , using circular dichroism ( see Materials and Methods ) . The far-UV CD spectra for NLS-CE , NLS-α and NLS-α-CE in water matched the characteristics of a random coil conformation . These spectra had a minimum of 195 nm for NLS-α , 197 nm for NLS-α-CE and 199 nm for NLS-CE ( Fig 7A–7C ) . The presence of tryptophan in these peptides resulted in a positive band between 220 to 235 nm . Similar results were observed in 50% TFE , which confirmed the unstructured nature of these peptides in this solvent . The CD spectra of NLS-α and NLS-α-CE peptides with 50% TFE displayed a decrease in negative ellipticity , suggesting that the TFE promoted aggregation of the peptides ( Fig 7A and 7B ) . The analysis of secondary structure content showed nearly identical fractions of α-helices , β-sheets and random-coil of peptides in water and in 50% TFE ( Table 3 ) . Interestingly , the CONTIN/LL estimates 30 to 35% content of β-sheets regardless of the solvent . The CD spectrum of the chimera peptide in water is characterized by a strong minimum shifted to 204 nm . A conformational change to a more ordered structure could be induced in the presence of 50% TFE . The positive maximum at 192 nm and a negative minimum at 207 nm together with a weak increase of negative ellipticity around 222 nm , suggested the presence of a mixture of α-helix and β-sheet in equilibrium with random coil conformation . Deconvolution of the CD spectrum of chimera in 50% TFE indicated some gain in α-helical content ( from 8 to 24% ) , with the concomitant decrease in both the β-sheet and random-coil contents ( Table 3 ) . Thus , none of the peptides showed a strong tendency towards a particular structure in water , and only the chimera peptide showed a slight tendency towards an α-helical structure in TFE .
We have presented a new strategy to design multifunctional peptides . Three multifunctional peptides were designed , chemically synthesized and structurally and biologically tested . The basic idea of our method is that proteins share particular activities ( e . g . CPP , AMP and DNA binding ) that are compatible , and consequently they can be placed together in multifunctional peptide sequences . This functional compatibility can be predicted at the protein sequence level by training a probabilistic classifier . In this work , we trained a classifier with CPPs and non-CPPs and discovered that both DNA-binding and AMP activities are compatible with CPP rules ( see S1 Fig and S1 Table ) . To validate this prediction , we tested if CPP rules could be used to design peptide sequences with AMP and DNA-binding activities and we showed that it is possible to do so . An important aspect of our designs is that two additional activities were incorporated ( pheromone and nuclear localization ) within these multifunctional peptides and we showed that these additional activities were retained , indicating that compatible functions enable multifunctional and moonlight polypeptides . As such , our approach may extend previous computational methods that effectively predict peptides with CPP activity [22–24] to effectively design multifunctional peptides [20–21] . Here , it became apparent that it is possible to gain antibacterial activity when mimicking CPPs . Interestingly , the antibacterial activity could be gained with the optimization of a single domain ( as shown by NLS-CE ) but was much stronger when including the α-factor sequence into the entire peptide sequence . As demonstrated by Iztli-1 and NLS-α , α-factor seems to have an intrinsic ability to be transformed into an antibacterial peptide by adding positive residues to it . For instance , embedding the α-factor within an AMP domain maintained the pheromone activity ( see Fig 5 ) and previous results showed that this is true for concentrations as low as 10μM [25 , 26] . Embedding the α-factor in any other multifunctional peptide optimized for CPP also maintained the pheromone activity ( see Fig 5B , 5C , 5E and 5F ) , yet at higher concentrations . This observation was true independent of the location ( N or C terminus ) of the extra residues added to the α-factor . These results suggest that the structure of the α-factor pheromone matches better the AMP rules than the CPP rules . Considering the known function of the α-factor to partially induce cell death on yeast cells , it may indeed be considered that the α-factor is more related to AMP activity than to CPP activity and consequently , its presence within multifunctional peptides optimized for CPP rules may explain the gain in AMP activity [37 , 38] . Additionally , any peptide sequence optimized to fit CPP rules ( NLS-CE , α-NLS-C , NLS-α-CE and chimera ) also showed effective DNA-binding activity ( see Fig 4A–4C ) . Thus , peptides optimized for CPP rules are compatible with DNA-binding activity , as expected from our analysis on the PFAM database ( see S1 Fig and S1 Table ) . This makes a strong argument for multifunctional designs of antimicrobial peptides since it is not only possible to do so , but the antibacterial activity can be amplified by an order of magnitude when including a specific set of compatible domains . In terms of secondary structure , the α-helical structure is frequently found among AMPs as well as in DNA-binding domains [39 , 40] . An important difference is that for AMPs the helical region participates directly in the biological activity while for DNA-binding , the binding may or may not take place at the helix . None of our designed peptides presented a strong preference to adopt an α-helical structure neither in water nor in TFE ( see Table 3 and Fig 7 ) , suggesting that the functionality of the designed peptides does not depend on this particular structure . Instead these results suggest that the peptides we designed posses some conformational freedom characteristic of multifunctional proteins [41] . It is not the goal of this work to establish the structural basis of the activity of these peptides , and the secondary structure of these peptides in complex with DNA or membranes remains to be elucidated . In any case , our results provide additional evidence for the conformational freedom of multifunctional peptides . Comparing the efficiencies to penetrate cells , to kill E . coli cells or to bind DNA among the three designed peptides ( α-NLS-C , NLS-α-CE and chimera ) , it may be observed that each peptide perform similarly in most of these activities , but some differences were noticed . For instance , NLS-α-CE was clearly the best antibacterial peptide , but had similar penetrating activity in yeast or mammalian cells than α-NLS-C . On the other hand , chimera had the lowest antibacterial and penetrating activities . Concerning the design of these peptides , NLS-α-CE sequence was optimized to match efficient CPP rules , while α-NLS-C was optimized to match CPP rules , without considering their efficiency . On the other hand , the chimera peptide included optimization to match membrane-binding proteins . Thus , these results suggest that the optimization to match efficient CPPs rendered highly efficient CPP and AMP activities , while optimization for membrane binding proteins reduced these activities . Yet , other aspects of these peptides varied as well ( e . g . , the length of the peptides ) , thus further studies are required to explain this observation . Our results pose two interesting questions . From an evolutionary perspective , compatible functions may facilitate a change of function or moonlighting activity . As such , there might be an evolutionary connection between compatible functions and multifunctional proteins , and such a relationship deserves to be further investigated . From a pharmacological perspective , it has been recognized that a shift from the one-target drug model to a multi-target paradigm is required , considering that drugs with clinical effects often have multiple targets [42] . The method we developed here might facilitate the design of poly-pharmacological peptides . Thus , it will be useful to investigate further the impact of creating this type of multifunctional peptides for treating complex diseases . In summary , we introduce a conceptual framework to design multifunctional CPPs based on the observed compatibility of certain peptide activities . We also present a design strategy and experimental evidence for three new multifunctional peptides that combine cell penetration , antimicrobial , pheromone and DNA binding activities . Our software is available at http://bis . ifc . unam . mx/en/software/dcf .
The bacterial strain used in this study was Escherichia coli DH5α ( ( F- endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoR nupG ϕ80dlacZΔM15Δ ( lacZYA-argF ) U169 , hsdR17 ( rK- mK+ ) , λ– ) ) . For the pheromone signaling we used the Saccharomyces cerevisiae BY4741 strain ( MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 ) . The mating pheromone α-factor from S . cerevisiae was purchased from Sigma-Aldrich ( catalog number T6901 ) . Anaspec , Inc . produced the Iztli peptide 1 ( Iztli-1 ) . The company verified the purity of these peptides using High-performance liquid chromatography ( HPLC ) and mass spectrometry ( data not shown ) . Optimized Fmoc and Boc methodologies were employed for peptide syntheses with free N- and C- termini . Delivered peptide powders were first dissolved in sterile ddH2O and their actual concentrations were determined at 280 nm using a Nanodrop apparatus ( Thermo Scientific , USA ) . A small fraction was subsequently diluted further to yield 200μM stock solutions kept at -80°C . All peptides used in this study included only L-amino acids . AnaSpec Inc also synthesized the TAMRA-labeled versions of the peptides used in this study . We purified these peptides using HPLC and verified their masses by mass spectrometry . The purity of these peptides was estimated as follows: 1 ) α-NLS-C: >95% , 2 ) α-NLS-C woRK >95% , 3 ) chimera >80% , 4 ) chimera woRK >50% , 5 ) NLS-α-CE >80% and 6 ) NLS-α-CE woRK >80% . Classification was performed on a compound data set of 1267 confirmed CPP sequences and 1267 random peptides from SWISS-PROT using the classification algorithms from alglib and dlib ( www . alglib . net , www . dlib . net ) . For CPP and CPP efficiency classification we used data sets obtained from http://www . imtech . res . in/raghava/cellppd/dataset . php . In particular , the data set for CPPs consisted of the CPPSite1 , Sanders and Dobchev sets [43] . The membrane-binding data set was obtained from a previously published list of trans-membrane protein fragments [44] . All training sets are also available in the Github repository at https://github . com/cdiener/dcf . For each of the sequences 27 physiochemical properties were calculated , consisting of the frequencies of each of the 20 amino acids and 7 additional properties . All of those additional properties were calculated across a sliding window of 8 amino acids and summarized by either the mean or the difference of maximum and minimum values . The additional properties were: mean charge , sliding window range of charge , hydrophobicity , isoelectric point , sliding window range of the hydrophobic moment , water-octanol partition coefficient and an approximation of the alpha-helical content in the sequence . Those properties were then used as the features for random forest and SVM classification . Probabilities for any new sequence to constitute a CPP could now be predicted from the trained models . For prediction of CPP efficiency and membrane-binding the models were trained in the same manner using data sets of highly efficient versus low efficiency CPPs and membrane-binding versus non membrane-binding domains , respectively . PFAM domains were analyzed by first obtaining version 27 . 0 of the PFAM database containing only the sequence parts used in family identification ( conserved regions ) . This data set was then split into the individual families and the CPP probabilities P ( CPP ) = P ( CPP|s ) were predicted for all individual sequences in the data set . Family-wise measures were obtained by calculating the mean probabilities to be a CPP , P ( CPP ) and mean probability to be a CPP and have high efficiency P ( CPP , eff ) = P ( CPP , eff|s ) for each family . Source code and all data for these classifications , can be found at http://bis . ifc . unam . mx/software/dcf ( DOI: 10 . 5281/zenodo . 30278 ) . Optimization of peptide fragments joining and flanking the desired domains was performed by a custom implementation of a Simulated Annealing algorithm in C++ called modes ( Multi-objective designer ) . New candidate sequences were generated by altering the current optimal sequence during each step of the optimization in one of three ways: amino-acid substitution , addition or deletion . Here , addition was performed by adding any of the valid amino acids in a random position in one of the fragments . Deletions were performed by randomly deleting one amino acid from each fragment . Substitution was performed by randomly mutating a position within a chosen fragment in accordance with BLOSUM80 substitution frequencies . Two strategies were employed to improve convergence . First , we implemented energy landscape pavement ( ELP ) to counteract trapping in local minima based on a histogram of previously visited energy values [45] . Second , we used a temperature schedule derived by varying the acceptance probability q in the edge case where one of the generated candidates has a low energy and all the other candidates a high energy: q=exp ( −Emin/T ) ( n−1 ) exp ( −Emin/T ) +exp ( −Emin/T ) ( 1 ) This acceptance probability was linearly varied between a value close to one and close to zero during the optimization , leading to an almost random acceptance in the beginning of the optimization and terminating by only allowing improvements to the solution in the final stage of optimization . The program source code together with the scripts to reproduce the conclusions and figures can be found at http://bis . ifc . unam . mx/software/dcf or https://github . com/cdiener/dcf ( DOI: 10 . 5281/zenodo . 30278 ) . 1μg of pGREG546 plasmid was pre-incubated with the required quantity of each peptide to result in a 1:1 or 1:4 peptide charge to DNA charge ratio . Samples were incubated for 30 minutes and loaded on a 1% Agarose gel in SB buffer . DNA affinity was quantified by measuring the mean pixel intensity of the DNA bands with Fiji ( http://fiji . sc ) and normalizing to the intensity range between an empty spot on the gel and the intensity of the pGREG546 plasmid without any added peptide . Analysis scripts for the EMSA experiments as well as the raw data can be found with the distributed source code at http://bis . ifc . unam . mx/software/dcf . Escherichia coli strain DH5α was grown for 12 hours to mid-log phase and diluted to 8·105 cells/ml by using the reference OD obtained by absorption spectroscopy . Each sample 100μL culture was treated with 50μL solution containing the respective peptide in the desired concentration or water in the case of the controls . Growth curves were measured on two 96 well half area plates with two replicates each followed for 24h using a plate reader ( Synergy , USA ) . In order to minimize plate-specific effects due to variations in loading time or initial cell numbers , different plates measurements were normalized by the time-point where the water control samples reached their half-maximal OD before combining . The respective minimum from each growth curve was subtracted from the normalized data from individual or combined plates . Areas under the curves were obtained by applying a linear interpolation to the OD curves , which was then integrated exactly . Growth rates were obtained by log-transforming the growth curves , identifying the exponential phase by the linear parts in that log-transformation ( OD600 between 0 . 05 and 0 . 5 ) and performing a linear regression in the exponential phase with the slope denoting the respective growth rate . All mentioned methodology was implemented in R ( www . r-project . org ) and is available as open source at http://github . com/cdiener/dcf . Using the slope of the linear regressions determined in these experiments , we estimated the minimal inhibitory concentration ( MIC ) for E . coli of our 3 CPP designed peptides ( α-NLS-C , NLS-α-CE and chimera ) and tested it experimentally . Pheromone signaling induction in Saccharomyces cerevisiae was measured by induction of eGFP fluorescence in a BY4741 strain containing the Fus1-eGFP construct in place of the wild type Fus1 ( MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 fus1::eGFP HIS3MX6 , obtained from the yeast GFP collection ) [33] . The BY4741 fus1::eGFP strain was grown overnight in YPD to mid-log phase and diluted to 106 cells/mL . Samples were treated with the respective peptide concentrations and incubated for 3h before mixing 1:1 with low-diffusivity agarose suspended on a covered glass slide . Images were taken with an Olympus FV10 confocal microscope using a 60x objective with 2 . 0 optical aperture . Images were analyzed using Fiji ( www . fiji . sc ) . The only modifications applied to the images were separation of the bright field and GFP channels , adjustment of brightness and contrast to 50% and addition of scale bars . All 6 fluorogenic peptides ( TAMRA-peptides , see Table 1 ) were prepared to 600μM in water as a 10X stock solution . Although all the peptides were apparently soluble in water , they all formed aggregates under the microscope . Thus , 3 cycles of sonications ( sonicator Bransonic 32 ) for 3 min each were performed to dissolve these aggregates . BY4741 ( MATa ) cells were grown overnight and diluted to reach an OD600 = 0 . 3 . These cells were pelleted and suspended in water in a final volume of 1mL . 13 . 5μL of these cells were put on a 26 . 4x76 . 2x1 . 2 mm microscope slide ( LAUKA ) and 1 . 5μL of the stock solution for each peptide was added to a final volume of 15μL and covered with a 22x22x1 . 5 mm cover glass ( Thomas scientific ) . These samples were observed under a multi-photon confocal microscope ( Olympus FV1000 ) and pictures were recorded at 0 , 10 and 20 min after the peptide addition . All images were taken as z-stacks with 10–14 z-levels . Bright field images were chosen from the in-focus z-stack , whereas TAMRA images denote the average fluorescence across the entire z-stack to obtain the fluorescence inside the entire cell volume . Images were analyzed using Fiji ( http://fiji . sc/Downloads#Fiji ) . Primary Mouse Embryo Fibroblasts ( MEFs ) were prepared from CD-1 mouse embryos of 17–18 days of gestation obtained from the animal house of the Instituto de Fisiología Celular , UNAM , following IACUC guidelines . MEFs were maintained in high glucose DMEM ( 10569 , Gibco ) containing 10% FBS ( 16000 , Gibco ) , 5000 U penicillin/streptomycin ( 15140 , Gibco ) and plated at a density of 7x104 cells/well in 12-well culture plate . MEFs were incubated at 37°C with 6 different peptides to a final concentration of 60μM . The peptides used were: α-NLS-C , α-NLS-C woRK , NLS-α-CE , NLS-α-CE woRK , chimera or chimera woRK ( the woRK peptides are negative controls for internalization; see Table 1 ) . Cells were incubated with the peptides for 20 min , then washed with PBS and fixed with 4% paraformaldehyde for 30 min at room temperature . Afterwards , samples were washed with PBS three times and then incubated with DAPI for 2 min at room temperature . Images were obtained by fluorescence microscopy ( Nikon ECLIPSE Ti-U ) using the NIS Elements , Basic Research software , Version 3 . 13 . Circular dichroism spectra were recorded on a ChirascanTM CD Spectrometer ( Applied Photophysics , UK ) equipped with a peltier temperature controller . 0 . 3mg of each peptide was dissolved in 1mL of water or 50% aqueous trifluoroethanol and placed in quartz cuvettes of 0 . 1cm path length . Far-UV CD spectra were collected from 185 to 250 nm at 20°C . Spectra were averaged over 5 scans and the averaged blank spectra of solvents were subtracted . Ellipticity is reported as mean residue molar ellipticity and recorded in terms of molar elipticity [θ] ( deg cm2 dmol−1 ) . CONTIN/LL within the CDPro analysis software program [46] was used to estimate the secondary structure content of the peptides .
|
Most proteins and peptides in nature display multiple activities either by fusing different domains ( with different activities ) or by evolving multiple activities in a single domain . Understanding which activities may be combined to render multifunctional proteins remains an open question relevant to understanding the organization of living organisms and to improve the design of pharmacological peptides . To address this problem , we introduce the concept of compatible activities , that is , activities that may combine without losing any of these in a single polypeptide chain . To identify compatible activities in peptide sequences , we used a machine-learning approach and discovered that a penetrating activity should be compatible with DNA-binding and antimicrobial activities . To test if these activities may combine without any functional loss , we designed peptide sequences that harbor two independent activities ( nuclear localization and pheromone ) and experimentally showed that all our designed peptides display penetrability , pheromone , antimicrobial and DNA-binding activities , supporting the idea that multifunctionality may be achieved combining compatible activities .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"antimicrobials",
"sequencing",
"techniques",
"medicine",
"and",
"health",
"sciences",
"drugs",
"microbiology",
"dna-binding",
"proteins",
"optimization",
"mathematics",
"antibacterials",
"pharmacology",
"molecular",
"biology",
"techniques",
"chemical",
"synthesis",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"peptide",
"synthesis",
"proteins",
"biological",
"databases",
"molecular",
"biology",
"biosynthetic",
"techniques",
"biochemistry",
"sequence",
"databases",
"peptides",
"database",
"and",
"informatics",
"methods",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"pheromones",
"physical",
"sciences"
] |
2016
|
Effective Design of Multifunctional Peptides by Combining Compatible Functions
|
Sugars are important nutrients for many animals , but are also proposed to contribute to overnutrition-derived metabolic diseases in humans . Understanding the genetic factors governing dietary sugar tolerance therefore has profound biological and medical significance . Paralogous Mondo transcription factors ChREBP and MondoA , with their common binding partner Mlx , are key sensors of intracellular glucose flux in mammals . Here we report analysis of the in vivo function of Drosophila melanogaster Mlx and its binding partner Mondo ( ChREBP ) in respect to tolerance to dietary sugars . Larvae lacking mlx or having reduced mondo expression show strikingly reduced survival on a diet with moderate or high levels of sucrose , glucose , and fructose . mlx null mutants display widespread changes in lipid and phospholipid profiles , signs of amino acid catabolism , as well as strongly elevated circulating glucose levels . Systematic loss-of-function analysis of Mlx target genes reveals that circulating glucose levels and dietary sugar tolerance can be genetically uncoupled: Krüppel-like transcription factor Cabut and carbonyl detoxifying enzyme Aldehyde dehydrogenase type III are essential for dietary sugar tolerance , but display no influence on circulating glucose levels . On the other hand , Phosphofructokinase 2 , a regulator of the glycolysis pathway , is needed for both dietary sugar tolerance and maintenance of circulating glucose homeostasis . Furthermore , we show evidence that fatty acid synthesis , which is a highly conserved Mondo-Mlx-regulated process , does not promote dietary sugar tolerance . In contrast , survival of larvae with reduced fatty acid synthase expression is sugar-dependent . Our data demonstrate that the transcriptional network regulated by Mondo-Mlx is a critical determinant of the healthful dietary spectrum allowing Drosophila to exploit sugar-rich nutrient sources .
Mono- and disaccharides , i . e . sugars , are an important source of nutritional energy , but animal species display marked differences in the degree of sugar utilization and tolerance . While the diet of carnivores is typically low in sugars , nectarivores , like hummingbirds , feed primarily on sugar-rich nectar [1] , [2] . Sugars from fruits and honey have been part of the ancestral human diet . However , the large quantities of refined sugars consumed by modern humans far exceed those available in natural sources [3] . In fact , it has been proposed that excessive added sugar in the diet , especially fructose , might contribute to the development of metabolic syndrome [3]–[5] . Yet the genetic factors governing the delicate balance between healthful dietary sugar utilization and the sugar overload-induced metabolic disturbance are poorly understood . Drosophila is a well-suited model for exploring the physiological consequences of sugar intake . Drosophila melanogaster is a generalist fruit breeder naturally performing well on a broad range of dietary sugars [6] . However , excessive intake of sugars has been shown to cause diabetes-like metabolic changes in D . melanogaster , including insulin resistance , elevated circulating glucose and increased adiposity [7] , [8] . Dietary sugars have also been shown to shorten Drosophila lifespan [9] . The sugar-induced insulin resistance has been attributed to the JNK-regulated lipocalin Neural Lazarillo [8] . Moreover , high sugar induced gene expression has been previously analysed [7] , [10] . However , beyond these observations , the functional interactions between genotype and dietary sugar have remained poorly understood . Elevated systemic glucose levels cause cellular stress and tissue damage [11] , [12] . Animals therefore rapidly adapt their metabolism to fluctuating sugar intake , maintaining circulating glucose levels constant . A postprandial increase in circulating glucose triggers the release of insulin , which induces the rapid uptake of excess glucose by metabolic tissues including muscle , adipose tissue , and liver [13] . Intracellular glucose is immediately converted into glucose-6-phosphate and further metabolized into glycogen and lipids or catabolised to release energy . Metabolic tissues are exposed to large variations in the flux of intracellular glucose and therefore need to regulate their metabolism accordingly . The basic helix-loop-helix transcription factor paralogs ChREBP ( Carbohydrate Response Element Binding Protein ) and MondoA act together with their common binding partner Mlx ( Max-like protein X ) to mediate transcriptional responses to intracellular glucose in mammals [14] . The ChREBP/MondoA-Mlx complex is activated by glucose-6-phosphate and other phosphorylated hexoses , and regulates gene expression by binding to target promoters containing a carbohydrate response element ( ChoRE ) [15]–[19] . ChREBP and MondoA regulate the majority of the global glucose-induced transcriptional responses and many of their target genes encode enzymes in glycolytic and lipogenic pathways [15] , [16] , [20] , [21] . ChREBP and MondoA play differential tissue-specific roles in mammals: ChREBP functions in the liver , adipose tissue and pancreatic beta cells [22]–[26] , while MondoA is predominantly expressed in the skeletal muscle [27] . Of the mammalian ChREBP/MondoA-Mlx complex , the role of ChREBP has been studied in a physiological setting using loss-of-function mice . While ChREBP is nonessential in terms of survival , the mutant mice display a number of metabolic phenotypes , including elevated plasma glucose and liver glycogen as well as reduced adiposity [25] , [28] . ChREBP−/− mice survive poorly on a diet with high levels of sugars , but the underlying reasons have remained unexplored [25] . ChREBP is known to regulate a range of metabolic genes , including those involved in de novo lipogenesis [15] , [21] . Which target genes contribute to the various physiological phenotypes and what is the causal interrelationship between the physiological phenotypes , are questions that require powerful genetics and are therefore challenging to address in vivo in mammals . Moreover , existence of another Mondo paralog , MondoA , might mask some phenotypes in the ChREBP−/− mouse . To better understand the physiological roles of the Mondo/ChREBP/-Mlx complex and its target genes , we have explored their role in Drosophila melanogaster . The Drosophila genome encodes one ortholog for each of ChREBP/MondoA and Mlx , which we call Mondo ( alternative identifiers: CG18362 , Mlx interactor , ChREBP ) and Mlx ( alternative identifiers: CG3350 , Bigmax ) , respectively [27] , [29] , [30] . We have generated mlx null mutant flies , which displayed lethality in the late pupal stage . D . melanogaster larvae can normally utilize high levels of dietary sugars [6] , but loss of Mlx or knockdown of Mondo caused striking intolerance towards sucrose , glucose and fructose . The mlx null mutant larvae also displayed extensive metabolic changes , with strongly elevated circulating glucose , signs of amino acid catabolism and altered lipid and phospholipid profiles . Systematic functional analysis of Mlx-regulated genes revealed three genes contributing to dietary sugar tolerance: cabut , encoding a Krüppel-like transcription factor , phosphofructokinase 2 , a regulator of the glycolytic pathway , and Aldehyde dehydrogenase type III , which is linked to detoxification of reactive aldehydes .
To study the physiological role of Drosophila Mlx , we generated a mutant allele using imprecise P-element ( P{XP}bigmaxd07258 ) excision . We recovered one mutant allele , mlx1 , which lacked the entire coding region of mlx as well as 17 C-terminal amino acids of the neighbouring gene CG3368 ( Figure 1A ) . As controls , we recovered lines from which the P-element had been excised precisely , leaving mlx intact . If not differently stated , the precise excision line is used as a control throughout the study . As expected , mlx1 mutant larvae expressed neither Mlx protein nor mRNA ( Figure 1B; Figure S1A ) . The mlx1 mutants displayed lethality at the late pupal ( pharate ) stage ( Figure S1B ) , and only a small number of adult flies could be recovered . mlx1 mutant flies also displayed a modest developmental delay when raised on our standard fly food ( Figure S1C ) . Experiments with defined nutrients revealed that mlx1 mutant larvae failed to survive on a diet with 20% sucrose as the sole nutrient source ( Figure 1C ) . To test if the mutant lethality was due to either the inability to utilize carbohydrates as energy source or intolerance towards sucrose , we supplemented protein-rich diet ( 20% yeast ) with increasing levels of sucrose . While mlx1 mutant larvae developed with similar kinetics as control animals on a high protein/low sugar diet , increasing the sucrose concentration gradually slowed down larval development of mlx1 mutants ( Figure 1D , E ) . At higher sucrose levels , mlx1 mutants failed to pupate and died as larvae , while control animals displayed no apparent change in pupation kinetics with respect to 0–15% sucrose . To confirm that the observed phenotypes were due to loss of mlx function , we used RNAi-mediated knockdown and transgenic rescue . Ubiquitous knockdown of Mlx by RNAi led to significantly slower pupation , and increased pupal lethality on protein rich food supplemented with 15% sucrose , while displaying no visible phenotype in the absence of added sucrose ( Figure 1F ) . Driver line without RNAi was used as a control . Moreover , sugar intolerance and pupal lethality of the mlx1 mutants were efficiently rescued by ubiquitous expression of transgenic mlx ( tub-GAL4>UAS-mlx ) ( Figure 1G , 1H ) . To further rule out the possibility that sugar intolerance was due to disturbed function of the neighbouring gene CG3368 , we used RNAi for ubiquitous knockdown . CG3368 was efficiently silenced , but no sugar intolerance was observed ( Figure S1D–S1F ) . Thus , mlx gene function is essential for tolerance to dietary sucrose . To test for specific intolerance towards glucose or fructose , we supplemented the protein-rich food with 10% of either monosaccharide . Both caused clear developmental delays of mlx1 mutants ( Figure 1I ) . Drosophila melanogaster is a dietary generalist , feeding on micro-organisms on decaying fruits and vegetables that have varying sugar content . To test whether the sugar intolerance of mlx1 mutants was relevant within the natural spectrum of D . melanogaster's diet , we allowed larvae to develop on pieces of red grape with baker's yeast inoculum . Indeed , mlx1 mutants were unable to pupate in these conditions , while >50% of the control larvae reached the pupal stage ( Figure 1J ) . In mammals , Mlx forms a functional complex with the Mondo paralogs , MondoA and ChREBP . We tested if the heterodimeric function of Mlx is conserved in Drosophila and essential for the sugar tolerance . Drosophila Mlx co-immunoprecipitated with Mondo when expressed in Drosophila S2 cells , suggesting heterodimeric function ( Figure 2A ) . Notably , Mlx from S2 cells runs as two distinct bands ( Figure 2A ) , which correspond to the two upper bands present in the in vivo sample ( Figure 1B and Figure S2 ) . The nature of these bands has remained unclear , no alternative splicing has been reported and both bands are resistant to alkaline phosphatase treatment ( not shown ) . Ubiquitous RNAi knockdown of Mondo ( tub-GAL4>Mondo RNAi ) led to delayed pupation ( Figure 2B ) and reduced pupal survival on high sugar diet ( 20% yeast-15% sucrose ) ( Figure 2C ) . In sum , the biochemical and genetic evidence implies conservation of the Mondo-Mlx heterodimer function in Drosophila . The diet-dependent phenotype of mlx1 mutants led us to perform a comprehensive survey of their metabolic status using mass-spectrometry based lipidomics and metabolomics . Lipidomics analysis revealed significant downregulation of key phospholipid groups , such as phosphatidylethanolamines ( PE ) and lysophosphatidylcholines ( LysoPC ) ( Figure 3A ) . Total triglyceride ( TG ) levels showed a lower trend in mlx1 mutants , but the difference to the controls was not statistically significant ( Figure 3B ) . However , mlx1 mutants showed significant enrichment in triglyceride species with long fatty acid tails ( Figure 3B ) . At the same time , mlx1 mutants showed strong downregulation of certain fatty acids , such as myristoleic acid and lauric acid ( Figure 3C ) . On the other hand , ceramide ( Cer ) levels were elevated in mlx1 mutants compared to controls ( Figure 3D ) . Together , mlx1 mutants display signs of severely altered lipid and phospholipid metabolism . In addition to the changes in lipid profiles , total amino acid levels were significantly reduced in mlx1 mutants ( Figure 3E ) , while concentration of urea was dramatically increased ( Figure 3F ) . This implies that mlx1 mutants might catabolise amino acids for energy . To study changes in glucose metabolism , we measured glucose and trehalose levels in the hemolymph of larvae raised on diets with varying sucrose content . Trehalose is a disaccharide released by gluconeogenesis and glycogenolysis in insects [31] . The levels of circulating glucose were moderately elevated in mlx1 mutant larvae raised on a low-sugar diet ( 20% yeast ) ( Figure 4A ) . However , increasing the dietary sucrose to 5% , which still sustains larval development of mlx1 mutants , led to a prominent increase of circulating glucose in mlx1 mutants while remaining constant in control animals ( Figure 4A ) . Trehalose levels were also significantly elevated in mlx1 mutants , but unlike glucose , trehalose levels were little affected by the dietary sucrose levels ( Figure 4B ) . Furthermore , glycogen levels were significantly elevated in mlx1 mutants ( Figure 4C ) , indicating that glucose catabolism , not cellular glucose intake , is limiting glucose clearance from circulation in mlx1 mutants . To verify that the elevated glucose levels were due to loss of Mlx function , we performed a transgenic rescue , which normalized circulating glucose levels ( Figure S3A ) . Further , RNAi-mediated knockdown of Mlx led to a clear increase in circulating glucose , trehalose and glycogen ( Figure S3B–S3D ) . In line with the expectation of heterodimeric function for Mondo and Mlx proteins , Mondo RNAi knockdown led to a prominent increase in circulating glucose and trehalose ( Figure 4D , E ) . Also the glycogen levels were significantly increased in Mondo RNAi larvae ( Figure 4F ) . Knockdown of CG3368 , the neighbouring gene of mlx , had no influence on circulating glucose or trehalose ( Figure S3B , S3C ) . In conclusion , mlx null mutant and mondo knockdown larvae display a reduced capacity to utilize circulating glucose leading to poor homeostatic adaptation to elevated dietary sugar . As mammalian MondoA-Mlx and ChREBP-Mlx appear to display tissue-specific functions , we wanted to determine the functionally important tissues for Drosophila Mondo-Mlx . We first analyzed their mRNA expression using quantitative RT-PCR . Expression of Mondo and Mlx were highly correlated; highest levels were detected in the fat body , gut and Malpighian tubules ( Figure 5A ) . To functionally dissect the contribution of different tissues to the sugar sensitivity , we used tissue-specific transgenic rescue . Restoring Mlx expression in neurons ( Elav-GAL4 ) or muscle ( Mef2-GAL4 ) did not significantly improve the sugar tolerance or survival of mlx1 mutants ( Figure 5B ) . However , targeted expression in the fat body with two independent driver lines ( ppl- and r4-GAL4 ) efficiently rescued survival on high sugar diet . Moreover , rescue of Mlx in the fat body , but not in muscle , was sufficient to normalize the levels of circulating glucose in mlx1 mutants ( Figure 5C ) . Notably , transgenic Mlx expression by tub-GAL4 ( Figure 1G ) or Mef2-GAL4 ( Figure 5B ) caused moderately reduced survival , which was also observed in the presence of intact endogenous mlx ( Figure S4 ) . To identify Mlx target genes , we performed a microarray gene expression profiling specifically in the fat body . Comparing gene expression between mlx1 mutant and control fat bodies from 3rd instar prewandering larvae raised on a moderate sucrose level diet ( 20% yeast-5% sucrose ) revealed 97 down- and 96 up-regulated genes ( >2-fold change and adjusted p-value<0 . 05 ) ( Figure 6A; Table S1 ) . As expected for a deletion mutant , mlx was identified as the most strongly downregulated gene on the microarray ( Figure 6A ) . Gene Set Enrichment Analysis ( GSEA ) revealed a significant enrichment of KEGG categories involved in metabolic regulation ( Figure 6B ) . For example , KEGG categories of fatty acid metabolism and nitrogen metabolism were strongly downregulated ( Figure 6C ) , which is in good agreement with the metabolomics data ( Figure 3 ) . Many of the genes downregulated in mlx1 mutant fat body also showed reduced expression during earlier larval stages in whole larval samples ( Figure 6D ) . Mlx-regulated genes include several key metabolic genes , such as glycerol-3-phosphate dehydrogenase-1 ( Gpdh , CG9042 ) , stearoyl-CoA 9-desaturase-1 ( desat1 , CG5887 ) , Glutamine synthetase-1 ( Gs1 , CG2718 ) , and 3-hydroxybutyrate dehydrogenase ( sro , CG12068 ) . The mean expression levels of three known ChREBP and MondoA targets in mammals , fatty acid synthase ( Fas , CG3523 ) , acetyl-CoA carboxylase ( ACC , CG11198 ) and phosphofructokinase 2 ( PFK2 , 6-phosphofructo-2-kinase/fructose-2 , 6-bisphosphatase , CG3400 ) , [20] , [32] , were also reduced , although they did not pass our strict microarray cut-offs ( Figure 6D ) . Interestingly , one of the most strongly downregulated genes in mlx1 mutant fat body was that encoding the Krüppel-like transcription factor Cabut ( cbt; CG4427 ) ( Figure 6A , 6D ) . To test in an unbiased way if any of the genes downregulated in mlx1 mutants had an essential role in maintaining organismal sugar tolerance , we systematically targeted 103 candidate genes by RNAi ( Table S2 ) . Intriguingly , ubiquitous knockdown of two Mlx-regulated genes identified on the microarray led to significant sugar intolerance . Transcription factor Cabut was among the most highly Mlx-regulated genes in the microarray . Ubiquitous knockdown of Cabut expression by RNAi during the larval stage caused a modest delay of pupation on low sugar diet . However , on high sugar diet ( 20% yeast-15% sucrose ) Cabut knockdown led to prominent developmental delay and impaired survival ( Figure 7A ) . This suggests that Mondo-Mlx activates a hierarchical transcriptional network to regulate dietary sugar tolerance with Cabut as an essential downstream effector . Another Mlx-regulated gene , which caused sugar intolerance upon ubiquitous knockdown , was Aldehyde dehydrogenase type III ( Aldh-III , CG11140; Figure 7B ) . Most Aldh-III knockdown animals reached the pharate stage on a low sugar diet , but died during early pupal stages on a high-sugar diet ( Figure S5A ) . Similar early pupal lethality was observed in mlx1 mutants on sugar concentrations that allowed pupation ( data not shown ) . Survival on a 20% sucrose-only diet was also significantly reduced upon Aldh-III knockdown ( Figure 7C ) . We also tested whether restoring Aldh-III activity by transgenic expression would be sufficient to rescue impaired mlx1 mutant survival on high sugar diet . While transgenic expression of Aldh-III did not rescue mlx1 mutant pupation on 20% yeast-15% sucrose diet ( data not shown ) , larval survival of mlx1 mutants on 20% sucrose-only diet was significantly improved by transgenic Aldh-III ( Figure 7D ) . The same was true when Aldh-III expression was rescued only in the fat body ( Figure S5B ) . In conclusion , Aldh-III is essential and sufficient for providing dietary sugar tolerance . We also explored whether the sugar intolerance observed upon knockdown of Cabut and Aldh-III was associated with elevated circulating glucose levels . Surprisingly , knockdown of either gene did not result in a significant increase in circulating glucose ( Figure 7E , 7F ) . This implies that impaired clearance of circulating glucose is not an essential prerequisite for intolerance to dietary sugars . Instead , these two Mondo-Mlx-regulated parameters can be uncoupled at the level of the downstream target genes . Based on these phenotypes , the Cabut-dependent branch of the transcriptional network mediates only a subset of Mondo-Mlx functions . We also studied the possibility that Cabut could be a direct regulator of Aldh-III , however the mRNA levels of Aldh-III were unchanged in Cabut RNAi larvae ( Figure S5C ) . Also the mRNA levels of Mondo and Mlx were unchanced in Cabut RNAi larvae , indicating that Cabut is not a feedback regulator of Mondo-Mlx . Notably , it is possible that Cabut is regulating a subset of common genes with Mondo-Mlx . Perhaps the best established function of mammalian ChREBP-Mlx is promotion of de novo lipogenesis in response to high carbohydrate intake [14] . This function is mediated through upregulation of lipogenic genes and it appears to be conserved in Drosophila ( Figure 6D; [33] ) . Two key targets of Mondo-Mlx involved in de novo lipogenesis are acetyl-CoA carboxylase ( ACC ) and fatty acid synthase ( Fas ) . Thus exploring their function in respect to dietary sugar will reveal whether de novo lipogenesis is functionally coupled to sugar tolerance . While knockdown of ACC was embryonic lethal ( data not shown ) , Fas knockdown animals displayed some degree of survival until pupal stage . Strikingly , Fas knockdown larvae displayed early larval lethality on high protein diet ( 20% yeast paste ) , but diet supplementation with 15% sucrose partially rescued the lethality allowing pupation ( Figure 7G ) . Thus , Mondo-Mlx-mediated regulation of fatty acid synthesis is a non-essential function for dietary sugar tolerance . In contrast , high dietary sugars promote survival upon compromised fatty acid synthesis . As our systematic analysis of sugar tolerance genes did not reveal a mechanism by which Mondo-Mlx maintains low circulating glucose , we analyzed glucose levels on additional Mondo-Mlx targets that have a metabolic function . Phosphofructokinase 2 ( PFK2 , Pfrx , CG3400 ) synthesizes and breaks down fructose-2 , 6-bisphosphate , which is an allosteric activator of phosphofructokinase 1 and hence promotes glycolysis . PFK2 expression was downregulated in mlx1 mutants ( Figure 6D ) . Knockdown of PFK2 led to elevated circulating glucose , showing that PFK2 is a Mondo-Mlx downstream target , which contributes to circulating glucose levels ( Figure 7H ) . Interestingly , PFK2 knockdown also reduced pupation on high sugar diet ( 20% yeast-15% sucrose ) ( Figure 7I ) , implying that Mondo-Mlx-mediated activation of the glycolytic pathway contributes to dietary sugar tolerance .
Our study demonstrates that interfering with the function of the Mondo-Mlx complex severely affects Drosophila energy metabolism , rendering animals highly intolerant to sugars in their diet . The sugar intolerance is likely due to a combined effect of several downstream effectors , since our systematic loss-of-function analysis revealed three Mlx target genes that are essential for survival on high sugar diet . Sugar tolerance is influenced by glycolysis , which also contributes to clearance of glucose from circulation . However , circulating glucose levels and sugar tolerance are phenotypes that can be uncoupled , as in the case of two Mlx targets , Cabut and Aldh-III , which only contribute to sugar tolerance . Mondo-Mlx shows a high degree of functional conservation between flies and mammals , as orthologs of many Drosophila Mlx-regulated genes are known targets of ChREBP/MondoA-Mlx in mammals . These include glycerol-3-phosphate dehydrogenase-1 , stearoyl-CoA 9-desaturase-1 , fatty acid synthase , acetyl-CoA carboxylase , and phosphofructokinase 2 [15] , [21] , [25] . Drosophila Mlx displayed an essential role in the fat body , which is the counterpart of mammalian adipose tissue and liver . Thus , it is conceivable that the liver and adipose tissue-specific ChREBP-Mlx , instead of the muscle-specific MondoA-Mlx , represents the ancestral function of the heterodimer . This study identifies Aldh-III as a novel gene contributing to dietary sugar tolerance . Aldh-III is the ortholog of mammalian Aldh3 family . Aldehyde dehydrogenases ( Aldhs ) are highly conserved NAD ( P ) + -dependent enzymes that oxidize aldehydes to the corresponding carboxylic acid and act on a broad range of substrates [34] . Aldehydes are highly reactive compounds forming adducts with nucleic acids and proteins , thus disturbing cellular functions . Reactive aldehydes can originate from exogenous sources or be products of cellular metabolism . Aldhs have been shown to provide protection against a number of ectopic stresses , including toxic chemicals , heat stress , and UV irradiation [34]–[37] . One of the best-established functions of Aldh proteins is neutralization of acetaldehyde , a toxic metabolite of ethanol . Acetaldehyde elimination is mainly mediated by Aldh2 [34] , [38] . Polymorphism in the Aldh2 gene is common is Asian populations and it leads to poor ethanol tolerance [39] . Our finding that another member of the Aldh family , Aldh-III , is essential for dietary sugar tolerance is intriguing , since there are multiple parallels between the hepatic pathophysiologies related to excessive ethanol and fructose consumption in humans [3] . Of note , we observed that Aldh-III is sufficient in rescuing the sugar intolerance of mlx1 mutants in the fat body , the insect counterpart of liver , suggesting that metabolic stress-induced dysfunction of the fat body contributes to sugar intolerance . Future studies should be aimed at understanding the generation of sugar-derived reactive aldehydes and their molecular targets . We also provide evidence that PFK2 expression is positively regulated by Mondo-Mlx in Drosophila and that the PFK2 levels are critical in managing circulating glucose levels and providing sugar tolerance . Thus , our data suggests that regulation of PFK2 gene expression might be a suitable strategy to manage hyperglycemia in diabetes . Elevated PFK2 expression has also been associated with the high rate of glycolytic flux in neoplastic tumors [40] . Exploring the contribution of Mondo and Mlx proteins in this setting is therefore warranted . In addition to being regulated transcriptionally , PFK2 activity in mammals is known to be posttranslationally regulated by insulin signalling through protein kinase AKT [41] , [42] . It will be interesting to learn , what is the contribution of insulin signalling pathway activity on the dietary sugar tolerance . The finding that Cabut is an essential secondary effector of Mondo-Mlx is intriguing . In fact , cabut expression has been reported to respond to other metabolism-related signals: it is upregulated upon inhibition of TOR complex 1 signalling [43] , which is likely mediated by activation of the Forkhead transcription factor , FoxA [44] . While the developmental role of Cabut has been studied [45] , [46] , its metabolic functions have remained unexplored . Our finding showing that Cabut plays an essential metabolic role in providing dietary sugar tolerance implies this topic deserves an in-depth survey in the future . Notably , the closest mammalian homologs of Cabut , Klf-10 and Klf-11 , have been linked to metabolic regulation . Mutations in the klf-11 locus are associated with risk of diabetes [47] , while Klf-10 appears to negatively regulate lipogenic genes in hepatocytes [48] . klf-10 expression is regulated both by circadian signals [49] as well as by ChREBP upon high glucose [48] . Thus , Klf-10 might be the functional ortholog of Cabut . Future studies should be aimed at identifying the Cabut target genes involved in its metabolic functions . Dietary sugar tolerance displays a wide natural variation , even in closely related animal species . For example , two Drosophila species , D . melanogaster and D . mojavensis , have strikingly different tolerance to dietary sugars [6] . In contrast to the fruit generalist D . melanogaster , D . mojavensis is a cactus breeder , which does not naturally encounter high levels of simple sugars and displays poor survival on sugar-rich diet [6] . Based on our data , it is possible to hypothesize that Mondo-Mlx-regulated transcriptional network contributes to the natural variation in sugar tolerance . Genetic differences in sugar tolerance are also observed in humans . For example , hereditary fructose intolerance ( HFI ) is caused by mutations in the aldolase-B gene [50] . A fructose-restricted diet renders HFI relatively benign , but ingestion of fructose or sucrose leads to strong symptoms , including nausea and vomiting as well as a risk of liver and kidney damage . It will be important to explore whether genetic changes in the ChREBP/MondoA-Mlx network influence individual's risk for sugar overload-induced metabolic disturbance . This study highlights the usefulness of Drosophila as a model for systematically exploring the genetic factors defining the range of healthful nutrient intake . Notably , the sugar intolerance in mlx1 mutants is not a pleiotropic consequence of a generally disturbed energy metabolism . Knockdown of other key transcriptional metabolic regulators , such as SREBP , did not cause notable sugar intolerance ( unpublished observation ) . The availability of genome-scale reagents , including in vivo RNAi lines , offers the possibility for the systematic dissection of genes contributing to dietary sugar tolerance . These genes might include novel members of the Mondo-Mlx-regulated genetic network , but they will also uncover whether parallel regulatory pathways are involved . The Drosophila model is also particularly useful in dissecting the function of transcription factors that are master regulators of several gene groups that contribute to distinct physiological outputs . While this study has focused on uncovering those Mondo-Mlx targets that contribute to sugar tolerance and circulating glucose levels , uncovering the downstream effectors behind the other metabolic phenotypes of mlx1 mutant fly awaits for future studies .
P{XP}bigmaxd07258 were obtained from Bloomington stock center . For generating UAS-mlx flies , the coding region of mlx cDNA was amplified by PCR and cloned into pUAST vector using BglII and XbaI restriction sites . FLAG-tag was incorporated into the C-terminus . Aldh-III coding region was cloned into pUAST vector using NotI and XhoI restriction sites . RNAi lines were obtained from Vienna Drosophila RNAi Center and from NIG-FLY Stock Center . The following GAL4 driver lines were used in this study: tub-GAL4 [51] , ppl-GAL4 [52] , r4-GAL4 [53] , Elav-GAL4 [54] and Mef2-GAL4 [55] . In standard conditions flies were maintained at 25°C on medium containing agar 0 . 6% ( w/v ) , semolina 3 . 2% ( w/v ) , malt 6 . 5% ( w/v ) , dry baker's yeast 1 . 8% ( w/v ) , propionic acid 0 . 7% ( v/v ) and Nipagin ( methylparaben ) 2 . 4% ( v/v ) . For defined nutrient studies , larvae were grown on food containing 20% ( w/v ) dry baker's yeast , 0 . 5% ( w/v ) agar , and 2 . 5% ( v/v ) Nipagin ( methylparaben ) in PBS supplemented with varying concentrations of sucrose , glucose or fructose . 1st instar larvae were collected from apple juice plates ( apple juice 33 . 33% ( v/v ) , agar 1 . 75% ( w/v ) , sugar 2 . 5% ( w/v ) and Nipagin ( methylparaben ) 2 . 0% ( v/v ) ) and larvae were grown at controlled density ( 30 larvae per vial ) . For generating anti-Mlx antibodies full-length Drosophila mlx cDNA was cloned into pGEX-4T2 . Recombinant GST-Mlx was purified using Glutathione-agarose ( Sigma ) . Anti-Mlx antiserum was raised by immunizing a guinea pig ( Storkbio Ltd ) . Drosophila S2 cells were grown at 25°C in standard Shields and Sang M3 medium ( Sigma ) containing 2% of fetal bovine serum ( Gibco ) , 1× insect medium supplement ( Sigma ) and penicillin/streptomycin ( Gibco ) . The transfections were performed using Effectene ( Qiagen ) , according to manufacturer's protocol . Expression of transfected genes was induced with 1 . 2 µM CuSO4 24 h post-transfection . For detecting endogenous Mlx in vivo , 3rd instar prewandering larvae were homogenized in Laemmli sample buffer and boiled for 5 min . Samples were resolved on SDS-PAGE and detected by Western blotting using anti-Mlx antibodies . For the pulldown experiment , cells were lysed in IP lysis buffer ( 10 mM Tris-HCl pH 8 . 0 , 150 mM NaCl , 0 , 1% NP40 ) and lysates were cleared by centrifugation . Lysate protein concentration was adjusted to 1 µg/µl . 1 ml of lysate was incubated o/n with 25 µl Strep-Tactin beads ( IBA ) . The beads were washed 5 times with IP lysis buffer , after which Laemmli sample buffer was added and samples were boiled for 5 min . Pulldown and lysate samples were resolved on SDS-PAGE , transferred to nitrocellulose and analyzed by Western blotting using anti-V5 ( Invitrogen ) , anti-Mlx and anti-Kinesin ( Cytoskeleton ) . Metabolomics was performed using prewandering 3rd instar larvae grown on 20% yeast-5% sucrose diet . Analysis was done in four biological replicas . Data was processed using Guineu [56] and MZmine 2 [57] software packages for small polar metabolites and molecular lipids , respectively . Detailed description is available in Protocol S1 . Metabolite assays were done using prewandering 3rd instar larvae grown on 20% yeast or 20% yeast-5% sucrose diet . All analyses were done at least in four biological replicas . Glucose , trehalose and glycogen measurements were conducted as described [58] , [59] . Staged 3rd instar prewandering control and mlx1 mutant larvae were grown on a 20% yeast-5% sucrose diet . RNA was extracted from 3–4 larval fat bodies per sample in four biological replicas using the Nucleospin RNA XS kit ( Macherey-Nagel ) . The Amino Allyl MessageAmp II aRNA Amplification kit ( Ambion ) was used for aRNA synthesis . Hybridization to Agilent Drosophila Gene Expression Microarray , 4×44K was performed according to the manufacturer's instructions . Data normalization and analysis was performed using R , taking advantage of packages from the Bioconductor repository . A full documentation of the data analysis is available in Protocol S1 . For quantitative RT-PCR , the RevertAid H Minus First Strand cDNA Synthesis Kit ( Fermentas ) with random hexamer primers was used for first strand cDNA synthesis . PCR was performed using Maxima SYBR Green qPCR Master Mix ( 2X ) ( Fermentas ) and analyzed on StepOnePlus ( Applied Biosystems ) real-time PCR system . Primer sequences are available as Table S3 . All expression profiling data is available under accession E-MTAB-699 at the ArrayExpress repository . Statistical significance for each experiment ( excluding metabolomics and microarray ) was determined with unpaired Student's t-test with unequal variance . All quantitative data are presented as mean ± SEM for a minimum of three independent biological replicates .
|
Diet displays extreme natural variation between animal species , which range from highly specialized carnivores , herbivores , and nectarivores to flexible dietary generalists . Humans are not identical in this respect either , but the genetic background likely defines the framework for a healthy diet . However , we understand poorly the genetic factors that define the spectrum of healthy diet for a given species or individual . Here we have explored the genetic basis of dietary sugar tolerance of Drosophila melanogaster . D . melanogaster is a generalist fruit breeder that feeds on micro-organisms on decaying fruits and vegetables with varying sugar content . However , mutants lacking the conserved Mondo-Mlx transcription factor complex display striking intolerance towards dietary sucrose , glucose , or fructose . This is manifested in the larvae by the inability to grow and pupate on sugar-rich food , including red grape , which belongs to the normal diet of wild D . melanogaster . Larvae lacking Mondo-Mlx show widespread metabolic imbalance , including highly elevated circulating glucose . Genome-wide gene expression analysis combined with systematic loss-of-function screening of Mlx targets reveal that the genetic network providing sugar tolerance includes a secondary transcriptional effector as well as regulators of glycolysis and detoxification of reactive metabolites .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"animal",
"genetics",
"genetic",
"mutation",
"molecular",
"cell",
"biology",
"genetic",
"screens",
"gene",
"expression",
"genetics",
"biology",
"gene",
"networks",
"molecular",
"biology",
"genetics",
"and",
"genomics",
"gene",
"function"
] |
2013
|
Mondo/ChREBP-Mlx-Regulated Transcriptional Network Is Essential for Dietary Sugar Tolerance in Drosophila
|
Dengue virus ( DENV ) is a pathogen with a high impact on human health . It replicates in a wide range of cells involved in the immune response . To efficiently infect humans , DENV must evade or inhibit fundamental elements of the innate immune system , namely the type I interferon response . DENV circumvents the host immune response by expressing proteins that antagonize the cellular innate immunity . We have recently documented the inhibition of type I IFN production by the proteolytic activity of DENV NS2B3 protease complex in human monocyte derived dendritic cells ( MDDCs ) . In the present report we identify the human adaptor molecule STING as a target of the NS2B3 protease complex . We characterize the mechanism of inhibition of type I IFN production in primary human MDDCs by this viral factor . Using different human and mouse primary cells lacking STING , we show enhanced DENV replication . Conversely , mutated versions of STING that cannot be cleaved by the DENV NS2B3 protease induced higher levels of type I IFN after infection with DENV . Additionally , we show that DENV NS2B3 is not able to degrade the mouse version of STING , a phenomenon that severely restricts the replication of DENV in mouse cells , suggesting that STING plays a key role in the inhibition of DENV infection and spread in mice .
Viral infections have a vast impact on human health , resulting in hundreds of thousands of deaths yearly . To replicate and spread , these intracellular pathogens subvert the host cellular defense machinery . Dengue virus ( DENV ) is the most prevalent arbovirus in humans , and productively infects cells that are involved in the immune response , such as monocytes , B cells , macrophages and dendritic cells ( DCs ) among others [1] , [2] , [3] , [4] , [5] . Like most viruses , DENV has evolved in order to inhibit or evade different aspects of the innate immune system , the first line of human defense against microbes . DCs are antigen presenting cells ( APCs ) and some of the first cells that interact with the virus after the bite of an infected mosquito . Infection of these cells induces their activation , which results in their migration to the lymph nodes where the virus can infect other susceptible cells . The kinetics of infection of different cells in the immune system is not well documented , due to the lack of immune-competent mouse models for dengue disease . Nevertheless , in mice defective for type I IFN signaling , one of the most accepted current models for dengue disease , it has been shown that DCs and macrophages are productively infected by DENV [3] , [4] , [5] , [6] reviewed in [7] . DENV is a single stranded RNA virus of positive polarity that , after entering the cytoplasm of the host cell , releases its genome and synthesizes a polyprotein using the cellular machinery , as a first event of the viral cycle . The DENV polyprotein is cleaved by the viral protease complex ( NS2B3 ) and cellular proteases , including furin [8] . After this processing , some of the viral proteins have the ability to rearrange the ER membrane and create the micro-environment necessary for the production of de novo synthesized viral genomic RNA . During this event , DENV accumulates products with conserved molecular structures , like RNA with 5′-triphosphate moiety or double stranded RNA , also referred to as pathogen associated molecular patterns ( PAMPs ) . These foreign molecules are ligands of different cellular receptors engaged in their recognition , known as pattern recognition receptors ( PRRs ) . PRRs are mainly divided into two different classes depending on their localization , associated with either the membrane or the cytoplasm . The Toll-like receptor ( TLR ) family is composed of membrane proteins with domains that are designed to detect extracellular PAMPs . On the other hand , the cytosolic DExD/Hbox RNA helicase proteins that contain caspase-recruiting domains ( CARDs ) , referred to as RIG-I and MDA-5 , can detect specific PAMPs present in the cytoplasm . These last two cytoplasmic sensors together with the TLR family members ( TLR3/TLR7/TLR8 ) have been described so far as the most relevant DENV sensors [9] , [10] , [11] . After recognition of the mentioned PAMPs by the C-terminal helicase domain of RIG-I and MDA-5 , these undergo a conformational change that exposes their CARD domains and promote the interaction with different down-stream molecules . One of the most well studied down-stream molecules , referred as IPS-1 ( also known as , MAVS , CARDIF or VISA ) , is located in the outer membrane of the mitochondria and transmits the signal via different molecules , including the tumor necrosis factor receptor associated factors 6 and 3 ( TRAF6 and TRAF3 ) and the IκB kinase ( IKK ) family members ( TBK1 , IKKα , IKKβ and IKKε ) among other cellular factors [reviewed in [12]] . Recently three different groups , using cDNA library screening of genes that induced the IFNβ promoter , described an adaptor protein that localizes in the endoplasmic reticulum ( ER ) . This protein was named as stimulator of the interferon gene ( STING ) [13] , mediator of IRF3 activation ( MITA ) [14] and endoplasmic reticulum IFN stimulator ( ERIS ) [15] . Also the same protein , referred to as MYPS , was previously identified as a mediator of anti-major histocompatibility complex II monoclonal antibody-induced apoptosis in B-lymphoma cells [16] . STING is highly expressed in several immune cells , including macrophages and DCs , as well as endothelial and epithelial cells [13] . This protein can interact with RIG-I and IPS-1 , but not with MDA-5 , and the signaling mediated by this adaptor is independent of the sensing by the TLR family members [17] . In two recent reports , it was documented that STING is involved in the pathway that mediates the detection of pathogens with DNA genomes [18] and has a role as a direct sensor of cyclic di-nucleotides , a signaling molecule produced exclusively by bacteria and archea [19] . Activation of STING by some of these stimuli leads to its relocalization with TBK1 from the ER to perinuclear vesicles containing the subunit of the exocyst complex 5 ( Sec5 ) followed by the phosphorylation of TBK-1 and the subsequent activation of the transcription factors IRF3/7 and NFκB , which translocate to the nucleus and complex with ATF2/c-Jun to induce the expression of type I IFN and pro-inflammatory cytokines [17] . A remarkable hallmark of highly virulent human pathogens is the ability , acquired through evolution , to inhibit this innate immune response by the expression of viral factors that affect one or several steps of the above described signaling cascade . Some of the most notorious examples are the influenza virus NS1 protein , that targets RIG-I for degradation , minimizing the sensing of influenza virus PAMPs by this PRR [20] or the Hepatitis C virus NS34A protease complex that cleaves the adaptor IPS-1 to interrupt the signaling that ends with the activation of IRF3 , NFκB and the subsequent production of type I IFN in human hepatocytes [21] . Our group has documented that DENV is a weak inducer of type I interferon in human DCs , in particular when compared with other viruses that competently produce these cytokines in large amounts , such as Newcastle disease virus ( NDV ) [22] and Semliki Forest virus ( SFV ) [23] . This lack of type I IFN production by DCs infected with DENV results in an impaired ability of those DCs to prime T cells toward Th1 immunity , an effect that can be reversed by the addition of IFNβ [24] . Nevertheless , DENV is able to induce the expression of some pro-inflammatory cytokines at early times post infection , which we hypothesize is a strategy used by this virus to attract more cells to the site of infection by allowing the expression of some chemo-attractants by infected cells . Our group described that the infection by DENV does not induce the phosphorylation of IRF3 in human primary cells , resulting in an inhibition of type I IFN production [24] . In a subsequent report , we examined the ability of DENV-infected DCs to respond to a variety of type I IFN-triggering signals using potent stimulators such as NDV , SeV , SFV , or TLR-3 ligand poly ( I∶C ) [25] . This effect is viral dose dependent and takes place as early as 2 hours after DENV infection . We also showed that the inhibition of IFNα/β production after NDV infection in DENV-infected DCs is not a bystander effect , implying an active role of the DENV-infected DC population in the inhibition of IFNα/β . By using an NDV vector strategy to express the individual DENV non-structural proteins ( NS2A , NS2B3 , NS4A and NS4B ) , we showed that only the recombinant NDV expressing the protease complex NS2B3 inhibited IFNα expression in infected MDDCs , as compared to NDV alone . Similar results were obtained using an IFNβ promoter activity assay in 293T cells . Catalytically inactive NS2B3 mutants showed a diminished inhibition of this phenotype , which highlighted the important role for the protease activity of the NS2B3 protein as inhibitor of the type I IFN production . Interestingly , the proteolytic core of NS2B3 , consisting of the last 40 amino acids of NS2B and the first 180 amino acids of NS3 , was enough to reduce the activation of the IFNβ promoter by a strong stimulus , such as Sendai virus ( SeV ) infection . DENV has also been shown to express inhibitors of the type I IFN signaling cascade [26] and has been shown to encode for at least four non-structural proteins NS2A , NS4A , NS4B and NS5 that target different components of this pathway . The most remarkable example is the proteasomal degradation of human STAT2 by the NS5 , a phenomenon that does not occur in mouse cells , which makes mouse STAT2 a restriction factor for DENV replication in these animals [27] , [28] , [29] , [30] , [31] . In summary , DENV can successfully inhibit two fundamental steps of the innate immune system , both the inhibition of the type I IFN production and the signaling . In this way , DENV reduces the expression of hundreds of interferon inducible genes that would otherwise establish the antiviral state and control the spread of the infection in the host . In the present report , we describe the mechanism of inhibition of type I IFN production by DENV in primary human and mouse cells and identify the human adaptor molecule STING as a target of the DENV NS2B3 protease complex . We demonstrate that the proteolytic activity of this viral factor is crucial for the cleavage and degradation of STING and this phenomenon impairs the production of type I IFN in DENV infected cells . Furthermore , we show that DENV NS2B3 is not able to cleave the mouse version of STING . Using STING double knockout mouse embryonic fibroblast ( MEFs ) and human dendritic cells , we demonstrate the relevant role of this host factor in the restriction of DENV replication in mouse cells . This is the first report showing STING as a target for cleavage and degradation by a viral protein to inhibit innate immune responses and as a host restriction factor for virus infection in primary cells .
Previous results from our laboratory showed that dengue virus inhibits type I IFN production in human primary dendritic cells and that this inhibition requires a proteolytically active NS2B3 protease complex [24] , [25] . In order to identify potential NS2B3 targets we performed a bioinformatic search for potential DENV protease cleavage sites contained within members of the type I IFN pathway [32] . We identified putative cleavage sites in several known members of the type I IFN pathway ( see table 1 ) . After testing the factors shown in table 1 for their susceptibility to be cleaved by the DENV protease NS2B3 , we observed that only STING was cleaved in our experimental set up ( data not shown and figure 1 ) . We have generated a wild type DENV-NS2B3 , and a proteolytically inactive version ( NS2B3-S135A ) , ( Figure 1A ) by direct mutagenesis [25] that were used to analyze the potential cleavage of STING by the DENV protease complex . When we compared the human amino acid sequence of human STING to its mouse counterpart we noticed that the putative NS2B3 cleavage site in hSTING which is situated at the beginning of transmembrane domain 3 ( Figure 1B ) , is absent in mouse STING . In order to test the susceptibility of human and mouse STING proteins to proteolytic cleavage or degradation , we co-expressed a C-terminally HA-tagged STING alongside a wild type or catalytically active DENV-NS2B3 , and a proteolytically inactive version ( NS2B3-S135A ) in 293T cells , ( Figure 1A ) and analyzed them by western blot ( Figures 1C , 1D ) . In the presence of WT NS2B3 we observed the full-length 42 kDa human STING and an additional band of about 32 KD , which is consistent with a C-terminal region product of a cleavage occurring within the first 96 aa of STING ( Figure 1C ) . The additional band was not visible when the mouse version of STING or a catalytically inactive NS2B3 was used ( Figure 1C and 1D ) . This putative cleavage site for the DENV NS2B3 lies very close to the conserved cysteine motif C88xxC91 , or redox motif , recently described to be required for dimerization of STING and subsequent signaling in the type I IFN production pathway [33] . Regardless of their susceptibility to cleavage by the DENV NS2B3 complex , both the human and mouse versions of STING co-immunoprecipitated with the WT DENV NS2B3 complex and the proteolytically inactive mutant NS2B3 S135A ( Figure 1E , lanes 2 , 3 , 6 and 7 ) . To map the putative cleavage site of human STING for the DENV NS2B3 complex , we mutated the sequence corresponding to the first three amino acids of the human site , RRG ( shown in figure 1B as hSTING in red ) with the sequence corresponding to the amino acids HCM found in the mouse version of STING ( shown in figure 1B as mSTING ) . These recombinant versions of STING were co-transfected into 293T cells with the WT and mutant version of the NS2B3 protease and the ability of the DENV protease to cleave STING ( Figure 1F , lane 5 ) was drastically reduced when the mouse sequence was present in hSTING ( Figure 1F , lane 2 ) . These data confirm the requirement for amino acids RRG for efficient cleavage of STING by the DENV NS2B3 . However , replacement of the corresponding amino acid sequence of mouse STING ( IHCM ) by the human putative cleavage sequence ( LRRG ) does not render mouse STING susceptible to cleavage by the DENV protease ( Figure 1G , lane 2 ) , suggesting that additional flanking amino acids are required for this cleavage . Altogether , these results strongly suggest that STING is a target for NS2B3 in human cells and possibly a restriction factor for DENV infection in the mouse . To test whether endogenous STING undergoes the same NS2B3-dependent processing as in overexpression experiments in 293T cells , we infected human MDDCs with DENV-2 ( 16681 strain ) and analyzed the cell lysates by western blot at different time points ( Figure 1H ) . Infection of MDDCs by DENV resulted in the degradation of STING that could be detected at 24 and 48 hours post infection ( hpi ) ( Figure 1H , lanes 9 and 10 ) which correlate with peak expression levels of the NS2B3 ( as detected with NS3 specific antibodies ) . As expected , this degradation of STING was not observed in MDDCs treated with UV-inactivated DENV or mock treated cells ( Figure 1H , lanes1–5 and 11–15 ) . These data demonstrate that DENV NS2B3-dependent cleavage of endogenous human STING occurs in cells relevant to DENV infection ( MDDCs ) , and therefore has the potential to play a crucial role in inhibition of type I IFN production . We next investigated whether STING cleavage by DENV NS2B3 had an impact on its ability to mediate the signaling necessary for type I IFN production . We transfected 293T cells with either hSTING ( Figures 2A and 2B ) or mSTING ( Figures 2C and 2D ) and the three different versions of the DENV protease: wild type , the proteolytically inactive version ( NS2B3-S135A ) and the proteolytic core ( NS2Bh-NS3pro ) alongside luciferase reporter constructs driven by either an IFNβ promoter ( IFNβ-Luc ) or by three IRF3/7 binding sites ( p55-C1B-Luc ) ( kindly provided by Dr . Megan Shaw and shown in figure 2E schematically ) [34] . As shown in figures 2A and 2B , cleavage of hSTING by the DENV NS2B3-WT and NS2Bh-NS3pro greatly inhibited its ability to activate both reporter constructs , while transfection of the mutant version of the protease did not . Conversely , the DENV NS2B3 had minimal or no impact on the ability of mSTING to induce activation of either of the reporter constructs used . We did not observe any DENV NS2B3-dependent inhibition when the human adaptor TBK1 was used as a positive control to induce the IFNβ promoter , demonstrating that the inhibition during dengue infection occurs upstream of this adaptor ( data not shown ) . Taking together , these data demonstrate that cleavage of hSTING by the DENV NS2B3 precludes the induction of type I IFN responses . Moreover , to validate the observed results in human primary cells , we used E . coli DNA , a known inducer of STING signaling [35] to treat human MDDCs previously infected with DENV , UV-inactivated DENV or mock treated ( as shown schematically in figure 2F ) . Figures 2G , 2H and 2I show that only live DENV but not UV inactivated DENV was able to inhibit the induction of IFNβ , IFNα or ISG15 by this ligand in human MDDCs . To validate the results described using p55-C1B and IFNβ promoter assays in a primary cell model , we measured IFNα/β production upon infection of MDDCs either with DENV or a Semliki forest virus ( SFV ) expressing the DENV NS2B3 protease complex or the mutant version of the protease ( NS2B3-S135A ) as a control [23] , [25] . Consistent with our earlier report [24] , human MDDCs infected with DENV-2 ( 16681 strain ) were unable to produce IFNα/β . Furthermore , SFV-NS2B3 induced significantly lower levels of IFNα/β mRNA than the SFV-NS2B3-S135A ( Figures 3A and 3B ) . Interestingly , SFV-NS2B3-WT induced significantly higher expression of TNFα at early times post-infection compared to SFV-NS2B3-S135A control ( Figure 3C ) . As shown previously , this would suggest an involvement of the NS2B3 protease complex in the expression of this pro-inflammatory cytokine [24] . As expected , the infection of MDDCs by DENV up-regulated the expression of STING in these cells ( Figure 3D ) . Figure 3E shows the kinetics of infection by DENV in MDDCs , with the peak of viral RNA at 48 hpi . In contrast , the SFV vectors used in these studies show low levels of viral RNA at late times after treatment ( Figure 3F ) , since these vectors are replication deficient [36] . Then we infected mouse bone marrow-DCs ( BM-DCs ) using the same viruses , and analyzed the gene induction profile in those cells . As expected , infection of BM-DCs by DENV was rapidly controlled , consistent with the inability of DENV to infect mouse cells , and showed an opposite kinetic of viral RNA synthesis compared to the observed pattern in human DCs , with a modest induction of cytokines ( Figure 3G to 3K ) . SFV is an alphavirus that can replicate in mouse cells , and DCs in particular . In this context , the SFV-NS2B3 exhibited a higher induction of both IFNα/β genes compared to the SFV-NS2B3-S135A control , showing an opposite profile than that observed in human DCs , in which SFV-NS2B3-S135A induced higher levels of IFNα/β genes ( Figure 3G and 3H ) . Again , the kinetics of infection of the SFV vectors in mouse DCs ( Figure 3L ) show very low levels of viral RNA , consistent with their lack of productive infection in these cells [36] . The observed phenomenon agrees with the inability of recombinant DENV-NS2B3 to cleave mouse STING and decrease the activity of the IFNβ and p55-C1B promoters induced by this adaptor protein ( Figure 2C and 2D ) . To explore STING's impact on the DENV replication in mouse cells , we used WT ( Sting +/+ ) and STING double knockout ( Sting −/− ) mouse embryonic fibroblasts ( MEFs ) . First , we infected WT and Sting −/− MEFs with two different DENV-2 strains , 16681 and NGC ( a strain that was obtained after several passages in mouse brain ) [37] , with an MOI of 5 . Then , we measured the ability of the two DENV-2 strains to induce IFNβ production , to replicate in these cells and to release infectious particles from those cells ( Figure 4A–4F ) . Both DENV-2 strains induced significantly higher levels of IFNβ in WT MEFs as compared to the Sting −/− MEFs ( Figure 4A and 4D ) , underlining the relevance of STING in the signaling of type I IFN upon the infection with DENV . Consistent with the observed low induction of IFNβ , Sting −/− MEFs were permissive to DENV replication while , despite the high MOI used , replication of DENV 16681 and NGC was rapidly controlled in WT MEFs ( Figure 4B and 4E ) . The production of infectious particles by the two DENV-2 strains in WT and Sting −/− MEFs was measured by plaque assay and shows that the KO MEFs were more permissive to DENV infection than the WT MEFs and have very different peaks of infection ( Figure 2C and 2F ) . To test whether our observation was independent from the high MOI and viral strains used , we repeated the infection using different DENV serotypes ( DENV-2 16681 strain , DENV-3 PR-6 strain and DENV-4 H-241 strain ) and an MOI of 1 , with similar results ( data not shown ) . These results are likely due to the inability of DENV to inhibit the type I IFN signaling in mouse cells and the establishment of the antiviral state [31] To confirm the relevance of STING cleavage by the DENV-NS2B3 on the inhibition of type I IFN production upon DENV infection , we transduced Sting −/− MEFs with lentiviruses expressing either WT human STING ( STING-WT ) or a mutant ( uncleavable ) version , that harbors the mouse STING sequence at the NS2B3-cleavage site ( STING-MUT ) ( Figure 1B and 1F ) . Twenty-four hours after transduction MEFs were infected with DENV-2 at an MOI of 1 ( strains 16681 and NGC ) or mock treated and the levels of IFNβ , IFNα and viral RNA were measured by qRT-PCR after total RNA extraction from the cells at different times post infection ( Figure 5A–5F ) . A schematic representation of the lentiviruses used is shown in figure 5G . For these experiments , as shown in figure 5A and 5D , MEFs expressing the uncleavable version of STING ( STING-MUT ) expressed significantly higher levels of IFNβ mRNA when compared to MEFs expressing WT STING ( STING-WT ) or to the control MEFs ( GFP ) confirming that the cleavage of STING by DENV-NS2B3 is necessary for the inhibition of IFNβ production in infected cells . The induction of IFNβ was detected as early as 2 hpi . Infection with the two DENV-2 strains also induced higher levels of IFNα mRNA in MEFs expressing STING-MUT than in the MEFs expressing STING-WT and the GFP control ( Figure 5B and 5E ) . Under these experimental conditions , the replication of the mouse adapted DENV-2 ( NGC ) was increased in Sting −/− MEFs expressing wild type STING as compared with STING-MUT ( Figure 5F ) . In the case of 16681 strain , a significant increase of replication was observed only with Sting −/− MEFs ( GFP ) at 48 hpi . , and no significant difference was observed in MEFs expressing the two versions of STING , presumably due to a high level of lentiviral-expressed STING that could overwhelm the ability of this non mouse-adapted DENV strain to replicate in this system ( Figure 5C ) . As shown in figure 5 , the lack of STING cleavage was sufficient to increase the expression of type I IFN in MEFs infected with DENV . To validate these results in primary human cells , we transduced MDDCs from three different donors with the STING-expressing lentiviruses and the GFP-only control ( Figure 5G ) . STING transduced DCs were then infected with DENV-2 at an MOI of 5 and we assessed the production of IFNβ in those cells . As shown in Figure 6A ( showing one representative donor out of three ) , there were no significant differences between the levels of STING-WT and STING-MUT mRNA . This demonstrates that any difference in antiviral effect observed with the two different versions of STING is independent of the expression levels for these proteins . Upon DENV-2 infection , DCs expressing STING-MUT produced higher levels of IFNβ when compared with STING-WT or GFP controls ( showing statistical significance at 2 and 12 hpi ) . The induction of IFNβ messenger RNA was detected as early as 2 hpi , which is in agreement with the results described with MEFs ( Figures 4 , 5 and 6B ) . These data confirm that detection of DENV infection by DCs takes place at early times post infection and that STING cleavage by DENV NS2B3 is fundamental to inhibit the signaling mediated by this adaptor in human cells . We next measured DENV replication kinetics and we found that viral RNA levels were significantly lower in DCs over-expressing STING-MUT when compared with STING-WT and GFP control ( Figure 6C ) . Suggesting that the inability of DENV to cleave mutant STING and inhibit the induction of type I IFN has a direct impact on its replication kinetic and the accumulation of viral RNA in those cells . To determine STING's impact on DENV replication in primary human MDDCs , we used RNA interference ( RNAi ) to silence its endogenous expression . A decrease of STING mRNA level was observed when specific siRNAs were used compared to two scrambled control siRNAs ( Figure 7A ) . As expected , the previously observed upregulation of STING after DENV infection was controlled by the STING siRNAs ( Figure 7A ) . As a consequence , the reduction in STING expression resulted in an increase of DENV replication , illustrated in Figure 7B . When the viral progeny released by those infected MDDCs was quantified by plaque assay , the six donors treated with STING specific siRNA , showed viral production under this experimental conditions , however when scrambled siRNA was used , only three out of the six donors released detectable viral progeny in the supernatant ( Figure 7C ) . Data shown in figures 7A and 7B correspond to donor 4 in figure 7C . Taken together , these data confirm that STING is a crucial restriction factor of DENV replication in human dendritic cells , since its silencing increases the levels of DENV replication in those cells . Different populations of cells were isolated from human blood and subsequently infected with DENV-2 at MOI of 1 and 12 h after infection supernatants were collected and RNA was extracted from cells . DENV-2 RNA was detected in all cells tested including plasmacytoid DCs ( pDCs ) , B cells , blood circulating DCs ( cDCs ) , monocytes as well as in monocyte-derived DCs ( MDDCs ) ( Figure 8A ) . We also analyzed the cytokine and chemokine expression profile in all those cells by qRT-PCR ( data not shown ) and by multiplex ELISA ( Figure 8B ) in the supernatants at 12 hpi . We observed a marked chemokine response ( IL-8 and MIP1β ) in monocytes , MDDCs , B cells and cDCs at this early time point , ( Figure 8B ) . However pDCs did not show any significant chemokine profile after DENV-2 infection at this time point ( Figure 8B ) . More interestingly , there was no significant type I IFN production observed in any of the cells tested by qRT-PCR ( data not shown ) or ELISA ( Figure 8B ) . These data suggest that there is a coordinated and distinct kinetic of infection of DENV-2 in different cell populations in blood and there is a lack of type I IFN production in those cells after infection with this virus , at least at this early time point . The early time point of 12 hpi was chosen to obtain sufficient numbers of pDCs , since these cells have short half-lives and downregulate their specific cell surface markers in a rapid fashion . Nevertheless , we have previously reported that as early as 8 hpi , pDCs are able to produce type I IFN after infection with other viruses , such as NDV [24] . To rule out that the lack of type I IFN production resulted from lack of cell to cell interactions , we infected whole PBMCs with DENV-2 ( MOI of 1 ) and 18 h after infection cell supernatants were collected . Figure 8C shows multiplex ELISA data of cell supernatants from those cultures . While there is a clear IL-8 response to DENV-2 infection in PBMCs , consistent with the strong IL-8 signature observed in sera from patients [38] , there was no detectable IFNα secretion from infected PBMCs ( Figure 8C ) . This suggests that DENV-2 may inhibit type I IFN production in susceptible cells within those cultures . Macrophages have been shown to support DENV infection in animal models , and have been proposed to play an important role during early phases of dengue virus infection [39] , [40] . We tested if monocyte-derived macrophages ( MDMs ) when infected with DENV were able to produce type I IFN . Figure 8D shows that macrophages are efficiently infected with DENV , with an early peak of replication and produce TNFα and IL-6 after DENV infection and after SFV expressing the WT and mutant versions of the NS2B3 DENV protease complex ( Figures 8E , 8F ) . Under these experimental conditions we were unable to detect IFNα released by macrophages after DENV infection . Interestingly , when compared to MDDCs infected with the same viruses ( Figure 8H ) , macrophages produce at least 10-fold lower levels of IFNα after SFV infection , and the inhibitory effect of the DENV protease in this system was less apparent ( 8G and 8H ) .
Activation of innate immunity due to the detection of viral replication products in the cell leads to the expression of hundreds of antiviral genes that controls the spread of the infection [12] . The inhibition of different steps implicated in these molecular pathways by viruses has been a matter of extensive study for several years . It has been demonstrated by others and by our group that DENV can inhibit both the production and signaling of type I IFN by the expression of viral proteins . In this way , DENV can mitigate the immune response induced by the host upon infection [24] , [25] , [29] . Here we have identified the human adaptor molecule STING as a protein with a predominant role in the recognition of DENV by the innate immune system . This adaptor protein was described to reside in the ER , a cellular organelle intimately related to the DENV replication process . Also , STING has been described as part of the TRAP ( translocon associated protein ) complex that can associate with RIG-I and IPS-1 , two proteins with relevant roles in viral detection [41] . Ishikawa et al . also described an inhibition of the STING mediated IFNβ production by the yellow fever virus ( YFV ) NS4B [42] . However , when we tried to replicate these results using the DENV NS4B , this viral protein was unable to decrease the induction of luciferase mediated by STING in an IFNβ promoter assay ( data not shown ) . By co-expression experiments of human STING with the DENV NS2B3 protease complex we observed a specific cleavage at ( 94-RRGA-99 ) , a site described as a putative target for DENV NS2B3 [32] that generated a cleaved band of approximately 32 KDa ( Figure 1C and 1F ) . Interestingly , by analysis of the sequence alignment between human STING and its mouse version we , observed a drastic difference in the amino acid sequence in this region ( 94-HCMA-99 ) ( shown in figure 1B ) and the inability to cleave the mouse STING by the DENV NS2B3 was confirmed by co-expression experiments ( Figure 1D and 1G ) . Furthermore , the impact that the STING cleavage by NS2B3 had on the signaling of IFNβ production pathway was subsequently demonstrated using IFNβ and p55-C1B promoter systems ( Figure 2 ) . In these experiments , a reduction in luciferase induction was only observed for human STING , suggesting that the cleavage confirmed by WB ( showed in Figures 1C and 1F ) impaired the ability of this adaptor to induce IFNβ . Recently , Jin et al . described a series of mutations in hSTING that were implicated in the activation/dimerization and subsequent induction of interferon . Interestingly , two cysteines located at C88XXC91 were fundamental for the proper induction of type I IFN after stimulation [33] . It could be interesting to investigate the presence of mutations at the cleavage site of STING for DENV NS2B3 in the human population , to identify a natural resistance to DENV infection . While this manuscript was under review , Yu and colleagues reported by overexpression experiments that the DENV protease can cleave the adaptor molecule MITA , [43] . In the present report we provide important data on the role of this adaptor molecule in primary human and mouse cells and during the context of DENV infection . We also confirmed the cleavage and degradation of STING by the DENV NS2B3 protease in human MDDCs in the context of DENV infection , since it is important to validate these findings in a relevant primary cell system and during virus infection ( Figure 1H ) . In primary human cells as well as in mouse cells , such as MEFs and DCs , we also found that the presence of human STING allowed for greater DENV replication and the presence of mouse STING seemed to restrict DENV replication ( Figure 3 , figure 4 and figure 7 ) . We also show that the NS2B3 protease of DENV has specificity for the human STING and not for the mouse homologue of this protein ( Figure 5 and figure 6 ) , suggesting that STING may be an important restriction factor in mice . Several viral proteases have been described as proteins that modulate cellular pathways , allowing many viruses to modify the intra and extracellular environment to promote optimal conditions for replication and spread . Some of the most remarkable characteristics observed at early times after infection by DENV are the lack of IFNα/β induction and a robust induction of pro-inflammatory cytokines like TNFα [24] , [25] . As it was described in our previous work , DENV infection can abrogate IRF3 phosphorylation , but has no impact on NF-kB activity [25] . Using recombinant viruses expressing DENV-NS2B3 we observed a clear effect on the induction of TNFα in human DCs , similar to that observed with DENV infection ( Figure 3 and figure 8 ) . Also , the inhibition of luciferase activity driven by p55-C1B promoter was considerably more efficient when compared with IFNβ-promoter , since p55-C1B only harbors sites for IRF3/7 transcription factors , and IFNβ-Luc has also has response elements for NF-kβ and AP-1 transcription factors ( Figures 2B and 2D ) . Furthermore , Ishikawa et al . overexpressed STING in 293T cells in the presence of different promoters driving the luciferase gene . Interestingly , STING stimulated IFNβ promoter up to 400-fold , IRF3 response element ( PRDIII-I-Luc ) up to 1 , 000-fold , and NF-kβ responsive promoter ( NF-kβ-Luc ) only up to 12-fold [41] . This observation suggested that STING is fundamentally involved in phosphorylation of IRF3 , and under these experimental conditions showed a 100 fold less influence on NF-kβ induction . Taken together , these observations suggest that DENV NS2B3 protease inhibits IFNβ production by cleavage of the adaptor STING without modifying the observed NF-kβ activity induced after infection by DENV . Further work exploring the impact that DENV-NS2B3 has on the induction of NF-kβ activity in infected cells would confirm a putative role of this viral factor in the modulation of innate immune response by the induction of pro-inflammatory cytokines , a hallmark phenomenon observed during infection by DENV [44] . It is becoming increasingly clear that STING is a crucial adaptor in immune cells after infection with different viruses , such as HIV and DENV , among others [35] , [45] , [46] . These viruses require activation of their target cells in order to establish infection , or in the case of DENV to induce viremia in the host . Nevertheless , all viruses need to limit or inhibit the production of type I IFN in infected cells to avoid the establishment of an antiviral state in those cells . STING could be instrumental in these types of virus infections , since it can discriminate between the induction of type I IFN and the activation of the NFkB pathway . Along those lines , this report shows a novel mechanism of inhibition of IFN production by an RNA virus , namely DENV targeting STING . By inhibiting only type I IFN but not the NF-κB pathway , DENV induces a specific profile in infected human MDDCs and other susceptible primary cells that allow the virus to efficiently reach the lymph nodes and spread in the infected host , culminating in the production of viremia . All our experiments were performed in the context of primary infections with DENV , since we believe that the early events in primary infections dictate the quality of adaptive immune responses and the outcome of the infection . By targeting DCs and inhibiting the production of type I IFN in those cells , DENV may be able to efficiently modulate the generation of adaptive immune responses and establish infection in the host [24] , [25] . It has been proposed that during DENV infection IPS-1 may be responsible for controlling early viral replication and type I IFN production [47] , while the IFN signaling pathway ( JAK/STAT ) may control late viral replication and type I IFN production in DENV infected cells [48] . It is possible that interactions between STING and IPS-1 [46] may be disrupted by the DENV NS2B3 targeting of STING . This mayinhibit type I IFN production early during DENV infection in susceptible cells , although the NS2B3 has not been shown to directly interact with IPS-1 . Further experiments are required to understand these complex interplays between different signaling molecules in human primary cells infected with DENV . Our experiments demonstrate that primary human cells implicated in dengue virus infection , such as dendritic cells , macrophages , monocytes and B cells can support DENV replication , although at different levels . Interestingly , DENV infection did not induce type I IFN production in any of those human primary cells tested ( Figure 8 ) . These different blood cells may play different roles during DENV infection in humans , such as being involved in the initial infection or in the final stage of viremia . Also , since the mouse models that support DENV replication and recapitulate dengue symptoms are deficient in type I IFN responses or are reconstituted with human immune cells [6] , [49] , the inhibition of type I IFN in infected cells seems to be crucial for the establishment of infection by DENV . Our data on different cells from blood also show that the inhibition of type I IFN production by DENV is not a DC specific phenomenon ( Figure 8 ) . The inability of DENV to replicate in wild type mouse cells is well documented , and many attempts have been made to develop a competent animal model to study DENV infection [50] . The data presented in this manuscript , showing the ability of DENV to replicate in Sting −/− MEF ( Figure 4 ) , open new approaches to develop a mouse model to study DENV infection and also highlights the requirement that type I IFN production has on the innate immune system and for the control of invading pathogens . Ashour et al . described the adaptor STAT2 as a restriction factor for DENV replication in mouse cells [31] . Based on our combined data , an interesting approach would be the development of a transgenic mouse model with humanized STING and STAT2 . This approach could provide an immune competent mouse model for DENV that eliminates two of the potential bottlenecks that exist for DENV replication in mice .
The animal protocol used in this study was reviewed and approved by the University of Miami Institutional Animal Care and Use Committee ( IACUC ) under IACUC protocol 11–181 “Host Defense and the Regulation of Interferon Production: STING . ” The University of Miami has an Animal Welfare Assurance on file with the Office of Laboratory Animal Welfare ( OLAW ) , National Institutes of Health . The assurance number is #A-3224-01 , effective July 11 , 2007 . Additionally , as of July 20 , 2010 , the Council on Accreditation of the Association for Assessment and Accreditation of Laboratory Animal Care ( AAALAC International ) has continued the University of Miami's full accreditation . Vero , 293T and mouse embryonic fibroblast ( MEFs ) , were cultured in Dulbecco's modified essential medium ( DMEM ) supplemented with 10% fetal bovine serum ( FBS ) . Baby hamster kidney cells ( BHK ) were grown in Glasgow minimal essential medium ( MEM ) supplemented with 10% FBS , and 20 mM HEPES . Mosquito cells derived from Aedes albopictus , clone C6/36 , were expanded at 33°C in RPMI medium with 10% FBS . All media were supplemented with 100 U/ml of L-glutamine and 100 µg/ml of penicillin-streptomycin . All tissue culture reagents were purchased from Invitrogen . Dengue virus serotype 2 ( DENV-2 ) strains 16681 and New Guinea C were used in this study . DENV was grown in C6/36 insect cells for 6 days as described elsewhere [51] . Briefly , C6/36 cells were infected at a multiplicity of infection ( MOI ) of 0 . 01 , and 6 days after infection , cell supernatants were collected , clarified , and stored at 80°C . The titers of DENV stocks were determined by limiting-dilution plaque assay on BHK cells [52] . Semliki Forest virus ( SFV ) expressing GFP and DENV-NS2B3 were generated as described previously [36] and titrated in BHK cells by immunofluorescence [53] . Lentiviral vector constructs were built using conventional molecular biology techniques . Briefly , human STING cDNA was PCR amplified from pcDNA3 . 1 hSTING [13] and cloned into a lentiviral vector derived from pHR SIN CSGW [54] ( Figure 5G ) . Mutations in the NS2B3 cleavage site at positions 94–96 of hSTING were obtained by overlap PCR . Residues RRG were changed to the corresponding murine sequence HCM . Lentiviral vector derived viruses were obtained by transfection of HEK 293T with 3 plasmids encoding STING , HIV-1 Gag-Pol , and VSV-G respectively [55] . Viral supernatants were harvested 48 and 72 hours post-transfection , 0 . 45 µm filtered , concentrated at 14 , 000 g for 6 hours over a 20% sucrose cushion and frozen at −80°C until used . Monocytes , pDCs , B cells and circulating CD11c+ DCs ( cDCs ) were isolated from blood of healthy donors ( New York Blood Center ) using Miltenyi isolation kits . CD14+ clinimacs , for monocytes , CD123/BDCA4 kit for pDCs and BDCA1 kit for cDCs . B cells were isolated as part of the BDCA1 kit for isolation of cDCs according to manufacturers' instructions . The purity of each cell population was tested by flow cytometry as described below and was routinely 85–95% for CD14+ cells , 87–90% for pDCs , and 95–99% for both MDDCs and cDCs . Samples of 5×105 isolated cell populations were infected with DENV-2 , 16681 at the indicated MOI in a total volume of 100 µl of DC media for 1 hour at 37 C . Then , DC media supplemented with 4% HS was added up to a final concentration of 106 cells/ml and cells were incubated for the remainder of the infections at 37 C . At the indicated times , cell supernatants were collected and cell pellets were used for RNA extractions . Whole PBMCs were used after ficoll centrifugation and samples of 60×106 PBMCs were infected with DENV-2 at MOI of 1 or left uninfected . After 1 hour DC media 4% HS was added . Eighteen hpi supernatants were collected and isolation of the different cell populations after DENV-2 infection was carried out as described above . Human MDDCs were obtained from healthy human blood donors ( New York Blood Center ) , following a standard protocol as previously described [24] and described above . Briefly , after Ficoll-Hypaque gradient centrifugation , CD14+ cells were isolated from the mononuclear fraction using a MACS CD14 isolation kit ( Milteny Biotec ) according to the manufacturer's directions . CD14+ cells were then differentiated to naïve DCs by incubation during 5 to 6 days in DC medium ( RPMI supplemented with 100 U/ml L-glutamine , 100 g/ml penicillin-streptomycin , and 1 mM sodium pyruvate ) with the presence of 500 U/ml human granulocyte-macrophage colony-stimulated factor ( GM-CSF ) ( PeproTech ) , 1 , 000 U/ml human interleukin 4 ( IL-4 ) ( PeproTech ) , and 10% FBS ( Hyclone ) . To generate MDMs , monocytes were cultured in the presence of 2000 U/ml human granulocyte-macrophage colony-stimulated factor ( GM-CSF ) for 10 days , and media was replenished ( with same concentration of GMCSF ) at days 2 , 5 and 8 . The purity of each cell population was confirmed by flow cytometry analysis and was at least 99% for MDDCs and 95% for MDMs . Femurs and tibia of wild-type C57BL/6 mice ( Jackson ) were soaked in 70% ethanol , washed with RPMI ( Invitrogen ) , and epiphyses were cut to expose the bone marrow . The bones were flushed with RPMI supplemented with 10% fetal bovine serum ( FBS ) to extract the bone marrow . Cells were pelleted by centrifugation , washed once with RPMI and resuspended in ammonium chloride red blood cell lysis buffer . RBC lysis was performed for 1 minute at room temperature , stopped with RPMI-FBS and cells were collected by centrifugation . Bone marrow cells were seeded in 6-well dishes in RPMI containing 10% FBS , 50 U/ml Penicillin ( Invitrogen ) , 50 µg/ml Streptomycin ( Invitrogen ) , 20 ng/ml GM-CSF ( Peprotech ) , 10 ng/ml IL-4 ( eBioscience ) , and 40 µM beta-mercaptoethanol ( BIO-RAD ) . Cells were cultured for 5 days at 37 degrees C , 5% CO2 and fresh media was added every 2–3 days . Human and mouse DCs were obtained as described above , and at day 5 of culture , samples of 1×106 cells were resuspended in 100 µl of DC-medium and were infected for 45 min at 37°C with the indicated MOI of virus ( diluted in DC media ) or with DC medium ( mock group ) in a total volume of 200 µl . After the adsorption period , DC medium supplemented with 10% FBS was added up to a final volume of 1 ml , and cells were incubated for the appropriate time at 37°C . 2 . 5×104 MDDCs were seeded per well in 96 well plates and transfected with the corresponding siRNA using the StemFect RNA transfection kit ( Stemgent ) , according to the manufacturer instructions . Chemically synthesized 27mer siRNA duplexes were obtained from OriGene Technologies , Inc . The sequences of the STING siRNA oligonucleotides used in this study are as follows: siSTING -1:5′-rGrGrCrArUrGrGrUrCrArUrArUrUrArCrArUrCrGrGrArUAT-3′ . siSTING-2:5′-rArCrCrUrGrUrGrArArArUrGrGrGrArUrCrArUrArArUrCAC-3′ . siSTING-3:5′-rGrGrArUrUrCrGrArArCrUrUrArCrArArUrCrArGrCrArUTA-3′ . Two random non-coding control siRNA were used: Sc-1 ( 5′-rUrArCrGrUrArCrUrArUrCrGrCrGrCrGrGAT-3 ) from Qiagen , and Sc-2 ( universal scrambled negative control siRNA duplex SR30004 ) from OriGene Technologies , INC . 48 h after transfection , cells were infected with Dengue virus at an MOI of 1 . Briefly , cells were centrifuged ( 400×g , 10 min ) , the media was removed and 25 µl of RPMI containing the appropriate amount of virus was added and the plates were incubated for 45 min at 37°C . Then , 75 µl of RPMI with 10% FBS were added and cells were incubated at 37°C for the indicated hours . Subsequently , cells were recovered by centrifugation for 10 min at 400×g , and the cell pellets were lysed for RNA isolation . Plated monocytes were transduced as previously described [56] . In brief , freshly isolated monocytes were transduced with VSV-G pseudo-typed SIV VLPs and each lentiviral vector construct for 3 h by spinoculation , in the presence of 2 µg/mL polybrene ( Sigma ) . Subsequently , cells were washed , resuspended in regular growth medium described before for the generation of monocytes derived human dendritic cells and incubated for 5 days at 37°C until stimulation . At day 5 post-transduction , MDDCs were infected with DENV as described before , and cell pellets were collected at the indicated time points , and lysed for RNA isolation . 293T cells were transfected by using Lipofectamine 2000 reagent ( Invitrogen ) according to the manufacturer's protocol . A type I IFN production antagonist assay was performed as described previously [25] using IFNβ-Luc and p55-C1B-Luc [34] , [57] . 293T cells seeded on 24-well plates were transiently transfected with 50 ng of the luciferase reporter plasmid together with a total of 400 ng of various expression plasmids or empty control plasmids . As an internal control , 50 ng pRL-TK was transfected simultaneously . Then , 24 or 48 h later , the luciferase activity in the total cell lysate was measured by using the Dual-Luciferase Reporter Assay System ( Promega ) according to the manufacturer's directions . Transfection of 293T cells and infection of human DCs was performed as described above . Cell lysates were obtained after incubation of cells with RIPA lysis buffer ( Sigma Aldrich ) supplemented with complete protease inhibitor ( Roche ) and resuspended in a total of 50 ml of Laemmli sample buffer ( Bio-Rad ) . Crude lysates were either boiled for 10 min or incubated at 42°C for 20 min and then kept on ice . Each sample was loaded in a polyacrylamide-SDS gel , and the proteins were electrophoretically separated by conventional methods . Proteins were transferred to nitrocellulose , and blots were incubated with anti-HA , anti-FLAG , anti-Actin anti-GAPDH ( Sigma Aldrich ) and rabbit polyclonal antibodies anti-hSTING [41] and anti-DENV NS3 ( kind gift of Dr . Andrea Gamarnik ) , and developed using SNAP ID detection system ( Millipore ) , following the manufacturer's instructions . Antibody-protein complexes were detected using a Western Lighting chemiluminescence system ( Perkin Elmer ) . RNA from different cells was extracted using Trizol ( Invitrogen ) , followed by a treatment with DNase using DNA-free Ambion . The concentration was evaluated in a spectrophotometer at 260 nm , and 500 ng of RNA were reverse transcribed using the iScript cDNA synthesis kit ( Bio-Rad ) according to the manufacturer's instructions . Evaluation of the expression of human and mouse cytokines from different cell types and viral RNA was carried out using iQ SYBR green Supermix ( Bio-Rad ) according to the manufacturer's instructions . The PCR temperature profile was 95°C for 10 min , followed by 40 cycles of 95°C for 10 s and 60°C for 60 s . Expression levels for individual mRNAs were calculated based on their CT values using two different housekeeping genes ( human: rps11 and α-tubulin genes ) and ( mouse: S18 and β-Actin ) to normalize the data . One paired two tailed Student's t-test was used to analyze data . Data considered significant demonstrated p values less than 0 . 05 .
|
Dengue virus ( DENV ) is a pathogen with a high impact in human health that replicates in a wide range of cells of the immune system . To efficiently infect humans , DENV must evade or inhibit fundamental elements of the innate immune system , namely the type I interferon response ( IFN ) . Thus , DENV can inhibit type I IFN signaling ( described by several groups ) , and type I IFN production ( described by our group ) . We documented the inhibition of type I IFN production in human monocyte derived dendritic cells ( MDDCs ) with an otherwise strong cytokine and chemokine profile in those cells and that the NS2B3 protease complex of DENV functions as an antagonist of type I IFN production , and its proteolytic activity is necessary for this event . Here we identify the human adaptor molecule STING as a target of the NS2B3 protease complex and characterize the mechanism of inhibition of the type I IFN production in primary human MDDCs mediated by this viral factor . We also describe that DENV NS2B3 cannot degrade the mouse version of STING , a phenomenon that strictly restricts the replication of DENV in mouse cells , suggesting that STING plays a key role in the inhibition of DENV infection and spread in mice .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"infectious",
"diseases",
"immunology",
"biology",
"microbiology"
] |
2012
|
DENV Inhibits Type I IFN Production in Infected Cells by Cleaving Human STING
|
Retroviral recombination is thought to play an important role in the generation of immune escape and multiple drug resistance by shuffling pre-existing mutations in the viral population . Current estimates of HIV-1 recombination rates are derived from measurements within reporter gene sequences or genetically divergent HIV sequences . These measurements do not mimic the recombination occurring in vivo , between closely related genomes . Additionally , the methods used to measure recombination make a variety of assumptions about the underlying process , and often fail to account adequately for issues such as co-infection of cells or the possibility of multiple template switches between recombination sites . We have developed a HIV-1 marker system by making a small number of codon modifications in gag which allow recombination to be measured over various lengths between closely related viral genomes . We have developed statistical tools to measure recombination rates that can compensate for the possibility of multiple template switches . Our results show that when multiple template switches are ignored the error is substantial , particularly when recombination rates are high , or the genomic distance is large . We demonstrate that this system is applicable to other studies to accurately measure the recombination rate and show that recombination does not occur randomly within the HIV genome .
Viral diversity is one of the major obstacles to the successful eradication of HIV [1] , [2] . It arises due to the interplay between mutations introduced by error-prone reverse transcription [3] , high levels of viral turnover [4] , retroviral recombination [5] and strong diversifying selection pressure from the immune system [2] . All retroviruses co-package two RNA genomes into each virion . Retroviral recombination occurs when the reverse transcriptase ( RT ) enzyme switches between co-packaged RNAs during reverse transcription ( reviewed in [6] , [7] ) . In HIV , recombination occurs much more frequently than mutation [8] , and is a major determinant of viral diversification . Within infected individuals , recombination allows sequential rounds of viral escape of both antibody and T-cell recognition , resulting in loss of immune control [9] , [10] . Furthermore , recombination can both promote and suppress the generation of multiple drug resistance , by creating or breaking linkages between drug resistance mutations [11]–[16] . Therefore , an accurate measurement of recombination rates directly within the HIV genome is fundamental to our understanding of HIV . Recombination has been studied extensively , by many groups , and is typically detected by monitoring the linking of marker points from co-packaged RNA genomes into a single DNA genome . One popular method of measuring recombination is through the use of retroviral reporter systems . These systems measure recombination within a ‘foreign’ gene insert , such as genes that code for antibiotic resistance proteins , surface protein markers , and/or fluorescent proteins [8] , [17]–[24] . Retroviral reporter systems have the advantage of being able to readily quantify a large number of recombination events within the gene insert . However , in vitro studies show that template sequence and nucleic acid structure are important determinants of the recombination process [25] , [26] . Therefore , measurements of recombination rates within non-HIV ‘foreign’ gene sequences will not recapitulate recombination rates within HIV sequence . Other groups utilize the genetic variation between and within HIV subtypes , and use sequencing to monitor recombination [8] , [21] , [22] . These systems provide the foundation to reveal recombination events within the HIV genome . However , the use of genetically divergent RNA templates does not reflect the situation in vivo , where the vast majority of infected individuals are infected with a single virus which rapidly diversifies into a viral quasispecies over the course of infection [27] . The use of divergent RNA sequences can lead to confounding differences in parameters known to affect recombination , including: overall RNA homology [28] , [29] , RNA packaging [30] , [31] , and the amino acid sequence of viral proteins , such as reverse transcriptase [32]–[34] . Therefore , the recombination events detected using divergent RNA sequences most likely reflect the special case of inter-subtype recombination . Hence , there is a real need to develop a retroviral recombination system which mimics the recombination that occurs between closely related , yet genetically distinct , viruses found within an infected individual . Recombination is detected by monitoring the linking of marker points from separate RNA genomes into a single DNA genome . Regardless of the system in which it is measured , recombination is either detected or undetected between any two marker points . This is generally interpreted as one or zero recombination events , respectively . However , with increasing genomic distance and/or recombination rate , there is an increased likelihood that there will be multiple template switches between any two marker points which go undetected . Consequently , with high rates of recombination and/or genomic distances between marker points , there is a greater chance of underestimating recombination rates due to multiple template switches . These possibilities have been mentioned previously [20] , [24] , [35] , [36] . However , there is no current standard method to calculate recombination rates over multiple genetic regions of varying lengths that also compensates for the possibility of multiple template switches between marker points . Additionally there exists no theoretical estimate for the error when recombination is measured without compensating for multiple template switches , as is often the case . Here , we present a novel experimental method based on limited codon modification of the HIV genome which does not change the infectivity of the virus or any viral protein . This allows the measurement of recombination between closely related genomes analogous to those found in the quasispecies of an infected individual . This system measures recombination in different gene segments , allowing the identification of possible recombination ‘hotspots’ , where template switches occur at higher frequencies . We then develop statistical tools to calculate an ‘optimal recombination rate’ that reproduces observed recombination frequencies , taking into account multiple template switches . These tools demonstrate the error in calculating crude recombination rates ( that do not consider multiple template switches ) and emphasize the necessity for careful data analysis . These tools also provide the basis to quantify statistical differences in recombination rates in various regions of the HIV genome , under different conditions , or infection with different target cells . Finally , our analysis allows for testing and subsequent validation of some inherent assumptions and sources of error in the experimental design . We compare our analytic procedure with previously published studies and find that our approach avoids some of the potential pitfalls of using reporter gene inserts .
Recombination is measured by analysing the cDNA that results from infection with non-identical ( heterozygous ) co-packaged RNA genomes . The positions in which the RNA genomes differ are called marker points . Recombination is detected only when the resulting cDNA contains a mixture of marker points from both RNA strands . It is tempting to conclude that one template switch has occurred every time recombination is detected between a set of marker points , and that no template switches occurred elsewhere . However , any even number of template switches between two fixed marker points will lead to us observing no recombination , and any odd number will result in us observing a single recombination event ( Figure 1A ) . An important consequence of this fact is that the probability of observing a recombination event is a function of the genomic distance between the markers and the overall recombination rate . We created a model of recombination which takes into account the possibility of not detecting recombination events ( see Materials and Methods ) . Our recombination rate calculation ( denoted ‘optimal’ recombination rate ) reveals the relationship between the overall recombination rate , distance between marker points and the probability of observing a recombination event ( Figures 1B and 1C ) . We show that for each genomic distance and overall rate of recombination , there is a unique probability of observing recombination . Furthermore , with high overall rates of recombination and large genomic distances , it becomes much more difficult to calculate the recombination rate accurately . Indeed , these probabilities eventually converge until it becomes impossible to derive the true rate of recombination because there is an equal chance of observing or not observing a recombination event . To demonstrate the consequences of ignoring multiple template switches , we utilized a simple equation ( denoted ‘crude’ recombination rate calculation ) : r = c/nl , where r is the rate of recombination events per nucleotide per round of infection ( REPN ) , c is the number of template switches detected , n is the number of sequences , and l is the genomic distance over which recombination is measured . This crude formula assumes that between marker points , at most one template switch can occur . To calculate the theoretical expected error of the ‘crude’ recombination rate we first use the ‘optimal’ recombination rate equation ( Eq . A ) to determine the probability of observing recombination over different genomic distances . We then use the ‘crude’ recombination rate calculation on these probabilities and find that this calculation consistently underestimates the real recombination rate . At an actual recombination rate of 0 . 001 REPN ( lower than the median recombination rate measured in T-cells in this study ) , the calculated crude recombination rate is 9% lower when measured over a distance of 100 nucleotides , and 37% lower when measured over a distance of 500 nucleotides ( Figure 1D ) . This error is even larger when the real recombination rate is 0 . 003 , where the crude rate is 25% and 68% lower than the actual rate when measured over a distance of 100 nucleotides and 500 nucleotides respectively ( Figure 1E ) . This error is a direct result of not considering multiple template switches , emphasizing the need for our optimal recombination rate calculation . We sought to measure the rate of recombination directly in the HIV genome . To this end , we made a marker virus ( MK ) by introducing 6 codon modifications into the gag gene of wild-type ( WT ) HIV . This creates 5 regions ( varying in length from 77 to 398 nucleotides ) over which we can directly measure the recombination rate of a full length HIV genome ( Figure 2A and 2B ) . These modifications neither affect the infectivity of the virus nor alter the amino acid sequence of any viral protein ( Figure S1 ) . The recombination process depends greatly on template sequence , RNA structure , the overall homology between sequences and the viral proteins involved in reverse transcription . Therefore , this system mimics the situation in vivo , where recombination occurs in the context of a quasispecies of highly related , yet genetically distinct viruses . In our experimental system , a template switch observed in the DNA provirus is most likely to be the result of viral recombination during reverse transcription of the two RNA molecules co-packaged in a heterozygous virion . However , it is possible that recombination could also have occurred during a number of steps in sample preparation and sequencing . To determine the potential bias within our experimental system , we quantified experimentally-induced recombination , as follows: Firstly , we tested the possibility of transfection-induced recombination which can occur via homologous recombination in the producer cell [37] . We measured this by direct sequencing of plasmid DNA extracted from co-transfected 293T cells ( Figure 2C ) . Of 182 sequences of plasmid DNA extracted from transfected cells we observed zero recombination events , suggesting that this is not a source of error in our system ( Table 1 ) . Secondly , we tested for PCR-induced recombination that may occur if the polymerase switches templates during PCR amplification of the viral sequences prior to sequencing . We measured this by performing two separate infections with either WT homozygous virus or MK homozygous virus ( Figure 2D ) . In this case , recombination occurs at the usual rate between co-packaged HIV RNA strands , but template switching between these identical copies of RNA cannot be detected . These homozygous samples are mixed prior to PCR . Thus , any observable recombination can be inferred to be an artifact of the PCR . 125 sequences were obtained and 3 recombination events were detected ( Table 1 ) . Finally , we measured the rate of ‘inter-virion’ recombination that may have occurred if the target cells were multiply infected , and retroviral recombination was occurring between the RNA molecules of different virions . To do this we co-infected cells with homozygous WT and homozygous MK virions . Thus , any intra-virion recombination would be undetected , but both inter-virion recombination and PCR-induced recombination would be detected ( Figure 2E ) . 128 sequences were obtained and 2 recombination events were detected ( Table 1 ) . To measure the biological rate of recombination we generated a mixture of heterozygous and homozygous virus by co-transfection . When equal amounts of two HIV plasmids are co-transfected , co-packaging of RNA into virions is random [38] . Therefore , when we co-transfected equal amounts of WT and MK plasmid , we expect 50% heterozygous virions , 25% homozygous WT virions and 25% homozygous MK virions ( Figure 2F ) . This mix was used to infect primary T-cells . 118 sequences were obtained and 58 recombination events were detected ( Table 1 ) . In determining recombination rates , it is easy to assume that transfection of equal amounts of WT and MK plasmid leads to the production of 50% heterozygous and 50% homozygous virus . However , variations in the level of co-transfection will lead to the production of a different proportion of heterozygous virions than expected . This will bias the calculation of recombination rates . Our design allows us to estimate the proportion of heterozygous virions in our experiments directly from the data ( as described in Materials and Methods ) . The estimated proportion of heterozygous virus was approximately 50% in our studies ( 48 . 6% , 45 . 1% , 49 . 7% and 46 . 1% for transfection , PCR , between virion and T-cell experiments respectively ) , indicating that there is no bias in infection rates between WT and MK virus or in the production of our heterozygous virions . We then calculated the recombination rates for each of our experimental conditions , using both our crude and optimal recombination rate calculations ( Table 1 ) . As we detected no recombination events in our transfection-induced recombination control , the crude and optimal recombination rates were 0 REPN . From 125 and 128 sequences for the PCR-induced recombination control and the PCR-induced plus inter-virion control , we observe 3 and 2 recombination events respectively . This corresponds to an optimal recombination rate of approximately 0 . 1×10−3 REPN . For our biological sample , the crude recombination rate was calculated to be 0 . 81×10−3 REPN , and the optimal recombination rate to be 1 . 45×10−3 REPN . Thus , the crude recombination rate underestimates the optimal rate by approximately 44% . This underlines the importance of calculating recombination rates using our ‘optimal’ recombination rate calculation instead of the ‘crude’ method commonly used in the literature , which does not compensate for multiple template switches . Using the above approach we are able to directly estimate the recombination rate from an experimental data set . However , the error of this estimate is affected by the number of sequences sampled , and their distribution . In order to determine confidence intervals for these estimates we generated probability distributions by bootstrapping the sequence data ( see Materials and Methods ) . The 95% confidence intervals of these distributions are calculated with the Percentile Method and are shown in Table 1 . Due to the high number of samples ( >118 for all datasets ) and relative symmetry of the bootstrap distributions ( data not shown ) , we assume very good coverage of these confidence intervals . We conclude that the recombination rates are significantly different ( at the 0 . 05 level ) when the 95% confidence intervals do not overlap . These distributions show that the recombination rate is not significantly different between PCR induced recombination and PCR induced plus inter-virion recombination . Thus , inter-virion recombination is not a significant factor in our experimental setup . However , recombination rates were significantly different between our controls and the rate of HIV RT-induced recombination in the biological sample . The true HIV RT-induced recombination rate was then calculated with a control correction method ( see Materials and Methods ) , that is approximately a subtraction of the two recombination rates . The RT-induced recombination rate alone is calculated to be 1 . 35×10−3 REPN in primary T-cells . Our experimental system allows recombination to be measured between closely related viral genomes . However , most recent recombination assays involve the insertion of fluorescent proteins into the HIV genome . In these systems two distinct defective genes , encoding a fluorescent protein , are inserted into different HIV genomes . A recombination event that eliminates the deactivating mutations recreates a functional fluorescent encoding gene . Recombination can then be measured via FACS analysis of infected and fluorescent protein expressing cells . This technique is capable of producing large quantitative datasets and has shown to be an effective tool to compare recombination rates under varying conditions . Generally , the extent of recombination in these systems is measured as a function of the multiplicity of infection ( MOI ) of fluorescent protein expressing cells and the MOI of viral infection . However these calculations are not easily comparable to calculations made for marker points separated by different genomic lengths . A clearer approach is to calculate the recombination rate in terms of ‘recombination events per nucleotide per round of infection’ ( REPN ) , as this rate allows the prediction of the number of recombination events that will occur over any length of RNA . To demonstrate how our recombination rate calculation method can be applied to fluorescent protein studies , and to make a direct comparison of these recombination rates to our own , we analysed the data from Rhodes et al . 2005 [24] . Table 2 , 3 , 4 , and 5 from Rhodes 2005 lists the total number of cells , infected cells , and green fluorescent protein positive ( GFP+ ) cells when the recombination is measured over a genomic distance of 588 , 300 , 288 and 103 base pairs , respectively . From these ratios the GFP+ MOI/infection MOI ( ratio denoted as M ) is calculated . This ratio represents the probability of a single infection event resulting in the reconstruction of a functional GFP protein ( see Materials and Methods ) . Recombination is only detectable from 50% of the virions ( those that are heterozygous ) . Thus , the probability that a heterozygous infection recreates a functional GFP is 2M . A functional GFP is only created when the two inactivating mutations are eliminated via recombination . However , the two inactivating mutations being ‘joined’ via recombination is equally likely . Thus , the probability that a heterozygous infection results in mosaic cDNA ( from a nucleotide sequence perspective ) is 4M . Using equation ( A ) ( Materials and Methods ) with R ( L ) = 4M converts M into the required recombination rate measured in REPN . Thus , taking into account the possibility of multiple template switches , the recombination rates for the data in Rhodes 2005 ranges from 0 . 49×10−3 to 0 . 97×10−3 ( table 2 ) . Note that the calculated optimal recombination rates in Rhodes 2005 are similar regardless of the genomic distance over which recombination is measured . This is because our analytical recombination rate calculation compensates for genomic distance when calculating the probability of multiple template switches . Our conversion of the data in Rhodes 2005 to a recombination rate per nucleotide per round of infection is in line with previous conversions by Suryavanshi and Dixit [36] , who used curve fitting techniques to estimate an average recombination rate over the different lengths . The advantage of our technique is that our procedure can be applied with a standard calculator and requires no curve fitting experience or software . Crossover sites of the HIV-1 RT may consist of RNA sequence determinants that direct the RT to switch templates , and it has been suggested that RNA-RNA interactions can promote recombination in vitro [28] , [29] . Unlike systems that measure recombination over only one region , our experimental design allows recombination to be studied in five gene segments in gag , which cover a total genomic distance of 917 base pairs . Our analytical recombination rate calculation allows us to calculate the optimal constant recombination rate that best describes our experimental data and compensates for the possibility of multiple template switches . This system allows us to determine: ( i ) if the variation in recombination along the gene is significantly different than that expected by random variation ( indicating whether recombination is a random event ) ; ( ii ) the optimal location for a recombination rate change ( determining the marker point that separates any recombination ‘hotspots’ and ‘coldspots’ ) ; and ( iii ) whether a two recombination rate model better describes the observed experimental recombination data . Together , these analyses will help us to determine whether recombination occurs randomly across the viral genome . We first use a chi-squared goodness of fit test to determine if the observed frequency of recombination and the expected frequency ( calculated from our optimal recombination rate and compensating for multiple template switches , equation ( B ) Materials and Methods ) are significantly different in each gene segment . Figure 3A profiles the experimentally observed and expected number of detected template switches that were recorded over the different sections of gag . The experimental data displays significant variation from the expected frequencies of recombination to the observed frequencies ( p = 0 . 02 ) suggesting that recombination rates vary along the gene segments . We then adjusted our mathematical model of HIV recombination to fit two optimal recombination rates along the gene segment . This was achieved by splitting the gene segment into two , and calculating each subsegments optimal recombination rate . The location of the split was optimised along marker positions 2 to 5 . We find that the recombination rate is higher towards the marker site 1 end and lower towards marker site 6 . The optimal location for recombination rate switch was at marker site 4 ( 1 . 95×10−3 and 0 . 49×10−3 REPN from sites 1–4 and 4–6 respectively ) ( Figure 3B ) . Comparing the dual recombination rate model to the original model with an F-test did not produce a significant p value ( p = 0 . 30 , Figure 3B ) , indicating that the dual recombination rate model did not fit significantly better to justify the additional parameters ( second recombination rate and switch location ) . To address this further , we analyzed a second set of data and sequenced 192 cDNA strands . Again , we found that recombination is higher towards marker site 1 and lower towards marker site 6 ( Figure 3C and 3D ) . However , an F-test comparing the one and two recombination rate models in this dataset , but this time applying the same switch location estimated in the first experiment ( one less parameter in the two recombination rate model ) , still did not achieve significance ( p = 0 . 068 ) . Thus , our data support a difference in recombination rate across the gene , but was unable to identify the precise ‘hotspots’ of higher recombination . The assumption of an equal recombination rate amongst all sequences predicts that the frequency of multiple recombination events should be Poisson distributed . However , due to the possibility of multiple template switches occurring between markers of varying genomic distances , and the possibility of varying recombination rate across the gag gene , the frequency of multiple detectable template switches does not follow a Poisson distribution . We calculated this distribution to compute the expected frequency of multiple detectable template switches and compare this to our experimental results ( Figure 4 ) . This calculation compensates for multiple crossovers and uses the individual recombination rate observed in each region . Our data indicates some variation from the expected frequency of multiple template switches , however this was not significant ( p = 0 . 096 ) . Finally , it is possible that the limited introduction of marker points into the HIV genome altered the RNA structure in such a way as to bias the recombination process . For example , reverse transcription commencing on the MK genome may be more likely to result in recombination than reverse transcription on the WT genome due to our codon modifications . This predicts that the probability of recombination will be different when the RT is reverse transcribing a WT or MK marker point . Therefore , we compared the proportion of recombination events where recombination occurred from MK to WT , versus from WT to MK in our sequences . Of the 90 template switches observed in the pooled dataset , 42 were MK to WT and 48 were WT to MV , consistent with the null expectation of 50∶50 ( p = 0 . 60 , binomial distribution ) . This illustrates that our codon-modified markers have not significantly altered the RNA structure so as to bias the observed recombination rate .
Recombination plays an instrumental role in the evolution of HIV [39] , [40] and continues to shape the global pandemic [41] . Despite the excellent progress made in understanding inter-subtype recombination [30] , [42] , [43] , the study of recombination occurring between closely related genomes within an infected individual has been hampered by the lack of an appropriate model . Existing recombination systems are based on foreign reporter sequences , inter-subtype HIV genomes and/or intra-subtype HIV genomes with variation in amino acid sequences . Therefore , we have developed a novel marker system and associated mathematical tools that: ( i ) measures the recombination rate directly on the HIV genome; ( ii ) controls for background recombination; ( iii ) corrects for multiple template switching . Our HIV recombination marker system uses genetic marker points based on the codon modification of the authentic full length HIV genome without altering the amino acid sequences . Other groups have previously measured recombination rates on the HIV genome using the divergent RNA sequences found between or within HIV subtypes or within non-viral reporter sequences . Our procedure has several advantages over previously published methods . First , we rationally introduced marker points into the HIV genome at well defined locations , by avoiding RNA sequences that are known to be important for HIV replication . These marker points allow recombination to be monitored , but do not affect the HIV replication cycle , even over multiple rounds of replication ( Figure S1 ) . This is in contrast to recombination systems using divergent RNA sequences from different viral strains , where the differential replication capacities of the virus may bias the outcome of recombination . Second , our marker system retains every virion protein and these are expressed in their correct biological context . In the case of the retroviral reporter systems , it is common to completely knockout one or more HIV proteins by replacing them with non-viral reporter protein sequence . Therefore , these reporter systems , even when attempts are made to reintroduce these proteins back into the virion , do not recapitulate the exact biological conditions occurring in the full length virus [22] , [24] . Third , our silent modifications do not change the amino acid sequence of the viral proteins . This is important in light of reports that the amino acid sequence of the HIV RT affects the rate of template switching [32] , [34] and that mutations in the Gag polyprotein can affect RNA packaging and recombination [44] . Therefore , it seems likely that variations in the amino acid sequence of any viral protein involved in either assembly or reverse transcription of the virus could have unintentional consequences on the rate of recombination . This would limit the utility of divergent RNA sequences , even from within the same subtype [45] . Fourth , by limiting our modifications to targeted regions of the genome , we aim to maintain overall RNA structure and homology , which are critical determinants of recombination [28] , [29] , [46] . We demonstrated that recombination occurred at an equal rate on our WT and MK genome; hence , our modifications do not change the rate of recombination . This indicates that the variations in the recombination rate we observe are due to differences in the RNA sequence between marker points , not to the marker points themselves . We acknowledge that there are experimental complexities associated with the direct measurement of recombination by sequencing that can lead to the inclusion of non-viral recombination artifacts . Therefore , we carefully controlled for transfection-induced recombination , PCR-induced recombination and the effects of co-infection due to inter-virion recombination . In our study , transfection-induced recombination can be excluded as a source of error . We also show that inter-virion recombination , due to multiple infections of a cell , is not a significant source of error . By contrast , most retroviral reporter systems are biased by multiple infections . That is , in most retroviral reporter systems , multiple infections cannot be distinguished from single infections . This decreases the apparent total number of infection events , which is required to accurately calculate the recombination rate . To overcome this , these systems make use of MOI calculations which compensates for multiple infections . However , MOI calculations assume that infection events are independent and random . This is problematic in light of reports that double-infection occurs more frequently than predicted from random chance alone [47] , [48] , although this effect has been challenged by other data [22] and mathematical analysis [49] . Nevertheless , our system has the advantage that the recombination rate calculations are not affected by the occurrence of multiple infections . Finally , we did detect some recombination due to PCR-induced recombination but were able to optimize our PCR cycling conditions to minimize its effects . In addition , our recombination rate calculation corrects for this background to reveal the true rate of recombination . This highlights the necessity of including appropriate controls , as the effect of PCR-induced recombination has been ignored in similar studies [8] , [22] , [50] , [51] . As recombination is measured by observing the linking of genetic marker points , all recombination systems are potentially biased by the occurrence of multiple template switches . A potential solution is to reduce the genomic distance between marker points and to evenly space them on the HIV genome . This effectively eliminates multiple template switches and any bias due to variations in genomic distance between marker points . However , it is impossible to modify the HIV genome in this way without drastically affecting the replication cycle . As a result , modifications that do not affect important RNA sequences or vary the amino acid sequences of viral proteins will always be unevenly spaced . Furthermore , increasing the frequency of marker points increases the genetic diversity between co-packaged RNAs . This is expected to decrease the observed recombination rate , as high levels of sequence identity between templates is required for efficient template switching [46] , [52] . Thus , whilst reducing the genomic distance between markers can improve the ability to detect recombination , it also biases the observation by decreasing the likelihood of template switching in the first place . As multiple template switches between any two marker points occur , by definition , between identical sequences , these switches take place under optimal conditions for recombination . Therefore , a better solution is to compensate for multiple crossovers when calculating the recombination rate , as we have done . We also calculate the theoretical estimate for the error when recombination is measured without compensating for multiple template switches and show that the width between marker points can dramatically affect the crude recombination rate estimation . For example , when the distance between marker points is 400 base pairs , an actual recombination rate of 0 . 001 REPN and 0 . 003 REPN would be crudely calculated to be 0 . 0007 REPN and 0 . 0011 REPN , respectively . This is a difference that could be interpreted as resulting from random variation alone . This effect becomes more important at higher recombination rates . This is especially significant as recombination has been reported to be 3-fold higher in macrophages than in T-cells [22] , although this has been disputed by another group [18] . Regardless , we find that HIV undergoes 1 . 35×10−3 REPN in primary T-cells , which is a high rate of recombination , equivalent to 12 . 5 recombination events per genome every replication cycle . This is higher than when we apply our method to the data in Rhodes 2005 [24] ( average recombination rate 0 . 69×10−3 ) . However the measurements in our study are based on the HIV genome rather than non-viral reporter genes in previous studies [18] , [22] . The utility of our optimal recombination rate calculation is demonstrated by the fact that the crude calculation underestimates the optimal recombination rate by 44% . In addition , our bootstrapping procedure determines the confidence intervals of the recombination rate estimate . As these confidence intervals are derived from the actual data set and take into account variable distances between markers , it enables the direct comparison of recombination rates under different experimental conditions as well as providing an additional level of accuracy to our estimation . Our measurements on the HIV genome demonstrate that recombination does not occur randomly . Firstly , our results suggests that two or more recombination events on the same RNA strand may be observed more frequently than expected , although this was not statistically significant ( p = 0 . 096 , Figure 4 ) . This is in line with previous work showing HIV recombination exhibiting negative interference , which is when a single recombination event increases the chance of a second recombination event taking place [8] , although this is not universally agreed upon by all researchers [18] . Secondly , in two independent datasets ( from different blood donors ) , the recombination rate appears to be lower towards marker position 6 compared to position 1 within the gag gene . We tested whether a dual recombination rate model fitted the data better , but this did not reach significance at the 0 . 05 level ( p = 0 . 068 ) . Therefore , the data imply that there are recombination ‘hot-’ and/or ‘cold-spots’ within the genome but this current dataset was not large enough to identify how , or precisely where , the recombination rate changes . Interestingly , comparative sequence analysis of inter-subtype recombinants also showed a reduction in recombination near the 3′ end of gag [43] . Although this could be due to selection , our study opens up the possibility that this region of the genome may be inherently less prone to recombination . Further studies with much larger sequence numbers will be required to determine the positions of various recombination ‘hot/coldspots’ , and their respective recombination rates . The requirement of large quantities of sequencing data is a major limitation of our analytical tool . However , with the availability of next generation sequencing technology , plus the design of a marker system that has more marker points ( higher level of resolution ) , these issues can be readily accommodated . We have now developed appropriate statistical tools to quantify the rate of retroviral recombination taking into account the experimental procedures involved in observing recombination . We have shown how this can be used to compare recombination rates and to identify recombination hotspots within the viral genome . We also test a number of underlying biological and analytical assumptions that are often overlooked . These methods take into account the experimental and biological complexities of measuring recombination , and will provide a strong quantitative foundation for future studies in this area .
Human primary cells were isolated from buffy packs from random ( identity blocked ) blood donors to the Red Cross Blood Bank . All biological samples were handled according to the Burnet Institute and the Alfred Hospital approved ethics guidelines that are in line with Australian Government regulation . Homozygous virus was produced by transfection of 293T cells with either WT or MK pNL4-3 . Heterozygous virus was produced by co-transfection of equal amounts of wild-type pNL4-3 and marker pNL4-3 into 293T cells . Transfections were carried out with polyethylenimine ( PEI; Polysciences ) , and transfection efficiencies were measured using a reverse transcriptase assay [53] , [54] . 36 hours post-transfection , virus containing media was harvested , clarified by centrifugation at 1 , 462×g for 30 minutes , and then passed through a 0 . 45µm filter to remove cellular debris . Purified virus was concentrated by ultracentrifugation at 100 , 000×g through a 20% sucrose cushion and stored at −80°C . Virus was treated with 90units/mL benzonase ( Sigma ) for 15 minutes at 37°C to remove contaminating plasmid DNA before use . Stimulated PBLs were infected with equal amounts of either homozygous or heterozygous virus , as determined by a HIV-1 antigen ( p24 CA ) micro ELISA assay ( Vironostika ) . Heat inactivated ( 2 hours at 56°C ) control infections were carried out to confirm efficient removal of plasmid DNA for each sample . 6 hours post-infection 10µg/mL T-20 ( Roche ) was added to the cells to prevent second round replications . 24 hours post-infection cells were pelleted , lysed and full length reverse transcriptase products were quantified , as previously described [55] . A 1kb fragment of gag was PCR amplified using the primers ( EcoRI ) NL3065s [GCAgaattcGAGCTAGAACGATTCGCAG] and ( BamHI ) NL4066a [TATggatccTGGATTTGTTACTTGGCTCATTG] and the following conditions: initial denaturation 98°C for 30 seconds , followed by 30 rounds of cycling at 98°C for 10 seconds and 72°C for 2 minutes . PCR amplification was done in the log-linear phase as determined by real-time PCR to minimize PCR induced recombination . The fragment was cloned into pGem7z ( Promega ) and sequenced using the M13F primer on an Applied Biosystems 3730×l ( Australian Genome Research Facility ) . Recombination events were identified by sequence analysis . Controls were carried out to quantify the background rate of recombination produced by the experimental protocol itself . Transfection induced recombination was measured by harvesting plasmid DNA 36 hours post transfection directly from 293T by alkaline lysis as in plasmid DNA preparation from bacterial cells [56] . Plasmid DNA was directly sequenced with NL2944 [AGAGATGGGTGCGAGAG] after isolation by transformation of E . coli . Wild-type HIV-1 pNL4-3 plasmid was obtained from the National Institutes of Health AIDS Research and Reference Reagent program , Division of AIDS , NIAID , NIH: pNL4-3 from Dr . Malcolm Martin [57] . Marker HIV-1 pNL4-3 plasmid was created through the introduction of six restriction sites in gag by site directed mutagenesis [58] , [59] , All six sites are codon optimized and have not changed the protein coding sequence , and are separated by 128 , 77 , 86 , 398 , 228 base pairs ( Figure 2A and 2B ) . The location of the marker points is determined , in part , by the limited number of locations on the HIV-1 genome where restriction sites can be successfully introduced without changing protein coding sequence . 293T cells were obtained from the American Type Culture Collection and maintained in DMEM media ( Invitrogen ) supplemented with 10% vol/vol CCS ( Hyclone ) and Pen/Strep ( Invitrogen ) . Primary human peripheral lymphocytes ( PBLs ) were isolated from two independent buffy coats of HIV-1 seronegative blood donors ( Red Cross Blood Bank Service , Melbourne ) by density gradient centrifugation over Ficol-Plaque Plus ( GE Healthcare ) . PBLs were isolated by counter-current elutriation . The purity of PBLs was assessed by flow cytometry ( FACs Calibur; Becton Dickinson ) and determined to be >95% pure based on forward scatter and side scatter characteristics . PBLs were stimulated for 2–3 days in RPMI-1640 ( Invitrogen ) supplemented with 10µg/mL phytohemagglutinin and transferred into fresh RPMI-1640 containing 50 units/mL Interleukin-2 ( Roche ) before infection . We co-transfect equal amounts of WT and MK DNA , in order to produce heterozygous virions . Assuming random co-packaging of viral RNA templates we expect that 50% of the synthesized cDNA to have derived from heterozygous sequences . However , differences in the proportions of the WT and MK sequences may affect the proportion of heterozygous virions ( resulting in incorrect estimation of recombination rate ) . We calculate the expected proportion of heterozygous virions from the experimental data as follows . Let PW and PM be the proportion of experimentally observed nucleotide sequence data that is completely WT and MK respectively . PW and PM represents cDNA derived from homozygous WT and MK virions , and also cDNA derived from heterozygous virions in which recombination was not observed . Now let F be the fraction of cDNA derived from heterozygous virions that did not observe recombination . We then havewhere w and m are the proportion of WT and MK constructs that were cotransfected into T cells to create the virions . Noting that m = 1-w allows for solving the expected proportion of cDNA derived from heterozygous virions , 2mw . Thus , we do not need to rely on the estimated proportion of WT:MK virus , but can directly estimate it from our data . We measure recombination by infection with two nearly identical HIV-1 viruses , denoted WT and MK , which differ at a number of marker positions in gag . Recombination is observed when a single sequence of DNA product contains both WT and MK markers . However , multiple template switches can occur between marker positions , and recombination can only be detected when there are an odd number of template switches . Thus , it is impossible to work out the exact frequency of recombination events . Rather , the data shows the probability of observing recombination ( a switch from WT to MK between markers or vice versa ) which is calculated as the number of recombination events observed divided by the number of sequences derived from heterozygous infection . Denote the probability of observing a recombination event between two marker positions separated by a genomic distance of L as R ( L ) . Denote the recombination rate per nucleotide per round of infection as r . These two quantities then satisfy the following . where [ ( L+1 ) /2] is the integer part of ( L+1 ) /2 and C ( L , i ) is the binomial coefficient for picking i unordered outcomes from L possibilities . Alternatively , when the genomic distance L is sufficiently large and the recombination rate r is sufficiently small ( as is generally the case with recombination experiments ) the following Poisson approximation holds [36] ( see ‘Poisson approximation’ ) which can be re-arranged to calculate the recombination rate as ( A ) Finally if recombination is studied over multiple regions of lengths L1 , L2 , L3 , … , Lk , then the recombination rate , r , is calculated as the r value that minimises the chi-square value ( B ) Where Oi and Ei is the observed and expected number of template switches that is detected in region i respectively . The expected number of template switches is calculated as the multiple of R ( Li ) and the number of heterozygous sequences . Probability distributions were generated by bootstrapping the sequence data as follows . In each bootstrap loop , sequence data was randomly sampled with replacement until the same number of sequences that were originally sampled , had been sampled in silico . From each new sample set the optimal rate of recombination was calculated as described above . This bootstrapping procedure was completed 10000 times and pooling each bootstrap loop generates a probability distribution for the recombination rate , r . Note that we sampled from the entire sequence pool , and thus this approach also incorporates the level of uncertainty in the proportion of heterozygous virus in the sample . The probability distributions of the recombination rate for different genetic constructs/target cells was used to compare rates . Let s ( n ) be the probability that a single cell has been infected n times with the HIV reporter virus . Let m be the MOI for the HIV reporter virus . Let pGFP be the probability of a single infection event resulting in the reconstruction of a functional GFP encoding region . That is , the probability that an infecting virion is heterozygous , and that recombination occurred between the two co-packaged RNAs such that both GFP deactivating mutations are eliminated . Then , the probability that a single cell has n infections that reconstitute a function GFP is given bywhich is the MOI formula ( Poisson distribution ) with MOI equal to the product , pGFPm . Note that C ( n+i , i ) is the binomial coefficient for picking i unordered outcomes from n+i possibilities . Thus , the GFP MOI equals pGFPm and division of the infection MOI , m , leaves pGFP , the probability that an infection event will reconstitute a functional GFP encoding region . In this assay , recombination can occur at two independent stages: The experimentally induced recombination , and the viral reverse transcription induced recombination . We measure the experimentally induced recombination alone , and the cumulative effect of experimentally induced recombination with the reverse transcription induced recombination . From this we calculate the reverse transcription induced recombination rate alone as follows . Let RE ( L ) and RR ( L ) be the probability of observing recombination over a genomic distance of L for the experimentally induced and RT induced recombination rates respectively . The cumulative probability of observing recombination after both effects , R ( L ) , is given byNote that RERR is subtracted once as RE and RR are independent and not mutually exclusive events , and subtracted a second time to eliminate the cases where PCR template switch nullifies an RT template switch . This is then re-arranged to giveThe recombination rate is calculated from equation ( A ) . If recombination is measured over multiple regions , as is the case in our experimental system , this should be applied to each region before calculating the recombination rate by minimizing the chi-square value ( equation B ) . The binomial terms Pi ( L ) above can be approximated by the Poisson distribution when the length , L , is sufficiently large , and the recombination rate , r , is sufficiently small . Under these conditions the Poisson coefficient is the product of the genomic length and recombination rate Lr . The probability of observing recombination is then approximated byThus , to calculate the recombination rate r from experimental data we re-arrange to givewhere R ( L ) can be measured from experimental data as the proportion of heterozygous sequences over which recombination was observed .
|
HIV's ability to generate and maintain high genetic diversity leads to multiple drug resistances and evasion from the immune system , eventually leading to immune failure and progression to AIDS . HIV maintains this diversity with a process of mutation ( incorrect copying of genetic information in viral replication ) and recombination ( mixing two viral genomes in the creation of viral offspring ) . Recombination is generally studied by inserting genes encoding non-viral fluorescent proteins . However , recombination in such modified HIV genomes may not accurately reflect the level of recombination occurring within a patient infected with HIV . Additionally , recombination will go undetected in regions where the parental genomes are identical , and this effect is often ignored . We have developed a novel experimental system which allows recombination to be measured between two very closely related HIV genomes . We have also developed statistical tools to accurately calculate the recombination rate , compensating for undetectable recombination in identical regions of the parental genomes . We show that our experimental system bypasses some of the pitfalls of fluorescent recombination experiments and our tools provide a strong quantitative foundation for future studies in this area .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases/hiv",
"infection",
"and",
"aids",
"mathematics/statistics",
"virology/viral",
"replication",
"and",
"gene",
"regulation",
"computational",
"biology"
] |
2010
|
Accurately Measuring Recombination between Closely Related HIV-1 Genomes
|
Advances in multi-unit recordings pave the way for statistical modeling of activity patterns in large neural populations . Recent studies have shown that the summed activity of all neurons strongly shapes the population response . A separate recent finding has been that neural populations also exhibit criticality , an anomalously large dynamic range for the probabilities of different population activity patterns . Motivated by these two observations , we introduce a class of probabilistic models which takes into account the prior knowledge that the neural population could be globally coupled and close to critical . These models consist of an energy function which parametrizes interactions between small groups of neurons , and an arbitrary positive , strictly increasing , and twice differentiable function which maps the energy of a population pattern to its probability . We show that: 1 ) augmenting a pairwise Ising model with a nonlinearity yields an accurate description of the activity of retinal ganglion cells which outperforms previous models based on the summed activity of neurons; 2 ) prior knowledge that the population is critical translates to prior expectations about the shape of the nonlinearity; 3 ) the nonlinearity admits an interpretation in terms of a continuous latent variable globally coupling the system whose distribution we can infer from data . Our method is independent of the underlying system’s state space; hence , it can be applied to other systems such as natural scenes or amino acid sequences of proteins which are also known to exhibit criticality .
We represent the response of a neural population with a binary vector s = {s1 , s2 , … , sN} ∈ {0 , 1}N identifying which of the N neurons elicited at least one action potential ( ‘1’ ) and which stayed silent ( ‘0’ ) during a short time window . Our goal is to build a model for the probability distribution of activity patterns , p ( s ) , given a limited number M of samples , D = { s ( 1 ) , … , s ( M ) } , observed in a typical recording session . The regime we are mainly interested in is the one where the dimensionality of the problem is sufficiently high that the distribution p cannot be directly sampled from data , i . e . , when 2N ≫ M . Note that we are looking to infer models for the unconditional distribution over neural activity patterns ( i . e . , the population “vocabulary” ) , explored in a number of recent papers [8 , 9 , 11 , 13–18 , 24 , 34] , rather than to construct stimulus-conditional models ( i . e . , the “encoding models” , which have a long tradition in computational neuroscience [1–3] ) . Previous approaches to modeling globally coupled populations focused on the total network activity , also known as synchrony , K ( s ) = ∑ i = 1 N s i . The importance of this quantity was first analyzed in the context of probabilistic models in Ref [11] where the authors showed that a K-pairwise model , which generalizes a pairwise maximum entropy model by placing constraints on the statistics of K ( s ) , is much better at explaining the observed population responses of 100+ salamander retinal ganglion cells than a pairwise model . Specifically , a pairwise model assumes that the covariance matrix between single neuron responses , Cij = 〈sisj〉 , which can be determined empirically from data D , is sufficient to estimate the probability of any population activity pattern . In the maximum entropy framework , this probability is given by the most unstructured ( or random ) distribution that reproduces exactly the measured Cij: p ( s ; J ) = 1 Z ( J ) exp ( ∑ i , j = 1 N J i j s i s j ) , ( 1 ) where Z ( J ) is a normalization constant , and J is a coupling matrix which is chosen so that samples from the model have the same covariance matrix as data . Note that because s i 2 = s i , the diagonal terms Jii of the coupling matrix correspond to single neuron biases , i . e . firing probabilities in the absence of spikes from other neurons ( previous work [11] used a representation si ∈ {−1 , 1} for which the single neuron biases need to be included as separate parameters and where Jii are all 0 ) . A K-pairwise model generalizes the pairwise model and has the form p ( s ; J , ϕ ) = 1 Z ( J , ϕ ) exp ( ∑ i , j = 1 N J i j s i s j + ∑ k = 0 N ϕ k δ k , K ( s ) ) . ( 2 ) The coupling matrix J has the same role as in a pairwise model while the additional parameters ϕ are chosen to match the probability distribution of K ( s ) under the model to that estimated from data . The “potentials” ϕk introduced into the K-pairwise probabilistic model , Eq ( 2 ) , globally couple the population , and cannot be reduced to low-order interactions between , e . g . , pairs or triplets , of neurons , except in very special cases . We will generically refer to probabilistic models that impose non-trivial constraints on population-level statistics ( of which the distribution of total network activity K is one particular example ) as “globally coupled” models . Here we introduce new semiparametric energy-based models that extend the notion of global coupling . These models are defined as follows: p ( s ; α , V ) = e - V ( E ( s ; α ) ) Z ( α , V ) , ( 3 ) where E ( s; α ) is some energy function parametrized by α , and V is an arbitrary increasing differentiable function which we will refer to simply as the “nonlinearity . ” The parametrization of the energy function should be chosen so as to reflect local interactions among neurons . Crucially , while it is necessary to choose a specific parametrization of the energy function , we do not make any assumptions on the shape of the nonlinearity—we let the shape be determined nonparametrically from data . Fig 1 schematically displays the relationship between the previously studied probabilistic models of population activity and two semiparametric energy-based models that we focus on in this paper , the semiparametric independent model ( which we also refer to as “V ( independent ) ” ) and the semiparametric pairwise model ( which we also refer to as “V ( pairwise ) ” ) . Our motivation for introducing the global coupling via the nonlinearity V traces back to the argument made in Ref [11] for choosing to constrain the statistics of synchrony , K ( s ) ; in short , the key intuition in earlier work has been that K ( s ) is a biologically relevant quantity which encodes information about the global state of a population . There are , however , many other quantities whose distributions could contain signatures of global coupling in a population . In particular , while most energy functions—e . g . , the pairwise energy function , E ( s; J ) = −∑i , j Jijsisj—are defined solely in terms of local interactions between small groups of neurons , the statistics of these same energy functions ( for instance , their moments ) are strongly shaped by global effects . Specifically , we show in Methods that the role of the nonlinearity in Eq ( 3 ) is precisely to match the probability density of the energy under the model to that estimated from data . In other words , once any energy function for Eq ( 3 ) has been chosen , the nonlinearity V will ensure that the distributions of that particular energy in the model and over data samples agree . Constraining the statistics of the energy E ( s; α ) is different from constraining the statistics of K ( s ) , used in previous work . First , the energy depends on a priori unknown parameters α which must be learned from data . Second , while K ( s ) is always an integer between 0 and N , the energy can take up to 2N distinct values; this allows for extra richness but also requires us to constrain the ( smoothed ) histogram of energy rather than the probability of every possible energy value , to prevent overfitting . As we discuss next , the statistics of the energy are also closely related to criticality , a formal , model-free property distinguishing large , globally-coupled neural populations . The notion of criticality originates in thermodynamics where it encompasses several different properties of systems undergoing a second-order phase transition [35] . Today , many other phenomena , such as power-law distributed sizes of “avalanches” in neural activity , have been termed critical [20] . Our definition , which we discuss below , is a restricted version of the thermodynamic criticality . We consider a sequence of probability distributions { p N } N = 1 ∞ over the responses of neural populations of increasing sizes , N . These probability distributions define the discrete random variable s ( the population response ) , but they can also be thought of simply as functions which map a population response to a number between 0 and 1 . Combining these two viewpoints , we can consider a real-valued random variable pN ( s ) ∈ ( 0 , 1 ) which is constructed by applying the function pN to the random variable s . The behavior of this random variable as N → ∞ is often universal , meaning that some of its features are independent of the precise form of pN . As is conventional , we work with the logarithm of pN ( s ) instead of the actual distribution . We call a population “critical” if the standard deviation of the random variable log pN ( s ) /N does not vanish as the population size becomes large , i . e . 1 N σ ( log p N ( s ) ) ↛ 0 as N → ∞ . ( 4 ) ( For completeness , we further exclude some degenerate cases such as when the probability density of log pN ( s ) /N converges to two equally sized delta functions . ) The above definition is related to criticality as studied in statistical physics . In thermodynamics , σ ( log p N ( s ) ) / N is proportional to the square root of the specific heat , which diverges in systems undergoing a second-order phase transition . While at a thermodynamical critical point σ ( log pN ( s ) ) /N scales as N−γ with γ ∈ ( 0 , 1/2 ) , here we are concerned with the extreme case of γ = 0 . Rather than being related to second-order phase transitions , this definition of criticality is related to the so-called Zipf law [23] . A pattern s can be assigned a rank by counting how many other patterns have a higher probability . In its original form , a probability distribution is said to satisfy Zipf law if the probability of a pattern is inversely proportional to its rank . No real probability distribution is actually expected to satisfy this definition precisely , but there is a weaker form of Zipf law which concerns very large populations , and which is much less restrictive . This weaker form can be stated as a smoothed version of the original Zipf law . Consider patterns whose rank is in some small interval [r , r + ΔN] , and denote pN ( r ) the average probability of these patterns . We generalize the notion of Zipf law to mean that for very large populations pN ( r ) ∝ r−1 ( ΔN is assumed to go to zero sufficiently quickly with N ) . As shown in Ref [23] , a system is critical in the sense of Eq ( 4 ) precisely when it follows this generalized Zipf law . Practically speaking , no experimentally studied population ever has an infinite size , and a typical way to check for signs of criticality is to see if a log-log plot of a pattern probability versus its rank resembles a straight line with slope −1 . Most systems are not expected to be critical . The simplest example is a population of identical and independent neurons , p N ( s ) = q ∑ i = 1 N s i ( 1 - q ) N - ∑ i = 1 N s i , ( 5 ) where q is the probability of eliciting a spike . For such population , 1 N σ ( log p N ( s ) ) = 1 N q ( 1 - q ) log q 1 - q , ( 6 ) which vanishes for very large number of neurons , and so the system is not critical . More generally , if pN ( s ) can be factorized into a product of probability distributions over smaller subpopulations which are independent of each other and whose number is proportional to N , then log pN ( s ) /N turns into an empirical average whose standard deviation is expected to vanish in the large N limit , and the population is not critical . Reversing this argument , signatures of criticality can be interpreted as evidence that the population is globally coupled , i . e . that it cannot be decomposed into independent parts . These preliminaries establish a direct link between criticality and semiparametric energy models of Eq ( 3 ) . Nonlinearity in semiparametric energy models makes sure that the statistics of the energy E ( s; α ) , and , since V ( E ) is monotone , also the statistics of log p ( s; α , V ) are modeled accurately ( see Methods ) . Because the behavior of log probability is crucial for criticality , as argued above , semiparametric energy models can capture accurately and efficiently the relevant statistical structure of any system that exhibits signs of criticality and/or global coupling . To fully specify semiparametric energy models , we need a procedure for constructing the nonlinearity V ( E ) . We cannot let this function be arbitrary because then the model could learn to assign nonzero probabilities only to the samples in the dataset , and hence it would overfit . To avoid such scenarios , we will restrict ourselves to functions which are increasing . We also require V ( E ) to be differentiable so that we can utilize its derivatives when fitting the model to data . The class of increasing differentiable functions is very large . It includes functions as diverse as the sigmoid , 1/ ( 1 + exp ( −E ) ) , and the square root , E ( for positive E ) , but we do not want to restrict ourselves to any such particular form—we want to estimate V ( E ) nonparametrically . Nonparametric estimation of monotone differentiable functions is a nontrivial yet very useful task ( for example , consider tracking the height of a child over time—the child is highly unlikely to shrink at any given time ) . We follow Ref [36] and restrict ourselves to the class of strictly monotone twice differentiable functions for which V′′/V′ is square-integrable . Any such function can be represented in terms of a square-integrable function W and two constants γ1 and γ2 as V ( E ) = γ 1 + γ 2 ∫ E 0 E exp ( ∫ E 0 E ′ W ( E ′ ′ ) d E ′ ′ ) d E ′ , ( 7 ) where E0 is arbitrary and sets the constants to γ1 = V ( E0 ) , γ2 = V′ ( E0 ) . The function is either everywhere increasing or everywhere decreasing ( depending on the sign of γ2 ) because the exponential is always positive . Eq ( 7 ) is easier to understand by noting that V ( E ) is a solution to the differential equation V′′ = WV′ . This means , for example , that on any interval on which W = 0 , the equation reduces to V′′ = 0 , and so V ( E ) is a linear function on this interval . If V ( E ) is increasing ( V′ > 0 ) , it also shows that the sign of W at a given point determines the sign of the second derivative of V at that point . An advantage of writing the nonlinearity in the form of Eq ( 7 ) is that we can parametrize it by expanding W in an arbitrary basis without imposing any constraints on the coefficients of the basis vectors yet V ( E ) is still guaranteed to be monotone and smooth . In particular , we will use piecewise-constant functions for W . This allows us to use unconstrained optimization techniques for fitting our models to data .
We start by considering one of the simplest models of the form Eq ( 3 ) , the semiparametric independent model: p ( s ; α , V ) = e - V ( - ∑ i = 1 N α i s i ) Z ( α , V ) . ( 8 ) If V were a linear function , the model would reduce to an independent model , i . e . a population of independent neurons with diverse firing rates . In general , however , V introduces interactions between the neurons that may not have a straightforward low-order representation . When fitted to our data , the nonlinearity V turns out to be a concave function ( see later sections on more complex models for a detailed discussion of the shape of the nonlinearity ) . Note that if V had a simple functional form such as a low order polynomial , then the model Eq ( 8 ) would be closely related to mean field models of ferromagnetism with heterogenous local magnetic field studied in physics . Our first goal is to use this simple model to verify our intuition that the nonlinearity helps to capture criticality . Many population patterns are observed several times during the course of the experiment , and so it is possible to estimate their probability simply by counting how often they occur in the data [19] . Given this empirical distribution , we construct a corresponding Zipf plot—a scatter plot of the frequency of a pattern vs its rank . For systems which are close to critical , this should yield a straight line with slope close to −1 on a log-log scale . We repeat the same procedure with samples generated from a semiparametric independent model as well as an independent model , which were both fitted to the responses of all 160 neurons . Fig 2 shows all three scatter plots . The independent model vastly deviates from the empirical Zipf plot; specifically , it greatly underestimates the probabilities of the most likely states . In contrast , the learned semiparametric independent model follows a similar trend to that observed in data . This does not mean that the semiparametric independent model itself is an excellent model for the detailed structure in the data , but it is one of the simplest possible extensions of the trivial independent model that qualitatively captures both global coupling and the signatures of criticality . Since the semiparametric independent model is able to capture the criticality of the data distribution , we also expect it to accurately model other features of the data which are related to the globally coupled nature of the population . To verify this , Fig 3A compares the empirical probability distribution of the total activity of the population K ( s ) = ∑i si to that predicted by the semiparametric independent model . The match is very accurate , especially when compared to the same distribution predicted by the independent model . This result goes hand in hand with the analysis in [39] which showed that interactions of all orders ( in our case mediated by the nonlinearity ) are necessary to model the wide-spread distribution of the total activity . The independent model is a maximum entropy model which constrains the mean responses , 〈si〉 , of all neurons . In other words , neurons sampled from the model would have the same firing rates as those in the data ( up to sampling noise ) . Even though the semiparametric independent model is strictly more general , it does not retain this property when the parameters α and the nonlinearity V are learned by maximizing the likelihood of data . Fig 3B demonstrates this point: although the predicted firing rates are approximately correct , there are slight deviations . On the other hand , the nonlinearity induces pairwise correlations between neurons which is something the independent model by construction cannot do . Fig 3C compares these predicted pairwise correlations to their data estimates . While there is some correlation between the predicted and observed covariances , the semiparametric independent model often underestimates the magnitude of the covariances and does not capture the fine details of their structure ( e . g . the largest covariance predicted by the semiparametric independent model is about 5× smaller than the largest covariance observed in the data ) . This is because a combination of independent terms and a single nonlinearity does not have sufficient expressive power , motivating us to look for a richer model . One way to augment the power of the semiparametric independent model that permits a clear comparison to previous work is by means of the semiparametric pairwise model: p ( s ; J , V ) = 1 Z ( J , V ) exp ( - V ( - ∑ i , j = 1 N J i j s i s j ) ) . ( 9 ) We fit this model to the responses of the various subpopulations of the 160 neurons , and we compare the resulting goodness-of-fit to that of a pairwise ( Eq ( 1 ) ) , K-pairwise ( Eq ( 2 ) ) , and semiparametric independent model ( Eq ( 8 ) ) . We measure goodness-of-fit as the improvement of the log-likelihood of data per neuron under the model relative to the pairwise model , as shown in Fig 4A . This measure reflects differences among models rather than differences among various subpopulations . The semiparametric pairwise model consistently outperforms the other models and this difference grows with the population size . To make sure that this improvement is not specific to this particular experiment , we also fitted the models to two additional recordings from the salamander retina which were also collected as part of the study [11] . One consists of 120 neurons responding to 69 repeats of a 30 second random checkerboard stimulus , and the other of 111 neurons responding to 98 repeats of a 10 second random full-field flicker stimulus . As shown in Fig 4B , the improvements of individual models on these datasets are consistent with the ones observed for the population stimulated with a natural movie . The advantage of using likelihood as a goodness-of-fit measure is its universal applicability which , however , comes hand-in-hand with the difficulty of interpreting the quantitative likelihood differences between various models . An alternative comparison measure that has more direct relevance to neuroscience asks about how well the activity of a single chosen neuron can be predicted from the activities of other neurons in the population . Given any probabilistic model for the population response , we use Bayes rule to calculate the probability of the ith neuron spiking ( si = 1 ) or being silent ( si = 0 ) conditioned on the activity of the rest of the population ( s−i ) as p ( s i | s - i ; α ) = p ( s ; α ) p ( s i = 1 , s - i ; α ) + p ( s i = 0 , s - i ; α ) . ( 10 ) We turn this probabilistic prediction into a nonrandom one by choosing whether the neuron is more likely to spike or be silent given the rest of the population , i . e . s i ( s - i ; α ) = argmax s i ∈ { 0 , 1 } p ( s i | s - i ; α ) . ( 11 ) In Fig 4C and 4D we compare such predictive single neuron models constructed from semiparametric pairwise , K-pairwise , pairwise , and semiparametric independent models learned from the data for populations of various sizes . Specifically , we ask how often these models would make a mistake in predicting whether a chosen single neuron has fired or not . Every population response in our dataset corresponds to 20 ms of an experiment and so we can report this accuracy as number of errors per unit of time . Predictions based on the semiparametric pairwise model are consistently the most accurate . Fig 5A shows the nonlinearities of the semiparametric pairwise models that we learned from data . In order to compare the nonlinearities inferred from populations of various sizes , we normalize the domain of the nonlinearity as well as its range by the number of neurons . Even though the nonlinearities could have turned out to have e . g . a sigmoidal shape , the general trend is that they are concave functions whose curvature—and thus departure from the linear V that signifies no global coupling—grows with the population size . The shape of these nonlinearities is reproducible over different subnetworks of the same size with very little variability . To further visualize the increasing curvature , we extrapolated what these nonlinearities might look like if the size of the population was very large ( the black curve in Fig 5A ) . This extrapolation was done by subtracting an offset from each curve so that V ( 0 ) = 0 , and then fitting a straight line to a plot of 1/N vs . the value of V at points uniformly spaced in the function’s domain . The plots of 1/N vs . V are only linear for N ≥ 80 , and so we only used these points for the extrapolation which is read out as the value of the fit when 1/N = 0 . To quantify the increasing curvature , Fig 5B shows the average absolute value of the second derivative of V across the function’s domain . The coupling matrix J of both the pairwise and the semiparametric pairwise models describes effective interactions between neurons , and so it is interesting to ask how the couplings predicted by these two models are related . While Fig 5C shows a strong dependency between the couplings in a network of N = 160 neurons , the dependency is not deterministic and , moreover , negative couplings tend to be amplified in the semiparametric pairwise model as compared to the pairwise model . Similarly to the semiparametric independent model , there is no guarantee that the semiparametric pairwise model will reproduce observed pairwise correlations among neurons exactly , even though pairwise model has this guarantee by virtue of being a maximum entropy model . Fig 5D shows that despite the lack of such a guarantee , semiparametric pairwise model predicts a large majority of the correlations accurately , with the possible exceptions of several very strongly correlated pairs . This is simply because the semiparametric paiwise model is very accurate–the inset of Fig 5D shows that it can also reproduce third moments of the responses . A K-pairwise model also has this capability but , as shown in Ref [11] , a pairwise model systematically mispredicts higher than second moments . Suppose we use the semiparametric pairwise model to analyze a very large population which is not globally coupled and can be divided into independent subpopulations . The only way the model in Eq ( 9 ) can be factorized into a product of probability distributions over the subpopulations is if the function V is linear . Therefore , the prior knowledge that the population is not globally coupled immediately implies the shape of the nonlinearity . Similarly , a prior knowledge that the population is critical also carries a lot of information about the shape of the nonlinearity . We show in Methods that if the parameters α are known , then the optimal nonlinearity in Eq ( 3 ) can be explicitly written as V ( E ) = log ρ ¯ ( E ; α ) - log p ¯ ^ ( E ; α ) , ( 12 ) where ρ ¯ ( E ; α ) is the density of states which counts the number of patterns s whose energy is within some narrow range [E , E + Δ] . The density of states is a central quantity in statistical physics that can be estimated also for neural activity patterns either directly from data or from inferred models [19] . Similarly , p ¯ ^ ( E ; α ) is the empirical probability density of the energy E ( s; α ) smoothed over the same scale Δ . Eq ( 12 ) follows from the relation p ¯ ^ ( E ; α ) ∝ ρ ¯ ( E ; α ) exp ( - V ( E ) ) , i . e . the probability of some energy level is just the number of states with this energy times the probability of each of these states ( see Methods ) . We would like to establish a prior expectation on what the large N limit of the nonlinearites in Fig 5A is . Adapting the same normalization as in the figure , we denote ϵ ( s; α ) = E ( s; α ) /N . Changing variables and rewriting Eq ( 12 ) in terms of the empirical probability density of the normalized energy p ϵ ¯ ^ ( ϵ ) = N p ¯ ^ ( ϵ N ; α ) yields V ( ϵ N ) = log ρ ¯ ( ϵ N ; α ) - log p ϵ ¯ ^ ( ϵ ) + log N . ( 13 ) For a system where si can take on two states , the total number of possible activity patterns is 2N , and so we expect the log of the density of states to be proportional to N . If the system is critical , then by virtue of Eq ( 4 ) σ ( log pN ( s ) ) is proportional to N , and similarly we also expect σ ( E ( s; α ) ) ∝ N . This means that σ ( ϵ ( s; α ) ) = σ ( E ( s; α ) ) /N converges to some finite , nonzero number , and therefore log p ϵ ¯ ^ ( ϵ ) also stays finite no matter how large the population is . Taken together , for large critical populations , the first term on the right hand side of Eq ( 13 ) is the only one which scales linearly with the population size , and hence it dominates the other terms: V ( E ) ≈ log ρ ¯ ( E ; α ) . ( 14 ) One of our important results is thus that for large critical populations , the nonlinearity should converge to the density of states of the inferred energy model . In other words , for critical systems as defined in Eq ( 4 ) , there is a precise matching relation between the nonlinearity V ( E ) and the energy function E ( s; α ) ; in theory this is exact as N → ∞ , but may hold approximately already at finite N . To verify that this is the case for our neural population that has previously been reported to be critical , we compare in Fig 6A the nonlinearity inferred with the semiparametric pairwise model ( Fig 5A ) to the density of states estimated using a Wang and Landau Monte Carlo algorithm [40] for a sequence of subpopulations of increasing size . As the population size increases , the nonlinearity indeed approaches the regime in which our prediction in Eq ( 14 ) holds . This convergence is further quantified in Fig 6B which shows the average squared distance between the density of states and the nonlinearity . The average is taken over the range of observed energies . The nonlinearities are only specified up to an additive constant which we chose so as to minimize the squared distance between the density of states and the nonlinearity . The link between global coupling and criticality is related to recent theoretical suggestions [28 , 29] , where global coupling between the neurons in the population emerges as a result of shared latent ( fluctuating ) variables that simultaneously act on extensive subsets of neurons . In particular , Ref [28] theoretically analyzed models with a multivariate continuous latent variable h distributed according to some probability density q ( h ) , whose influence on the population is described by the conditional probability distribution p N ( s | h ) = e - ∑ j h j O j ( N ) ( s ) Z N ( h ) , ( 15 ) where ZN ( h ) is a normalization constant , and O j ( N ) ( s ) are global quantities which sum over the whole population . The authors showed that under mild conditions on the probability density q ( h ) of h , and the scaling of O j ( N ) ( s ) with N , the sequence of models p N ( s ) = ∫ q ( h ) p N ( s | h ) d h ( 16 ) is critical in the sense of Eq ( 4 ) . If the latent variable is one-dimensional , i . e . h = h , then the models in Eq ( 16 ) have exactly the form of models in Eq ( 3 ) with E ( s; α ) = O ( s ) , i . e . given a probability density q ( h ) of the latent variable , we can always find a nonlinearity V ( E ) such that 1 Z ( α ) e - V ( E ( s ; α ) ) = ∫ 0 ∞ q ( h ) e - h E ( s ; α ) Z ( h ; α ) d h . ( 17 ) The reverse problem of finding a latent variable for a given function V ( E ) such that this equation is satisfied does not always have a solution . The condition for this mapping to exist is that the function exp ( −V ( E ) ) is totally monotone [41] , which , among other things , requires that it is convex . While our models allow for more general nonlinearites , we showed in Fig 5A that the inferred functions V ( E ) are concave and so we expect this mapping to be at least approximately possible ( see below ) . The mapping in Eq ( 17 ) is based on a Laplace transformation , a technique commonly used for example in the study of differential equations . Laplace transformations are also often used in statistical physics where they relate the partition function of a system to its density of states . While the mathematics of Laplace transformations yields conditions on the function V ( E ) so that it is possible to map it to a latent variable ( i . e . , exp ( −V ( E ) ) must be totally monotone ) , analytically constructing this mapping is possible only in very special cases . We can gain a limited amount of intuition for this mapping by considering the case when the latent variable h is a narrow gaussian with mean h0 and variance σ2 . For small σ2 , one can show that V ( E ) ≈ h 0 E - σ 2 ( E - E 0 ) 2 , ( 18 ) where E0 is the average energy if σ2 = 0 , and the approximation holds only in a small neighborhood of E0 ( |E − E0| ≪ σ ) . This approximation shows that the curvature of V ( E ) is proportional to the size of the fluctuations of the latent variable which , in turn , is expected to correlate with the amount of global coupling among neurons . This relationship to global coupling can be understood from the right hand side of Eq ( 17 ) . When the energy function is , for example , a weighted sum of individual neurons as in the semiparametric independent model of Eq ( 8 ) , then we can think of Eq ( 17 ) as a latent variable h ( perhaps reflecting the stimulus ) coupled to every neuron , and hence inducing a coupling between the whole population . A non-neuroscience example is that of a scene with s representing the luminance in each pixel , and the latent h representing the lighting conditions which influence all the pixels simultaneously . We used the right hand side of Eq ( 17 ) ( see Methods ) to infer the shapes of the probability densities of the latent variables which correspond to the nonlinearities in the semiparametric pairwise models learned from data . These probability densities are shown in Fig 6C . A notable difference to the formulation in Eq ( 16 ) is that the inferred latent variables scale with the population size; in particular , the inset to Fig 6C shows that the entropy of the inferred latent variable increases with the population size . Entropy is a more appropriate measure of the “broadness” of a probability density than standard deviation when the density is multimodal . Taken together with the results in Fig 4A , this suggests that global coupling is especially important for larger populations . However , it is also possible that the latents are becoming broader because the model is trying to compensate for limited capacity , and that the entropy of the latent would saturate if we had a more expressive energy function . Larger datasets and/or further improvements in probabilistic models are necessary to make more detailed conclusions . Interestingly , the probability densities of the latent variables consist of two modes at approximately h = 0 . 7 and h = 1 . 3 . We hypothesize that these modes reflect a discrete-like nature of the population dynamics which consist of bursts of activity interspaced with periods of approximate silence . These bursts are demonstrated in Fig 7A where we show the time dependence of the total network activity . Unfortunately , closer inspection reveals that the total network activity cannot be used in a straightforward manner to classify the population as active or inactive . The reason is that neurons are noisy and if we defined a population as inactive when the total network activity is 0 , then such definition is not robust to noise . In fact , the probability distribution of the total network activity ( Fig 3A ) is such that there is no obvious choice of a threshold , and so quantifying the discreteness of the population dynamics based on the total network activity would be arbitrary . To circumvent these problems and enable a robust classification of the population state as active or inactive , we can use the most likely value of the latent variable given a population response , i . e . h * ( s ) = argmax h p ( h | s ) = argmax h p ( s | h ) q ( h ) = argmax h q ( h ) e - h E ( s ; α ) Z ( h ; α ) . ( 19 ) Fig 7A shows the time dependence of h* , and Fig 7B its probability density ( estimated by collecting h* ( s ) over all repeats and times ) . The probability density of h* has two modes separated by an inaccessible region , so one can easily classify a population response s as active or inactive based on which mode h* ( s ) belongs to . Fig 7C and 7D show that a population pattern with , for example , 5 active neurons can have very different values for h* ( s ) , demonstrating that any measure based on the total network activity would easily confuse which state the population is in .
Criticality is a theoretical concept which depends crucially on how the probability distribution over population activity patterns scales with the population size . Constructing this scaling directly from data is complicated , and necessarily involves extrapolating to large population sizes [10 , 30] . As a consequence , answering the question whether a population is critical or “how close to critical” it is , is difficult . Here we took a different approach—we used the theoretical notion of criticality to guide our intuition about what models are useful for analyzing populations that exhibit signs of criticality such as an approximate Zipf law . From the standpoint of fitting statistical models , it is irrelevant whether or not the studied population is really critical given some operational realization of the large population size limit because our models can be used either way , and their accuracy can be evaluated using standard model selection techniques . In particular , our approach is agnostic to the origins of the signatures of criticality which have been hotly debated [25–30 , 42] . Our reasoning is thus very pragmatic: we on purpose avoided the controversial ( albeit interesting ) issues of whether the observed critical behavior in real data is “trivial” or not and what may be its mechanistic explanation , and focused rather on making use of the observation itself to design better probabilistic models for neural code . This pragmatic approach is driven by the rapid development of experimental techniques for recording the activity of large neural populations , which is posing a challenge for data analysis . The number of neurons that we can measure simultaneously is growing much faster than the time period over which we can record from these neurons . Therefore , we might soon be in a regime where the number of available samples is comparable to the population size . To make meaningful conclusions from such datasets , our models will need to take maximal advantage of the prior knowledge about the dependency structure among neurons . The prior knowledge that the distribution of activity could be close to critical and that the population could be globally coupled are two macroscopic features of the neural code that future models should be able to reproduce without extreme tuning of many parameters . Our semiparametric energy models directly utilize this prior knowledge , and because the complexity of the nonlinearity is held fixed for all population sizes , it can be easily used in models with arbitrary number of neurons . While today’s neuroscience provides us with sufficient data to build accurate models of neural populations , it is also important that these models generate new hypotheses and shape the direction of future research . For example , our goal was not to trace the origins of the observed Zipf law , but we nevertheless believe that the pursuit of these origins can only happen in a data-driven context to which our models will further contribute . There are many toy models that reproduce Zipf law , several of which have been proposed in the neuroscience context to additionally account for related signatures of criticality , e . g . , the behavior of the heat capacity . Some of these models invoked the particular structure of the observed pairwise correlations , ascribed specific importance to fluctuating ( latent ) variables ( see Discussion in [19] ) which could ( or not ) be directly related to the stimulus itself , or suggested that the processes of model construction , inference , or scaling to large N generate spurious signatures of criticality . The issue is thus not the lack of possible explanations . Rather , it is that these explanations account qualitatively for only one selected aspect of the actual data , while not truly testing whether the proposed explanation is quantitatively consistent with all of the reported phenomena and measured statistics . Here , we took seriously the idea that the signatures of criticality could be due to a global coupling to a hidden ( latent ) fluctuating variable , as proposed and discussed in the context of a blowfly motion-sensitive neuron in Ref [28] , and we have shown that the proposed mechanism is viable in a model that precisely accounts for a real and well-studied dataset [11] . It is important to stress that the identified latent variable is only an effective description of the data , and so , without further experiments , we cannot interpret it in terms of some biophysical mechanisms , nor can we claim , for example , that the population is critical because of this latent variable . However , knowing that this latent variable is a useful statistic describing the population should be a motivation for designing future experiments so that we can correlate it with more detailed mechanisms on the level of neural circuits , and possibly gain insight into its bimodal structure . It also suggests that we should analyze populations responding to various stimuli so that we can understand the latent variable’s stimulus dependence . The scaling of the latent variable shown in Fig 6C also suggests that we should reexamine whether we could find even better description of the data with more than one latent variables . This could be done by studying models with multiple or with multidimensional nonlinearities . Generally , these models have the form log p ( s ) ∝ V ( E1 ( s ) , E2 ( s ) , … ) , and a particularly interesting special case is when each “energy” function Ei is a simple linear projection of the responses as in the semiparametric independent model . These models offer an avenue for both improving the accuracy and reducing the number of parameters . In light of the theoretical analysis in Ref [28] , each dimension of the nonlinearity could possibly be interpreted as a separate latent variable . While we are not aware of general conditions which would guarantee that a multidimensional nonlinearity can be mapped to a multidimensional latent variable , intuition suggests that as the dimension of the nonlinearity increases , the space of nonlinearities which allow for this inversion becomes smaller . This means that if we fit a model with a general multidimensional nonlinearity to data , and we find that this nonlinearity can be mapped to a multidimensional latent variable , then it is an evidence that these latent variables can be correlated with actual physical mechanisms which can be sought for in future experiments . There exist alternative ways of modeling global coupling ( and thus likely capturing signatures of criticality ) in neural populations . Hidden-Markov-Model-type ( HMM ) models have been considered for the retinal data [43] , where the discrete hidden states correspond to collective modes of activity that , due to noise in neural spiking , map probabilistically into observed activity patterns of spiking and silence . In contrast , our model can be interpreted as having a single ( but continuous ) hidden variable—although we empirically find that the distribution of this latent variable is actually bimodal , highlighting the basic distinction between the “silent” or “inactive” state of the retina , and the “active” state [44] . The HMM models were introduced to capture more flexibly collective modes of activity first observed in pairwise and K-pairwise models [10 , 11] . Unlike the semiparametric pairwise model , they take into account the observed temporal dynamics , and they are also parametrically richer . Furthermore , their learned hidden states show interesting correspondence to the displayed stimuli even though the model is a priori agnostic about the stimulus . On the other hand , the HMM models admit no clear link to and interpretations of the signatures of criticality , which was our motivation in this paper . Related to the HMMs , [45 , 46] discuss another classes of accurate models which capture the temporal dynamics of the population . Unlike HMMs and related models , this paper is concerned with modeling the stationary distribution rather than the precise time-dependence of the population . While this discards a lot of information , and hence the resulting models are possibly less accurate , there are advantages to focusing on stationary models . On the technical side , temporal models require more parameters and associated decisions about how to represent the stimulus and its interactions with the population , and so they are harder to scale to datasets with large numbers of neurons . More importantly , however , it was precisely by disregarding the temporal information that the ubiquity of criticality and the role of weak pairwise correlations [8] in neural populations were discovered . It is thus possible that discarding the temporal information allows us to make more general observations about neural codes . This is an important hypothesis . For example , the models we consider in this work , as well as most of published models , are accurate only when applied to data collected in a very narrow experimental context , and it is unclear if/how much would these models generalize to novel stimuli/experimental conditions , nor is it obvious how to design experiments so that we can infer models which generalize as much as possible . While it remains to be tested , it is an intriguing hypothesis that stationary models have more potential for generalization across experiments . In the domain of stationary models , Restricted Boltzmann Machines ( RBMs ) and their derivatives [34] are also classes of energy-based models for population activity that could capture global coupling by latent variables . RBMs are universal learners that , given sufficient data , can reproduce any distribution—including a critical one; like HMM models , however , making a generic link between their parameters and criticality appears difficult . We note that the RBM structure is not incompatible with the structure of semiparametric energy-based models: one could consider a “semiparametric RBM model , ” where E in Eq 3 is defined by a RBM , whose parameters are learned jointly with the nonlinearity , V ( E ) . A different class of models that has been demonstrated to capture criticality consists of various derivatives of the dichotomized Gaussian model [26 , 39 , 47] . A comparison between the dichotomized Gaussian , pairwise , and K-pairwise models on the same dataset as we consider in this work was done in [11] . They showed that while the dichotomized Gaussian is comparable to the pairwise model , the K-pairwise model , and hence also the semiparametric pairwise model , are more accurate . The analysis in [48 , 49] shows that the distribution of the total network activity ( as in Fig 3 ) can often be fitted using a generalization of the dichotomized Gaussian model in which the inputs are q-Gaussians , but they assume that all neurons are the same and do not aim to model more detailed statistics of the neural responses . More recently [50] discusses how to extend the dichotomized q-Gaussian model to heterogeneous populations . However , they only show how to use this model to match the observed means and pairwise correlations while keeping the q parameter fixed , and they do not discuss how to perform maximum likelihood inference on all parameters simultaneously . Since these studies on the dichotomized q-Gaussian model showed that the q parameter is relevant for statistics related to global coupling , it would be an interesting research direction to develop a procedure for maximum likelihood inference of this model , and compare it to the semiparametric pairwise model . The observations of criticality in real data are not specific to neuroscience . Datasets in many other fields such as luminance in natural images [31] , or amino acid sequences of proteins [33] have been shown to exhibit Zipf law . In particular , models of the form Eqs ( 3 ) and ( 17 ) have been used to model the statistics of small image patches under the name elliptically symmetric distributions and Gaussian scale mixtures [51 , 52] although the motivation for using these models had nothing to do with criticality . These models are much easier to analyze than the models we consider in this paper because the variables si are continuous rather than discrete . Our discussion regarding Eq ( 14 ) and the prior expectations about the shape of the nonlinearity is valid even in the continuous case . In particular , elliptically symmetric distributions are essentially the same as our semiparametric pairwise models , Eq ( 9 ) , only with continuous variables . Because si are continuous , we can analytically evaluate the density of states , ρ ( E ; J ) ∝ E N 2 - 1 , ( 20 ) and so the optimal nonlinearity for an elliptically symmetric distribution fitted to a large system which exhibits criticality ( e . g . image patches ) is expected to be V ( E ) = ( N/2 − 1 ) log E + const . Another connection between our models and a substantial body of theoretical work is in the context of nonextensive statistical mechanics . Physicists have considered models of the form Eqs ( 3 ) and ( 17 ) as models of systems whose entropy grows sublinearly with the system size [53] . It is difficult to make these connections explicit because nonextensive statistical mechanics has been studied mostly through toy models rather than data-driven generative models that we examine here; furthermore , in the toy models the latent variables are usually assumed to converge to a delta function as the population size grows which is in stark contrast with our findings in Fig 6 . Nevertheless , deepening the connection between models inferred from data , the maximum entropy formalism itself ( e . g . , considering the possibility that our semiparametric energy models of Eq ( 3 ) can be derived from the maximization of a generalized version of the standard entropy ) , and nonextensive statistical mechanics is an interesting topic for further research .
Let ρ ( E′; α ) = ∑s δE′ , E ( s; α ) count the number of states which map to the same energy E′ . The probability distribution of E ( s; α ) when s is distributed according to Eq ( 3 ) is p ( E ′ ; α , V ) = ∑ s p ( s ; α , V ) δ E ′ , E ( s ; α ) = ρ ( E ′ ; α ) e - V ( E ′ ) Z ( α , V ) . ( 21 ) Given data D = { s ( i ) } i = 1 M , let p ^ ( E ′ ; α ) = 1 M ∑ i = 1 M δ E ′ , E ( s ( i ) ; α ) be the data distribution of the energy , and let Ωα be the image of E ( s; α ) . The average log-likelihood of the data can be rewritten as L ( α , V ) = - log Z ( α , V ) - 1 M ∑ i = 1 M V ( E ( s ( i ) ; α ) ) = - log Z ( α , V ) - ∑ E ′ ∈ Ω α p ^ ( E ′ ; α ) V ( E ′ ) = - ∑ E ′ ∈ Ω α p ^ ( E ′ ; α ) log ρ ( E ′ ; α ) + ∑ E ′ ∈ Ω α p ^ ( E ′ ; α ) log p ( E ′ ; α , V ) , ( 22 ) where the third line follows by substituting the logarithm of Eq ( 21 ) . Eq ( 22 ) has a simple interpretation . The last term , which is the only one depending on V , is the average log-likelihood of the samples { E ( s ( i ) ; α ) } i = 1 M under the model p ( E; α , V ) , and so , for any α , the purpose of the nonlinearity is to reproduce the data probability distribution of the energy . Our restriction that V is a twice differentiable increasing function can be seen as a way of regularizing learning . The last term in Eq ( 22 ) is the negative cross entropy between p ^ ( E ; α ) and p ( E; α , V ) and it is well known that this term is maximal if p ^ ( E ; α ) = p ( E ; α , V ) . According to Eq ( 21 ) , if V was arbitrary , then , for any α , we can satisfy this equality with any ( possibly infinite ) function V such that V ( E ) = log ρ ( E ; α ) - log p ^ ( E ; α ) + const . for all E ∈ Ω α . ( 23 ) If the energy function assigns distinct energies to distinct states , then the choice in Eq ( 23 ) leads to a model which exactly reproduces the empirical distribution of data , and hence overfits . An alternative way of regularizing would be to assume that V is a piecewise constant function . In that case , the analog of Eq ( 23 ) is V ( E ) = log ρ ¯ ( E ; α ) - log p ¯ ^ ( E ; α ) + const . , ( 24 ) where , for every bin on which V is constant , the density of states ρ ¯ ( E ; α ) counts the number of states whose energy maps to this bin divided by the bin width . Similarly , the empirical energy density p ¯ ^ ( E ; α ) counts the number of samples whose energy maps to this bin divided by the bin width . All models were trained using a variation of Persistent Contrastive Divergence [38] which performs an approximate gradient ascent on the log-likelihood for any model of the form p ( s; α ) = exp ( −F ( s; α ) ) /Z ( α ) , where F ( s; α ) is a computationally tractable function differentiable in the parameters α , and Z ( α ) is a normalization constant . Given an initial guess of the parameters α0 , and a list of Ms samples drawn from p ( s; α0 ) , the algorithm can be summarized as for t ≔ 1 to L αt = αt−1 + η ( E[∇α F ( s; αt−1 ) ]samplest−1 − E[∇α F ( s; αt−1 ) ]data ) samplest = GIBBSn ( samplest−1 , αt ) where L is the number of iterations , η is the learning rate , E[⋅]list denotes an average over the list of states , and GIBBSn represents n applications of the Gibbs sampling transition operator . Pairwise and K-pairwise models were trained using η = 1 , n = 2N , and with initial parameters drawn from a normal distribution with 0 mean and 0 . 1 standard deviation . We iterated the algorithm two times , first with L = 10000 , Ms = 3 × 104 , then with L = 10000 , Ms = 3 × 105 . Semiparametric independent and pairwise models were trained using η = 5 × 10−5 for the parameters of the function V ( see below ) , and η = 1 for all other parameters . We initialized the model with parameters corresponding to the learned independent ( pairwise ) models , and trained for L = 10000 iterations with Ms = 3 × 104 samples . The function V is parametrized through a function W ( see Eq ( 7 ) ) . We use piecewise constant functions to parametrize W . Let [E0 , E1] be an interval containing the range of energies E ( s; α ) which we expect to encounter during learning . We divide the interval [E0 , E1] into Q non-overlapping bins of the same width with indicator functions Ii , i . e . Ii ( E ) = 1 if E is in the ith bin , otherwise Ii ( E ) = 0 , and we set W ( E ) ≡ W ( E ; β ) = ∑ i = 1 Q β i I i ( E ) . We used Q = 20 bins in all experiments . This was a conservative choice: increasing Q did not result in a higher training or validation likelihood . The integrals in Eq ( 7 ) can be carried out analytically for this choice of W yielding an exact expression for V as a function of γ and β . For E < E0 , we have V ( E; γ , β ) = γ1 + γ2 ( E − E0 ) . For E > E0 we have V ( E; γ , β ) = γ1 + γ2 f ( E; β ) , where f ( E ; β ) = ∫ E 0 E exp ( ∫ E 0 E ′ W ( E ′ ′ ; β ) d E ′ ′ ) d E ′ = ∑ i = 1 [ E ] - 1 exp ( Δ ∑ j = 1 i - 1 β j ) exp ( Δ β i ) - 1 β i + exp ( Δ ∑ j = 1 [ E ] - 1 β j ) exp ( β [ E ] ( E - ( [ E ] - 1 ) Δ ) ) - 1 β [ E ] . ( 25 ) We define [E] as the number of the bin that contains E . If E > E1 , then we define [E] = Q + 1 , and βQ+1 = 0 . Using this expression we can calculate the gradients ∇α F ( s; α ) in the algorithm exactly . This calculation is straightforward although the resulting expressions are cumbersome . For the semiparametric pairwise model , we have F ( s ; γ , β , J ) = V ( ∑ i , j = 1 N J i j s i s j ; γ , β ) . The gradient with respect to the couplings is ∂ F ( s ; γ , β , J ) ∂ J k l = V ′ ( ∑ i , j = 1 N J i j s i s j ; γ , β ) s k s l . ( 26 ) The gradients with respect to γ and β are just the gradients of V ( E; γ , β ) with respect to these parameters and they are as follows: ∂ V ( E ; γ , β ) ∂ γ 1 = 1 , ( 27 ) ∂ V ( E ; γ , β ) ∂ γ 2 = f ( E ; β ) , ( 28 ) ∂ V ( E ; γ , β ) ∂ β k = γ 2 f ( E ; β ) ∂ β k . ( 29 ) If k > [E] , then ∂ f ( E ; β ) ∂ β k = 0 . ( 30 ) If k = [E] , then ∂ f ( E ; β ) ∂ β k = exp ( Δ ∑ j = 1 [ E ] - 1 β j ) exp ( Δ β [ E ] ) Δ β [ E ] - exp ( Δ β [ E ] ) + 1 β [ E ] 2 . ( 31 ) If k < [E] , then ∂ f ( E ; β ) ∂ β k = exp ( Δ ∑ j = 1 k - 1 β j ) exp ( Δ β k ) Δ β k - exp ( Δ β k ) + 1 β k 2 + Δ ∑ i = k + 1 [ E ] - 1 exp ( Δ ∑ j = 1 i - 1 β j ) exp ( Δ β i ) - 1 β i + Δ exp ( Δ ∑ j = 1 [ E ] - 1 β j ) exp ( β [ E ] ) ( E - ( [ E ] - 1 ) Δ ) - 1 β [ E ] . ( 32 ) Data likelihoods cannot be evaluated exactly because the normalization constants Z are intractable . We resorted to Monte Carlo method known as thermodynamic integration in physics [54] , and annealed importance sampling in machine learning , to estimate the normalization constants [55] . The initial model for annealed importance sampling was always the independent model for which the partition function can be calculated exactly . The sampling procedure consisted of 104 intermediate distributions which uniformly interpolated from the independent model to the model of interest . Each partition function was estimated using 104 samples . All reported likelihoods were evaluated on held-out data . A simple cross-validation also showed that our models did not suffer from overfitting . Density of states was estimated using the Wang and Landau algorithm [11 , 40] . The accuracy parameter ( the smallest increment size for the log of the density of states ) was 10−7 . The energy range was estimated during the first few thousand steps of the algorithm . This range was divided into ∼ 104 bins . We decreased the increment size every ∼ 108 iterations instead of checking energy histogram flatness since the later is hard to do when some energy bins are inaccessible . We inferred the probability densities of the latent variables by considering the model in Eq ( 17 ) with fixed J which corresponds to the coupling matrix of the previously learned semiparametric pairwise model . The domain of the latent variable was set to [0 , 5] . We approximated the integral with a sum by dividing this domain into 400 bins , and the value of the probability density q ( h ) was inferred by maximizing the likelihood of data subject to the constraint that q ( h ) integrates to 1 . To make the computation tractable , we needed an expression for Z ( h; J ) . This can be obtained from the estimated density of states ρ ( E; J ) of the energy as Z ( h ; J ) = ∑ s e - h E ( s ; J ) = ∫ ρ ( E ; J ) e - h E d E . ( 33 )
|
Populations of sensory neurons represent information about the outside environment in a collective fashion . A salient property of this distributed neural code is criticality . Yet most models used to date to analyze recordings from large neural populations do not take this observation explicitly into account . Here we aim to bridge this gap by designing probabilistic models whose structure reflects the expectation that the population is close to critical . We show that such principled approach improves previously considered models , and we demonstrate a connection between our models and the presence of continuous latent variables which is a recently proposed mechanism underlying criticality in many natural systems .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"linguistics",
"social",
"sciences",
"random",
"variables",
"neuroscience",
"covariance",
"probability",
"distribution",
"mathematics",
"statistics",
"(mathematics)",
"thermodynamics",
"entropy",
"animal",
"cells",
"probability",
"density",
"probability",
"theory",
"physics",
"statistical",
"models",
"cellular",
"neuroscience",
"cell",
"biology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"computational",
"linguistics"
] |
2017
|
Probabilistic models for neural populations that naturally capture global coupling and criticality
|
Extracellular trypanosomes can cause a wide range of diseases and pathological complications in a broad range of mammalian hosts . One common feature of trypanosomosis is the occurrence of anemia , caused by an imbalance between erythropoiesis and red blood cell clearance of aging erythrocytes . In murine models for T . brucei trypanosomosis , anemia is marked by a very sudden non-hemolytic loss of RBCs during the first-peak parasitemia control , followed by a short recovery phase and the subsequent gradual occurrence of an ever-increasing level of anemia . Using a newly developed quantitative pHrodo based in vitro erythrophagocytosis assay , combined with FACS-based ex vivo and in vivo results , we show that activated liver monocytic cells and neutrophils as well as activated splenic macrophages are the main cells involved in the occurrence of the early-stage acute anemia . In addition , we show that trypanosomosis itself leads to a rapid alteration of RBC membrane stability , priming the cells for accelerated phagocytosis .
Extracellular trypanosomes including Trypanosoma brucei , T . evansi , T . congolense and T . vivax , are parasites that affect a very broad host range , and combined , threaten human and animal health throughout various continents . Despite the incredibly wide range of trypanosomosis-associated diseases and pathological complications , one common feature of trypanosomosis is the occurrence of anemia , which is seen in human infections [1] as well as non-human primate trypanosomosis [2 , 3] , and most other wildlife and livestock trypanosome infections [4] . Anemia is a condition in which an imbalance occurs between erythropoiesis and red blood cell clearance . RBC are either destined for clearance ( e . g . by senescence , antibody coating or damage ) or are cleared as “innocent bystanders” ( e . g . during hemorrhage ) [5] . A loss of membrane phospholipid asymmetry has been recognized as a key trigger that can lead to recognition and extravascular removal of senescent RBCs , disordered RBCs , and transfused RBC that have been stored for a long time , by cells of the myeloid phagocyte system [6–8] . Hence , infection-associated complications that affect either of these processes will lead to anemia . Trypanosomosis is suggested to 1 ) hamper erythropoiesis , 2 ) enhance erythrophagocytosis ( also termed extravascular hemolysis ) and 3 ) in some cases , eg T . vivax infections , cause intravascular hemolysis . In murine models for T . brucei trypanosomosis , anemia is marked by a very sudden non-hemolytic loss of RBCs during the first-peak parasitemia control , followed by a short recovery phase and the subsequent gradual occurrence of an ever increasing level of anemia , reminiscent of ‘anemia of chronic infection’ [9–13] . Interestingly , as anemia occurs in B-cell deficient μMT mice with similar kinetics as WT mice , the process involved appears antibody independent [14 , 15] . This contrasts a previous in-vitro based hypothesis that cross-reactive anti-VSG antibodies might contribute to a complement-mediated hemolysis event [16] . Based on combined recent data , the most plausible explanation for the initiation of trypanosomosis-associated anemia is the occurrence of enhanced RBC phagocytosis , resulting from a pro-inflammatory cytokine storm occurring during the early stage of infection , leading to macrophage hyper-activation and enhanced erythrophagocytosis [12 , 13 , 17–20] . However , till now two main obstacles have hampered the in depth assessment of this hypothesis as ( i ) previous methods for RBC phagocytosis have difficulties differentiating between actual RBC uptake and RBC adherence to phagocytozing cells , and ( ii ) quantification of phagocytozed RBC numbers with simultaneous characterization of in vivo RBC phagocytozing cells has been virtually impossible . In order to address these issues , we now used a newly developed pHrodo based in vitro erythrophagocytosis assay , as well as an ex vivo FACS based analysis using the same substrate . Unique in this approach is that the visualization of RBC labeling is pH dependent and only becomes traceable in the acidic environment of the lysosome of phagocytozing cells . Hence , we were able to show ex vivo that activated liver monocytes , monocyte-derived macrophages as well as neutrophils are the main cells contributing to trypanosomosis-associated acute stage erythrophagocytosis . In addition , we show that trypanosomosis itself leads to a rapid alteration of RBC membrane stability , priming the cells for accelerated phagocytosis .
7–8 week old female C57Bl/6 mice purchased from Janvier as well as ubiquitin-GFP ( Jackson Laboratories ) mice bred in-house were housed at the animal facility of the Vrije Universiteit Brussel . All experiments complied with the ECPVA guidelines and were approved by the ETHICAL COMMITTEE for ANIMAL EXPERIMENTS ( ECAE ) at the Vrije Universiteit Brussel ( protocol #14–220–23 and #12–220–2 ) . Mice were infected by intraperitoneal ( i . p . ) injection of 5000 pleomorphic Trypanosoma brucei AnTat1 . 1E parasites , which were a kind gift from N . Van Meirvenne ( Institute for Tropical Medicine , Belgium ) . RBC counts were determined via a hematocytometer at two-day intervals on 2 , 5μl blood sample collected from the tail vein of non-infected and infected animals . Anemia was expressed as the percentage of reduction in RBC counts compared to non-infected animals . Peritoneal exudate cells ( PECs ) , spleen and liver were harvested from CO2 euthanized non-infected and day 6 infected mice . Livers were minced in 10 ml digestive media ( 0 . 05% collagenase type A in Hanks’ Balanced Salt Solution ( HBSS ) without calcium or magnesium; Invitrogen ) and incubation at 37°C for 30 minutes , the digested tissue was homogenized and filtered ( 40 μm pore filter ) . Spleen cells were obtained by homogenizing the organs in 10 ml RPMI medium containing 5% foetal calf serum ( FCS ) with or without collagenase type A , Next , the liver and spleen cell suspension was centrifuged ( 7 minutes , 300×g , 4°C ) and the pellet treated with RBC lysis buffer ( 0 . 15 M NH4Cl , 1 . 0 mM KHCO3 , 0 . 1 mM Na2-EDTA ) . Subsequently , the cells were resuspended in ME—medium ( RPMI medium , 5% FCS , 1% L-glutamine and non essential amino acids , 1% Penicillin-Streptomycin and β-mercaptoethanol ) and 4 105 cells were put in co-culture with or without 2 107 labeled or unlabeled RBCs in polypropylene tubes ( BD Biosciences ) . Co-cultures were incubated overnight at 37°C and 5% CO2 with or without lipopolysaccharide ( LPS ) stimulation ( 1μg/ml ) . An overview of the isolation protocol and erythrophagocytosis assay is given in Fig . 1 . 7–8 week old female C57Bl/6 non-infected or T . brucei infected ( day 6 p . i . ) mice were injected intravenously ( i . v . ) with 109 pHrodo labeled or unlabeled RBCs ( from either C57Bl/6 or Ubiquitin-GFP mice ) in 200 μl PBS . After 18 hours , mice were CO2 euthanized and spleen and livers were isolated and processed into single cell suspension as described above . Next , the cells were analyzed via flow cytometry as described further . Blood was harvested from CO2 euthanized mice by cardiac puncture using 50μl 1000 U/ml heparin . RBCs were counted and 109 RBCs were washed twice with 15 ml PBS , 2000 rpm , 7 minutes . Next , RBCs were labeled with 2 μl pHrodo Red succinimidyl ester ( pHrodo Red , Life Technologies ) in a final volume of 1 ml PBS for 60 minutes at 37°C . Subsequently , labeled RBCs were incubated for 15 minutes with 10 ml RPMI/5% FCS at 37°C and washed twice with the same medium , 2000 rpm , 7 minutes . Labeled RBCs were resuspended in RPMI/5% FCS at a final concentration of 109 RBCs/ml and put in co-culture with isolated leukocytes ( 20μl blood/condition ) . As negative control , RBCs were treated in the same manner without addition of the pHrodo dye . After overnight co-culturing ( see above ) cells were subjected to flow-cytometrical analysis . Briefly , the cells were washed with FACS medium ( 5% FCS in RPMI ) and non-specific binding sites were blocked by incubating 20 minutes at 4°C with an Fc-blocking antibody ( anti-CD16/32 , clone 2 . 4G2 ) . Next , cell suspensions were stained with fluorescent conjugated antibodies for 30 minutes at 4°C . Fluorescent antibodies: CD11b PE-Cy7 clone M1/70 , F4/80 FITC clone C1:A3-A , Ly6C APC clone AL-21 , Ly6G PerCP-Cy5 . 5 clone 1A8 , CD45 APC-Cy7 clone 30-F11 ( BD Biosciences ) , CD64 PE clone X54–5/7 . 1 . ( BioLegend ) , CCR2 PE clone 475301 , MerTK PE clone 108928 ( R&D systems ) , Ly6B clone 7/4 ( AbD Serotec ) . Following washing with FACS buffer they were analyzed on a FACS Canto II flow cytometer ( BD Biosciences ) and data was processed using FlowJo software ( Tree Star Inc . ) . The osmotic fragility of erythrocytes was determined using the adjusted protocol from Meurs et al . [21] . Solutions with decreasing concentrations of NaCl were prepared by mixing distilled H2O and HBSS solutions . These solutions change in isotonicity as the NaCl concentration decreases , resulting in hemolysis of a fraction of the erythrocytes and the red coloration of the solution ( hemoglobin content ) can be measured spectrophotometrically . A gradient of solutions was prepared and total hemolysis was determined by exposure to 100% distilled H2O , 0% hemolysis was determined by exposure to 100% HBSS-solution . 3μl of blood was added to 300μl of each solution , using flat bottom 96 well plates ( BD Biosciences ) . To determine if the labeling had an effect on RBC fragility , also pHrodo labeled RBCs and unlabeled RBCs analyzed in the same manner . After mixing by gently pipetting up and down , the solutions were incubated at room temperature for 20 minutes . Subsequently , erythrocytes ( remnants ) were pelleted by centrifugation at 2000 rpm for 10 minutes and 150μl supernatant was transferred to a new 96 well plate . The absorbance of the solutions was determined at 550 nm . The percentage of hemolysis was plotted against the concentration of NaCl in the medium and the NaCl concentrations corresponding with 50% hemolysis were determined . OD’s were normalized by using the smallest value of the data sets as 0% ( this is the OD measured in PBS ) and 100% is defined as the largest value of the data sets ( OD measured in H2O ) . The fatty acid composition of RBCs was analyzed by gas chromatography . RBCs from non-infected and T . brucei infected mice were stained for Ter119 PE ( clone TER-119 ) and CD71 APC ( clone R17217 ) ( BD biosciences ) . Next , 107 Ter119+ CD71- RBCs were sorted ( FACS AriaTM ) and lyophilized ( Flexi-dry μP Microprocessor control , FTSTM systems ) according to the manufacturers protocol . Subsequently , fatty acids were extracted and methylated using a method described elsewhere [22] and analyzed by gas chromatography . Statistical analysis was performed using Student-test and GraphPad Prism software ( GraphPad 6 , San Diego , CA ) . Values are expressed as mean ± standard error mean ( SEM ) . Values of p≤ 0 . 05 are considered to be statistically significant .
Multiple assays for erythrophagocytosis have been developed in the past , each set-up having its own drawbacks such as the need for radioactive reagents . Here we use the pHrodo dye for the labeling of RBCs . The dye reacts with the primary amines on the RBC to yield a covalently linked pH probe , which increases in fluorescence as the pH of the surroundings becomes more acidic . Due to the low pH of the phagolysosome , phagocytozed RBCs can be visualized without being mistaken for RBCs merely sticking to the outside of the phagocytozing cell . Hence this technique enables the straightforward quantification of erythrophagocytosis using FACS and , if required , confirmation of the obtained result by fluorescence microscopy . As shown in Fig . 1 , we used this technique to monitor erythrophagocytosis in different cellular contexts , always using the same experimental layout . To validate this erythrophagocytosis assay , pHrodo labeled RBCs and PECs from non-infected mice were put in co-culture overnight with or without LPS stimulation and subsequently stained and analyzed via flow cytometry to distinguish phagocytozing cells . Following gating on peritoneal macrophages , i . e . CD11bhi F4/80hi cells ( Fig . 2A , left panel ) , a shift in the fluorescence signal in co-cultures of PECs with pHrodo labeled RBCs occurs ( Fig . 2A , right panel ) indicating erythrophagocytosis by these cells . The shift in fluorescent signal expressed as delta median fluorescence intensity ( ΔMFI ) clearly shows that the peritoneal macrophages are the only cells involved in RBC uptake ( Fig . 2B ) . Treatment of PECs with LPS enhances the erythrophagocytozing ability of these cells . Uptake of labeled RBCs by PECs was confirmed by fluorescence microscopy ( Fig . 2C ) . Both liver and spleen are important organs for maintenance of RBC homeostasis and degradation of senescent RBCs under homeostatic conditions [23–25] . Hence , whole liver and spleen cultures were analyzed in the pHrodo erythrophagocytosis set-up to determine which cells are phagocytozing labeled RBCs . In the liver , CD11b+ F4/80+ myeloid cells ( i . e . resident macrophages or Kupffer cells ) were the most efficient erythrophagocytozing cells under steady state conditions ( Fig . 3A , gating strategy described in S1 Fig A-D ) . In the spleen , CD11b+ F4/80+ myeloid cells ( i . e . metallophilic and marginal zone macrophages ) as well as CD11b+ Ly6C+ Ly6G- monocytes and CD11b+ Ly6cint Ly6G+ polymorphonuclear ( PMN , granulocytes/neutrophils ) cells were able to phagocytose RBCs under steady state conditions ( Fig . 3B , gating strategy described in S2 Fig A-D ) . Of note , the Rest fraction of both spleen and liver , which consisted mainly of B- and T-cells and patrolling monocytes and NK cells respectively , did not exhibit significant erythrophagocytosis . In order to evaluate the process of erythrophagocytosis under inflammatory conditions , cells were stimulated with LPS in vitro . We observed that LPS stimulation of liver cells resulted in a 5 . 5-fold increase in RBC phagocytosis by the neutrophils , a 1 . 8 fold increase by the CD11b+ F4/80+ myeloid cells fraction and a 2 . 7 fold increase by the monocyte fraction ( Fig . 3A ) . LPS stimulation of spleen cells resulted in a similar enhanced RBC phagocytosis by the neutrophils ( 3 . 7 fold increase ) , CD11b+ F4/80+ myeloid cells ( 1 . 8 fold increase ) and monocytes ( Fig . 3B ) ( 3 fold increase ) . During the course of trypanosome infection anemia develops , which can be divided into different stages . During the acute stage of T . b . brucei infection , C57Bl/6 mice develop acute anemia , which is typically observed between day 5–8 post infection ( Fig . 4A ) . After a slight recovery phase ( i . e . between day 8–10 ) , the reduction in the RBC percentage persists throughout the chronic phase of infection . Here , using the pHrodo in vitro erythrophagocytosis assay we determined whether enhanced erythrophagocytosis could be responsible for this severe reduction in RBCs observed during the early stage of infection . Whole liver and spleen cells of day 6 infected mice were put in co-culture with labeled RBCs of the corresponding mice and the erythrophagocytozing potential was analyzed and compared to that of non-infected animals . Trypanosome infection causes an alteration of the liver myeloid cell composition , therefore a more elaborate gating strategy ( S3 Fig ) allows to distinguish monocyte-derived macrophages ( CD11b+ Ly6C+ MHC-II+ ) and resident macrophages ( CD11b+ F4/80+ Ly6C- MHC-II+ ) in addition to neutrophils and monocytes ( CD11b+ Ly6C+ MHC-II- ) . In the liver , neutrophils and monocytes show a remarkable increase in their phagocytozing potential ( Fig . 4B & S4 Fig ) during infection while the phagocytozing potential of the monocyte-derived macrophages was unaltered . In contrast , the phagocytozing potential of resident macrophages diminished and the Rest fraction ( consisting of NK cells and patrolling monocytes ) only played a minor role in erythrophagocytosis during infection . In contrast to the situation in the liver , spleen neutrophil-mediated erythrophagocytosis was not enhanced upon infection ( Fig . 4C ) . Monocytes on the other hand exhibited increased erythrophagocytosis and the CD11b+ F4/80+ myeloid cell-mediated erythrophagocytosis was reduced during infection ( Fig . 4C ) . When investigating the effect of T . brucei infection on erythrophagocytosis in vitro it appears that within the liver and spleen , neutrophils and monocytes are responsible for enhanced RBC clearance . The RBC clearance capacity of the liver Kupffer cells and spleen CD11b+ F4/80+ myeloid cells decreased at day 6 of infection . Besides anemia development , trypanosomosis infection has been shown to have a vast effect on spleen and liver immune cell populations [26–28] . Indeed , as mentioned earlier we observed that during the early stage of infection there was an increase ( in percentage as well as absolute numbers ) of monocytic cells and neutrophils in the liver and a decrease in resident macrophages ( Fig . 5A ) . In the spleen , a similar increase in monocytes and neutrophils is observed as well as an increase in CD11b+ F4/80+ myeloid cells ( Fig . 5B ) . Therefore , at the level of the spleen it seems that besides monocytes also the CD11b+ F4/80+ myeloid cells ( i . e . red pulp macrophages ) can significantly contribute to erythrophagocytosis during infection . Hence both organs seem to contribute to erythrophagocytosis during the acute stage of infection . It is generally known that the amount of senescent RBCs increases during infection , resulting in an enhanced erythrophagocytosis . Hereby , the RBC membrane displays enormous plasticity and deformability to cope with changes in pressure and shear stress in the microcirculation [21] . 40% of the erythrocyte membrane is composed of lipids in the form of phospholipids , glycolipids and un-esterified cholesterol [23] . These lipids exchange easily with plasma lipoproteins by a continuous exchange mechanism [29] . Hence , serum cholesterol levels or , e . g . insertion of pathogen lipids during infection , influence the lipid composition of the erythrocyte membrane , altering its physical properties and hereby affecting RBC survival [21] . Therefore , an altered RBC membrane composition could also play a role in the induction of acute anemia during murine trypanosome infection . To assess this , a quantitative fatty acid determination was performed on FACS sorted RBCs from both non-infected and trypanosome infected animals ( S5 Fig FACS gating ) . Fig . 6A shows that RBCs from infected animal ( i . e . iRBC ) have increased C16:0 ( palmitate ) fatty acids and reduced C18:0 ( stearate ) , C18:1 ( oleate ) , C18:2 ( linoleate ) and C22:1 ( erucic acid ) fatty acids compared to RBCs from non-infected animals , indicating that trypanosome infection indeed alters the RBC membrane composition . Next , the effect of trypanosomosis on RBC membrane rigidity was checked , using resistance to osmolarity changes as a readout . As indicated in Fig . 6B , a functional change in the RBC membranes of infected animals occurred , as the cells became significantly more susceptible to lysis . The assay was performed on day 6 post infection , concomitantly with the acute drop in RBC percentages . Of note , by comparing the RBC fragility of pHrodo labeled RBCs with unlabeled RBCs undergoing the same processed we could demonstrate that the pHrodo labeling had no effect on the RBC membrane rigidity neither for RBCs from non-infected or infected mice ( S6 Fig ) . Finally , the pHrodo in vitro erythrophagocytosis assay was used to assess a possible link between RBC membrane changes and susceptibility to accelerated erythrophagocytosis . Here , PECs from non-infected mice were co-cultured with pHrodo labeled RBCs from non-infected and trypanosome infected mice . PECs from non-infected mice showed an increased phagocytozing potential when co-cultured with RBCs from infected animals ( ΔMFI: 4936 ± 235 ) compared to co-cultures with RBCs from non-infected animals ( ΔMFI: 2970 ± 105 ) , implying an altered RBC state during acute African Trypanosome infection ( Fig . 6C , left panel ) . In addition , it cannot be excluded that differences in the myeloid cell activation state occurring during infection can play a role in the enhanced RBC uptake . This suggestion was confirmed by the observation that PECs from infected mice showed an increased phagocytozing potential when co-cultured with RBCs from infected animals ( ΔMFI: 6536 ± 1114 ) compared to co-cultures with RBCs from non-infected animals ( ΔMFI: 3537 ± 264 . 1 ) . In addition , the phagocytozing potential of PECs from infected mice was found to be higher than that of non-infected animals ( Fig . 6C , right panel ) . Therefore , not only an altered RBC membrane state but also a difference in myeloid cell activation during acute African Trypanosome infection seems to contribute to enhance RBC uptake . In order to validate the biological significance of the in vitro/ ex vivo observations we setup an in vivo erythrophagocytosis assay ( Fig . 7A ) , whereby pHrodo labeled RBCs from non-infected and infected mice were injected into non-infected or T . brucei infected ( day 6 p . i . ) mice respectively . Subsequently , 18 hours later the uptake of RBCs was tested by flow cytometry . To compare this new in vivo set up to a known technical approach , we performed a side-by-side experiment using RBC of Ubiquitin-GFP mice ( i . e . GFP+RBC ) . When comparing these two systems we could establish that although both systems show Ter119+ signals when gating on liver resident macrophages of non-infected mice , only the pHrodo labeling system due to its pH-sensitivity allows to determine that indeed erythrophagocytosis is occurring ( S7 Fig ) . As shown in Fig . 7B , in non-infected mice the resident macrophages were the main cells involved in RBC uptake . Yet , during the acute stage of infection , monocytes and monocyte-derived macrophages ( collectively termed monocytic cells ) as well as neutrophils exhibit enhanced erythrophagocytosis while the resident macrophages have a reduced erythrophagocytozing capacity . At the level of the spleen no significant enhanced RBC uptake was observed at the level of the monocytes and neutrophils , while there was a significant increase in erythrophagocytozing capacity of the CD11b+ F4/80+ myeloid cells ( Fig . 7C ) . Collectively , these data indicate that liver monocytic cells and neutrophils as well as spleen CD11b+ F4/80+ myeloid cells exhibit an enhanced RBC uptake at this acute stage of infection and hence seem responsible for the induction of acute anemia .
Infection-associated anemia is considered one of the most important pathological features of trypanosomosis and constitutes a major cause of death in bovine African trypanosomosis [30] . Anemia is also a prominent pathological feature of murine trypanosome infection , offering a good model to identify the mechanisms that mediate this phenomenon . T . b . brucei infection of C57Bl/6 mice elicits a severe reduction of RBCs between day 5 and day 8 post infection . Given the acuteness of this phenomenon , a consumptive process seems to be implicated , in particular , as no hemolysis appears to occur . Previously , we have reported a strong increase of cell surface receptors involved in uptake of RBCs and iron-containing compounds in liver tissue . Therefore , liver-associated erythrophagocytosis mediated by cytokine-activated macrophages ( M1 cells ) is a likely contributor to the aggressive anemia during the acute phase of infection [12] . Multiple experimental set-ups have been used to study erythrophagocytosis , among which using sheep-opsonized RBCs [31] labeling of RBCs with 51Cr [32] or cell labeling with fluorescent cell tracking dyes such as PKH26 and carboxy-fluorescein succinimidyl ester ( CFSE ) [33–35] . With respect to the first assay , the disadvantages are the handling of radioisotopes and the fact that fresh sheep RBCs ( foreign RBCs ) need to be opsonized with IgG prior to incubation with phagocytes and phagocytozed RBCs need to be counted or measured in an ELISA reader . In addition , sheep RBCs need to be bought and the transport of the material might affect RBC-membrane stability [36] . Furthermore , this assay allows determining Fc-mediated RBC uptake and does not allow drawing any conclusions with respect to non-Fc mediated RBC uptake mediated via for instance CD36 ( recognizing phosphatidylserine exposure on aged/damaged RBCs ( SIRP1alpha/CD47 ) [37] . The downside to the use of the lipophilic membrane dye PKH26 , which intercalates itself into the cell lipid bilayer , is that it could incorporate itself in other cells through a trogocytosis-like cell-cell membrane transfer mechanism , giving rise to specific binding and therefore masking ‘real’ phagocytosis [38 , 39] . Dyes such as CFSE which form random covalent bonds with amino groups on cellular proteins have an advantage regarding this drawback , but here it is impossible to distinguish real endocytosed RBCs from RBCs sticking to the outside of the macrophage membrane . Using instead RBCs labeled with the cell permeable dye Calcein AM [40] or RBCs from Green fluorescent protein ( GFP ) +/+ mice entail the same disadvantage regarding the distinguishing of surface bound or phagocytozed RBCs . In each of the above-mentioned RBC uptake setups there is an overestimation of the real erythrophagocytosis and therefore the RBC-specific Ter119 antibody needs to be included to allow distinguishing between surface-bound or phagocytozed RBCs . Other phagocytosis assays using labeled dextran beads , do not represent the real in vivo situation and only reveal the phagocytozing potential without giving specific information about erythrophagocytosis . In the past , we designed an in vitro adherent liver cell co-culture assay whereby a monolayer of liver cells was plated in vitro and incubated with a monolayer of RBCs . Clearance plaques were then visualized and counted and served as an indication of erythrophagocytosis [13] . The downside of this assay is that it is impossible to discriminate between erythrophagocytosis and RBC lysis . We now used a different setup to study erythrophagocytosis whereby we labeled RBCs with pHrodo , an amine-reactive ester dye that forms covalent bonds with proteins on the RBCs . This dye has been used successfully before for the labeling of phagocytozed apoptotic lymphocytes [41] and platelets [42] . pHrodo labeling offers the advantage that it is only fluorescent at acidic conditions , hence it will be fluorescent in the phagolysosome ( pH 4 . 5–5 ) but not in the cytoplasm ( pH 7–7 . 4 ) or on the outside of the cell . We confirmed , using a comparative study of GFP+RBCs from ubiquitin-GFP mice and pHrodo-labeled RBCs , that the latter technique allows a more specific determination of erythrophagocytosis ( see S7 Fig ) . In addition , we showed that the pHrodo labeling procedure did not affect the RBC membrane rigidity of either non-infected or infected animals . Using a co-culturing system of pHrodo labeled RBCs from non-infected mice with PECs , spleen or liver cells we could observe via flow cytometry and immunofluorescence microscopy that indeed RBCs are taken up by myeloid cells . Furthermore , the phagocytozing potential can be increased following LPS-stimulation . To validate the assay in a biological relevant situation the pHrodo labeled RBCs were ex vivo put in co-culture with whole spleen or liver cell suspensions from non-infected or infected ( day 6 p . i . ) animals . We observed that in the liver resident macrophages or Kupffer cells seem to be the most prominent erythrophagocytozing cells during steady state situations . However , during the acute phase of infection the erythrophagocytozing capacity of these cells decreased . Yet , an alternative explanation might rely in the fact that these myeloid cells have already taken up a lot of RBCs in vivo and hence cannot take up much more RBCs in vitro . In addition , also in absolute numbers the amount of Kupffer cells decreased , which might be due to massive apoptosis [43] . In contrast , the erythrophagocytozing capacity of monocytes and neutrophils was significantly enhanced during infection . Moreover , given that at this stage of infection there is a massive influx of monocytic cells and neutrophils in the liver , these cells can significantly contribute to anemia development . Of note , the influx of monocytic cells into the liver during the early stages of infection has been reported before and was shown to contribute to parasite control via release of pro-inflammatory molecules such as TNF on one hand and pathogenicity development ( liver damage ) on the other [27 , 44] . It has also been shown that the myeloperoxidase activity ( MPO ) increased drastically during the acute stage of infection [45] which infers that indeed neutrophils become more activated already at the early stages of infection . In the spleen , all myeloid cells were able to phagocytose RBCs during steady state situations . Yet , during infection the erythrophagocytosis ability seems to be unaltered for splenic neutrophils , decreased for CD11b+ F4/80+ myeloid cells ( i . e . red pulp macrophages ) and increased for monocytes . However , when taking into account the total number of cells , both organs can significantly contribute to phagocytosis of RBCs . In a second validation set-up , an in vivo pHrodo based erythrophagocytosis assay was established and shown to corroborate largely the in vitro/ex vivo obtained results . Yet , the phagocytic capacity of liver monocyte-derived macrophages as well as splenic CD11b+ F4/80+ myeloid cells was underestimated in the ex vivo approach . A possible explanation for this discrepancy could be that in the ex vivo assay conditions ( i . e . 18 hours of incubation ) the cells from infected animals are more susceptible to apoptosis , while in the in vivo setup cells are measured directly after isolation . Taken together , it seems that activation of both liver and spleen neutrophils and monocytic cells , as well as splenic CD11b+F4/80+ myeloid cells leads to enhanced erythrophagocytosis and hence can explain the occurrence of severe acute-stage non-hemolytic anemia observed in T . brucei trypanosomosis . Besides differences in erythrophagocytozing potential occurring during infection at the level of the myeloid cells , we also observed that RBCs from infected animals are phagocytozed more efficiently than non-infected RBCs , suggesting an alteration of the RBC membrane during trypanosome infection . Upon analysis of the RBCs , we observed an enhanced osmolytic fragility and an altered fatty acid membrane composition in RBCs from day 6 infected mice compared to non-infected WT mice . Hence , it could be that the modification of RBC properties contributes to enhanced RBC uptake by phagocytozing cells . Interestingly , in the T . evansi model it has been shown that lipid peroxidation causes membrane injury and osmotic fragility resulting in RBC destruction [46] . In T . congolense infection on the other hand , anemia has been correlated to RBC de-galactosylation [47] . In the T . vivax model on the other hand it has been shown that parasite-derived trans-sialidases released during the acute phase of infection triggers erythrophagocytosis by desialylating the major surface erythrocytes sialoglycoproteins , the glycophorins [40] . Besides differences in RBC membrane composition , we also observed that the activation state of myeloid cells plays also an important role in the observed enhanced RBC uptake . Hence , both RBC membrane deformability and enhanced myeloid cells activation can contribute to enhanced RBC uptake . In conclusion , the pHrodo labeling method for RBCs used herein provides a sensitive , reproducible and accurate method for the determination of both ex vivo and in vivo erythrophagocytosis . Results obtained indicate that infiltrating monocytic cells and neutrophils are undergoing infection-associated hyper-activation , resulting in increased erythrophagocytosis of membrane-modified RBCs . This coincides with the appearance of a type 1 cytokine storm that has been described to hallmark the first wave of trypanosomosis control . Given that ( i ) the percentage of resident macrophages within the CD45 hematopoietic compartment of the liver decreases drastically and coincides with reduced erythrophagocytozing potential , that ( ii ) there is a massive influx of monocytic cells and neutrophils into the liver and spleen that exhibit an enhanced erythrophagocytozing activity , and ( iii ) that splenic CD11b+F4/80+ myeloid cells exhibit enhanced erythrophagocytosis we propose that liver and spleen-associated monocytic cells and neutrophils as well as splenic CD11b+ F4/80+ myeloid cell activation plays a previously underestimated role in the acute phase of anemia development .
|
Extracellular trypanosomes , causative agents of sleeping sickness and Nagana , threaten human and animal health throughout the world . Anemia is a hallmark feature of virtually every type of trypanosome infection . During the early phase of experimental murine trypanosomosis , acute anemia occurs as witnessed by a 50% reduction in red blood cells within a 48 hour time span . The acute nature of this phenomenon suggests the implication of a consumptive process such as erythrophagocytosis . However , due to the multiple significant drawbacks of the presently used phagocytosis techniques , this has never been straightforwardly demonstrated . Here we developed a new erythrophagocytosis assay based on the labeling of red blood cells with the acid-sensitive dye pHrodo . This assay unequivocally distinguishes erythrophagocytozing cells in vivo and in vitro via flow cytometry and fluorescent microscopy . Using this new assay , we show that the acute anemia during experimental trypanosomosis is a result of enhanced erythrophagocytosis by activated liver monocytic cells and neutrophils as well as by activated splenic macrophages . Moreover , the red blood cell membrane composition and stability are altered during the infection , priming them for enhanced clearance by the myeloid phagocytic system .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Development of a pHrodo-Based Assay for the Assessment of In Vitro and In Vivo Erythrophagocytosis during Experimental Trypanosomosis
|
A cross sectional serological survey of arboviral infections in humans was conducted on the three islands of the Union of Comoros , Indian Ocean , in order to test a previously suggested contrasted exposure of the three neighboring islands to arthropod-borne epidemics . Four hundred human sera were collected on Ngazidja ( Grande Comore ) , Mwali ( Mohéli ) and Ndzouani ( Anjouan ) , and were tested by ELISA for IgM and/or IgG antibodies to Dengue ( DENV ) , Chikungunya ( CHIKV ) , Rift Valley fever ( RVFV ) , West Nile ( WNV ) , Tick borne encephalitis ( TBEV ) and Yellow fever ( YFV ) viruses and for neutralizing antibodies to DENV serotypes 1–4 . Very few sera were positive for IgM antibodies to the tested viruses indicating that the sero-survey was performed during an inter epidemic phase for the investigated arbovirus infections , except for RVF which showed evidence of recent infections on all three islands . IgG reactivity with at least one arbovirus was observed in almost 85% of tested sera , with seropositivity rates increasing with age , indicative of an intense and long lasting exposure of the Comorian population to arboviral risk . Interestingly , the positivity rates for IgG antibodies to DENV and CHIKV were significantly higher on Ngazidja , confirming the previously suggested prominent exposure of this island to these arboviruses , while serological traces of WNV infection were detected most frequently on Mwali suggesting some transmission specificities associated with this island only . The study provides the first evidence for circulation of RVFV in human populations from the Union of Comoros and further suggests that the virus is currently circulating on the three islands in an inconspicuous manner . This study supports contrasted exposure of the islands of the Comoros archipelago to arboviral infections . The observation is discussed in terms of ecological factors that may affect the abundance and distribution of vector populations on the three islands as well as concurring anthropogenic factors that may impact arbovirus transmission in this diverse island ecosystem .
Vector-borne infections are mostly sensitive to environmental changes whether natural or anthropogenic . Slight variations in ecologic conditions may severely affect pathogen transmission capacity by impacting the diversity , abundance and/or behavior of vectors , their competence for transmission , together with several traits of the parasite itself [1 , 2] . Unraveling these factors helps understanding how vector borne diseases may express contrasted dynamics in different geographic locations and identifying drivers of emergence acting either at local , regional or large distance scales . Island ecosystems provide ideal in natura conditions allowing to uncouple local transmission from long distance spread [3] . The peculiarities of such ecosystems ( geographic isolation , limited land surface area , low species richness and often-high levels of endemism ) , make oceanic islands most suitable for comprehensive entomological surveys and investigations of original host/pathogen interactions [3] . The Comoros archipelago ( 2144 Km2 ) consists of four volcanic islands that have emerged de novo from the South-Western Indian Ocean ( SWIO ) floor at the Comoros hot spot . They are distant of 40–60 Km from each other and are aligned on a Northwest to Southeast line , at the north end of the Mozambique Channel . The three northern islands constitute the Union of Comoros that ranks amongst the least developed countries in the world . The southern island of Mayotte is administered by France and exhibits significantly higher development indices . The islands of the Union of Comoros are small and sparsely populated . The largest island ( 1146Km2 ) sheltering the capital is named Ngazidja ( Grande Comore in French ) and has 380 000 inhabitants . The second larger island , Ndzouani ( Anjouan ) ( 424 Km2 ) has 300 000 inhabitants , while Mwali ( Moheli ) is the smallest island ( 290Km2 ) and is home to less than 45 000 inhabitants . The geographic proximity to Africa exposes the Comoros archipelago to arboviral emergence as evidenced by a number of epidemics recorded in the last decades . Hence , dengue , chikungunya and Rift Valley fever outbreaks flared up on the Comoros islands in first place before spreading to other islands of the SWIO . Most interestingly , there is evidence that these epidemics have had contrasted courses and severity on the 3 islands of the Union of the Comoros , despite their cultural homogeneity and the frequent inter island mobility of people and livestock . The different exposure of the three islands to the arboviral risk is most evidenced by the last epidemics of dengue and chikungunya . In 1993 , a large epidemic due to Dengue virus serotype 1 ( DENV1 ) affected Ngazidja , with a prevalence of 26% of IgM antibodies to Dengue virus ( DENV ) [4] . In contrast , the outbreak was much less severe on the sister islands of Ndzouani and Mwali where a ten times lower IgM positivity rate was detected [4] , and no dengue epidemic was reported at the same time on Mayotte . In March 2010 , a Dengue virus serotype 3 ( DENV3 ) epidemic was reported on Ngazidja where 1805 suspected cases were registered while the number of suspected cases was only 18 and 4 on Mwali and Ndzouani respectively , as informally reported by the Comorian Health Authorities ( www . reseausega-coi . org/system/files/06-Mlindasse-DENV3_0 . pdf15 ) . Only 76 cases of dengue were confirmed on Mayotte despite an efficient surveillance system [5] . In 2004 , a CHIK outbreak flared up in Lamu , Kenya [6] then spread to Ngazidja where 5202 cases were reported from January to May 2005 . A sero-survey conducted on the island during the epidemic showed an attack rate of 63% with 60% IgM and 27% IgG positive samples [7] According to the Comorian Health Authorities [8] , the outbreak appeared much less severe on the other Comorian islands with 207 cases reported on Ndzouani and only 1 on Mwali . On Mayotte , the epidemic curve delineated two waves [9]: the first wave ( April-June 2005 ) was concomitant to the epidemic on Ngazidja but of minor intensity as assessed by the very low prevalence ( 1 . 6% ) detected on October 2005; the second wave was however explosive: it started on early 2006 and peaked on March-April with a dramatic rise of CHIK virus ( CHIKV ) seroconversion rate . This second wave then expanded into a huge epidemic that involved all the SWIO islands , i . e . Madagascar , Reunion , Mauritius , and Seychelles [10] . Interestingly the second explosive outbreak completely spared the three islands of Union of Comoros where the CHIK dynamics was unimodal . Though previous investigations suggest a differentiated impact of arboviral diseases on the three islands , the available information supporting this view are of unequal solidity with regard to the three islands . Hence , in order to get an unequivocal assessment of this point , we realized a serological investigation in the Union of Comoros in order to delineate the arboviral risk and check whether this assumption is relevant to all viral species or only to some of them .
The present study was implemented at the initiative of Health authorities of the Union of Comoros ( Ministère de la Santé , de la Solidarité et de la Promotion du Genre ) in order to evaluate the impact of vector borne diseases on the general population . The National Malaria Control Program ( PNLP ) was identified by the Health authorities as the promoter of the study to supervise the serological investigation according to a protocol ensuring that all samples were anonymized and the study conducted ethically ( #1175/MSSPG/DNS ) . The present study used excess sera collected from the PNLP and from private laboratories established on Ngazidja , Ndzouani and Mwali . The PNLP has approved that only an oral informed consent should be required from the participants , in agreement with the Comoros cultural norms . The consent was collected by one of the co-authors ( RS ) , who is member of the PNLP and explained the objectives of the investigation , how it will be realized and that anonymity will be strictly respected . Anonymization was achieved before transmission to CRVOI ( Reunion Island ) and the only information made available to this laboratory was the age , gender and the island of origin for each sample . This is a cross sectional study involving 400 individuals living on Ngazidja ( n = 196 ) , Ndzouani ( 88 ) and Mwali ( 116 ) , the three main islands of the Union of Comoros . Anonymous participants were 325 individuals suffering diverse medical conditions , consulting either private laboratories of the three islands or the Surveillance Laboratory for Malaria ( PNLP ) located on Ngazidja . Patients consulting private laboratories had heterogeneous conditions and were mainly suffering from chronic diseases ( diabetes , hypertension etc . ) or occasionally fever , while PNLP patients were mostly febrile . Seventy-five healthy individuals accompanying patients also accepted to participate to the study and represented 22 , 13 and 18% of all individuals enrolled on Ngazidja , Mwali and Ndzouani , respectively . The study was conducted from August 1–October 8 , 2011 and was based on excess sera remaining after laboratory tests motivating the visit to the medical center were performed . The eligibility criteria were acceptance to participate to the study and age over 15 years . Sera were stored at -20°C until testing . The number of required sera from each island was calculated taking into account expected seropositivity rates based on literature [4 , 7 , 11] in order to be able to show a difference of 20% between islands for CHIKV and DENV , and of 6% for RVFV . Hence , the minimum sample size was estimated to 100 per island for CHIKV and DENV and to 190 per island for RVFV ( S1 Table ) . The numbers of sera actually collected and included in the present study were 196 , 116 and 88 originating from Ngazidja , Mwali and Ndzouani , respectively . DENV serotyping was performed using an ELISA-format micro neutralizations test as previously described [15] In this assay , viral neutralization is not measured by the reduction of the number of lysis foci , but rather by the reduction of viral proteins produced in the plate , as detected with a spectrophotometer through an indirect immunoperoxydase antibody test . This assay has been shown to provide monotypic responses similar to the standard serotype-specific PRNT assay in serum post primary infection [16] . Briefly , viruses were grown in 96-well plates in Vero E6 monolayers in the presence of two-fold serial dilutions of serum ranging from 1/40 to 1/2560 . Seven days post-infection , supernatants were removed; cells were fixed with 4% para-formaldehyde and permeabilized with 0 . 5% Triton X-100 . Viral proteins were detected spectrophotometrically using a commercially available pan-flavivirus anti-NS5 monoclonal antibody from hybridoma H86 . 13 B4A supernatant , HRP-conjugated anti-mouse secondary antibodies and TMB as peroxidase substrate . For each serum and virus , the neutralization titer was the reciprocal of the highest serum dilution that inhibited virus protein production , i . e . leading to a mean OD below the positive cut-off value as determined with the corresponding control test DENV . The DENV serotype that is neutralized at the highest serum titer is referred as the dominant serotype . Negative controls were performed using four true negative samples ( courtesy of the French National Reference Centre for Arboviruses ) . The four DENV serotypes used in the assay were H/IMTSSA/98/060 , H/IMTSSA-MART/98-703 , H87 , and Dak HD 34 460 , respectively [16–19] Sera with neutralizing titer ≥40 were considered positive . Seropositivity rates were first calculated for each arbovirus on the whole sample , then according to island , age , and clinical condition ( “healthy” versus “unhealthy” individuals ) and compared together using chi-square or Fisher's exact test .
A total of 300 sera tested positive through ELISA for IgG antibodies to DENV . This consisted of 174 , 77 and 49 sera from Ngazidja , Mwali and Ndzouani , corresponding to 88 . 8% , 66 . 4% and 55 . 7% seropositivity rates , respectively . The difference in seropositivity according to island was highly significant ( p<0 . 0001 ) . As several dengue epidemics hit the Comoros archipelago during the last 70 years , we typed sera against each of the 4 DENV serotypes using the ELISA-format micro neutralization test derived from the plaque reduction neutralization test ( PRNT ) . A subgroup of 90 sera testing ELISA positive for DENV was randomly selected from persons aged less than 50 . Sera from individuals over age 50 were not considered for this test since interpretation of the neutralization profiles is more complicated in older age classes which are more likely to have experienced sequential infections by different DENV serotypes , thus inducing stronger heterotypic DENV neutralization responses . Four samples testing ELISA negative were used as controls . Twelve sera tested negative by the ELISA-format micro neutralization test despite being positive through ELISA to DENV antigen , and were not further considered . Seventy-eight sera ( 54 , 18 and 6 originating from Ngazidja , Ndzouani and Mwali , respectively ) had individual titers over 40 to one or several DENV serotype ( s ) and were considered positive to the reactive serotype ( Table 4 ) . The mean titer of reactivity was 1093 , 835 , 566 and 718 , for serotypes 1 to 4 , respectively . Noteworthy , only few sera ( 14 out of 78 ) , reacted with one single serotype ( Table 5 ) If one considers only the dominant serotypes ( i . e . the serotype identified at the highest titer by each serum sample ) , the most frequently recognized serotypes were DENV serotypes 1 and 3 , particularly on Ngazidja ( Table 4 ) . Noteworthy , serotype 3 was mostly recognized in sera from young persons , and serotype 1 from older individuals . A total of 48 sera were positive for IgG antibodies to CHIKV . A significant difference was observed according to the sampled island ( p = 0 , 0001 ) , with a seropositivity rate higher on Ngazidja ( 18 . 9% ) than on Mwali ( 7 . 8% ) and Ndzouani ( 2 . 3% ) . No difference in seroprevalence was observed according to age . Forty-three sera ( 10 . 7% ) were positive for IgG antibodies to RVFV with seropositivity rates being again statistically different according to the sampled island ( p<0 . 001 ) . Sera from Ngazidja , Mwali and Ndzouani displayed IgG seropositivity rates to RVFV of 15 . 8% , 5 . 2% and 6 . 8% ( see Table 2 ) , respectively . No significant difference in seroprevalence was observed according to age , although it tended to be higher in adults >50 yo ( 18 . 9% ) than in the three other age groups ( 10 . 4% , 9 . 8% and 7 . 6% , respectively ) . Finally , 29 sera were positive for IgG antibodies to WNV by ELISA , with most of them ( 21/29 ) originating from Mwali ( Table 2 ) . Using the most stringent criteria to discriminate between infections by members of flaviviridae family ( See Material and Methods section ) , 21 among these 29 sera were unambiguously positive for WNV ( S1 Table ) , 7 sera yielded ambiguous results between WNV and DENV and one serum between WNV and TBEV . Interestingly , most ( 15/21 ) sera unambiguously testing positive for WNV still originated from Mwali . This was also the case for 5 out of the seven sera with ambiguous positivity to both DENV and WNV ( S1 Table ) . These results support the selectivity of WNV infection on Mwali . Three sera were positive for TBEV and two for YFV and they were all originated from Ngazidja . Altogether , among the 400 sera composing the serobank , 338 ( 84 . 5% ) reacted with at least one of the investigated arboviruses . The seropositivity rate to at least one arbovirus showed a strong association with age ( p<0 . 0001 ) since it was 68 . 7% , 81 . 2% , 90 . 1%% and 95 . 7% in the four successive age groups ( <15yo; 15-30yo; 31-50yo; >50yo ) , respectively .
Ngazidja , the closest to the African continent among the archipelago islands , is directly exposed to the East African coast , a hot spot for the emergence of zoonotic and vector borne diseases [36] Moreover , Ngazidja is the entry point for active trade with continental African countries conveying pathogens and vectors . For example , the bilateral trade bill established in 2000 with Tanzania had allowed importation of cattle to Ngazidja and was soon followed by the emergence of an epidemic of theileriosis on this island likely due to the introduction of the infected tick vector [37 , 38] . Similarly , the higher IgG seropositivity rates to CHIKV , DENV and RVFV measured on Ngazidja in the present study may simply reflect this geographical proximity . However , the higher IgM positivity rates to RVFV in sera from Mwali is in keeping with the fact that livestock was mostly infected on Mwali [29 , 30] . Comorian islands are volcanic islands spanning along a South/North line which ages range from 7 . 7–15 yo in the eldest Mayotte to 0 . 1–0 . 5 yo in the youngest Ngazidja [3] . These stretched ages impact their geologic structure and shape the ecologic conditions that prevail on these islands . For instance , Ngazidja has a very shallow soil layer , which cannot hold water , while the other islands , geologically older , support more advanced soil lateritisation that allows better persistence of surface water . The absence of perennial river on Ngazidja due to the highly porous soil imposes rainfall water to be stored in catchment tanks , creating anthropogenic conditions that are conducive to the proliferation of mosquito larvae . Subtle differences may exist between mosquito populations in the different islands . For instance , several mitochondrial haplotypes displaying some level of geographical structuration have been characterized among natural populations of Ae . aegypti sampled throughout other islands of the South West Indian Ocean and these haplotypes may be associated with distinct morphological and ecological traits [39] . Such information , as well as studies on the vector competence of these distinct lineages towards CHIKV and DENV are not available on the Union of the Comoros and should help understanding the human serological figures highlighted here . The present study shows that Mwali is clearly more exposed to WNV infection than the two sister islands . It is unlikely that this peculiarity relies on significantly different exposure to the bites of Culex spp . , the usual vectors of WNV . It rather suggests , if one considers the major role played by birds in the transmission of this virus , some specificities of Mwali with regard to local birds populations . In fact , Mwali is known to host the major seabird reproduction colonies of the whole Comorian islands [40] . Further investigation of migratory routes together with an analysis of the serological status of both migratory and local bird populations will help clarifying this issue . The main limitation of our investigation is in the study design , which for feasibility reasons was based on the use of excess sera collected from laboratory consultants . Although enrolled participants reside in either of the three main islands , they were not representative of the whole Comorian population . However our results are in keeping with the scarce studies conducted so far in this country indicating an intense exposure of all islands of the Union of Comoros to arboviral infections with salient disparities among islands with regard to some arboviruses . They stress the importance of Ngazidja as the potential entry point for arboviral infections to the whole south Western Indian Ocean region and raise interesting working hypothesis to account for the local specificities of arboviral epidemiology in this tropical island ecosystem . Controlled studies are needed to identify risk factors to various arbovirus infections to which the population of the different islands are exposed .
|
Peculiarities of Island ecosystems make oceanic islands most suitable areas for analyzing how variations in local ecological conditions may impact on the intensity of arboviral transmission . The Comoros archipelago ( 2144 Km2 ) consists of four volcanic islands that are distant of 40–60 Km apart , located at the northern end of the Mozambique Channel . The three northern islands constitute the Union of Comoros that ranks amongst the least developed countries in the world . The geographic proximity to Africa exposes the Comoros archipelago to arboviral emergence as evidenced by a number of epidemics recorded in the last decades ( i . e . dengue , chikungunya , and Rift Valley fever ) . Most interestingly , there is historical evidence that these epidemics have had contrasted courses and severity on the 3 islands of the Union of the Comoros , despite their cultural homogeneity and the frequent inter island mobility of people and livestock . Based on these observations , we conducted a serological investigation in order to delineate the risk to arboviral infection on the 3 islands of the Union of Comoros and assess whether it is relevant to all arboviral species or only to some of them . Our results confirm the differential exposure of the Comoros islands to the arboviral risk . The ecological and anthropological factors that may account for this contrasted epidemiology are discussed .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"rift",
"valley",
"fever",
"virus",
"enzyme-linked",
"immunoassays",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"togaviruses",
"pathogens",
"immunology",
"geographical",
"locations",
"microbiology",
"alphaviruses",
"viruses",
"chikungunya",
"virus",
"rna",
"viruses",
"comoros",
"antibodies",
"immunologic",
"techniques",
"bunyaviruses",
"africa",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"immunoassays",
"arboviral",
"infections",
"people",
"and",
"places",
"biochemistry",
"west",
"nile",
"virus",
"arboviruses",
"flaviviruses",
"viral",
"pathogens",
"physiology",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"organisms"
] |
2016
|
Serological Evidence of Contrasted Exposure to Arboviral Infections between Islands of the Union of Comoros (Indian Ocean)
|
Phytoplankton are key components of aquatic ecosystems , fixing CO2 from the atmosphere through photosynthesis and supporting secondary production , yet relatively little is known about how future global warming might alter their biodiversity and associated ecosystem functioning . Here , we explore how the structure , function , and biodiversity of a planktonic metacommunity was altered after five years of experimental warming . Our outdoor mesocosm experiment was open to natural dispersal from the regional species pool , allowing us to explore the effects of experimental warming in the context of metacommunity dynamics . Warming of 4°C led to a 67% increase in the species richness of the phytoplankton , more evenly-distributed abundance , and higher rates of gross primary productivity . Warming elevated productivity indirectly , by increasing the biodiversity and biomass of the local phytoplankton communities . Warming also systematically shifted the taxonomic and functional trait composition of the phytoplankton , favoring large , colonial , inedible phytoplankton taxa , suggesting stronger top-down control , mediated by zooplankton grazing played an important role . Overall , our findings suggest that temperature can modulate species coexistence , and through such mechanisms , global warming could , in some cases , increase the species richness and productivity of phytoplankton communities .
Phytoplankton fix CO2 from the atmosphere through photosynthesis and underpin the secondary production of many of the world’s aquatic ecosystems [1] , yet the effects of global warming on their biodiversity ( the number of distinct taxa in a given location ) and productivity remains largely unknown . Temperature sets the pace of metabolism [2] and , consequently , a host of life history traits and attributes that determine fitness , including population growth rate , abundance , resource acquisition rate , mortality , and interspecific interactions [3–7] . Global warming could therefore substantially alter local phytoplankton biodiversity through its effects on ecological dynamics ( e . g . , competition , predation ) . For example , it could reduce local biodiversity by increasing metabolic rates ( and hence resource demands ) of individuals , resulting in competitive exclusion through increased resource competition . Alternatively , it could increase local biodiversity by magnifying effects of keystone- or frequency-dependent predation ( e . g . , where consumers switch between resources preferentially to predate upon the most abundant taxa ) , allowing inferior competitors to persist [8–10] . Temperature-mediated effects on species coexistence have received relatively little attention and remain poorly understood ( but see [9] for a notable exception ) . While the experiments conducted thus far suggest that phytoplankton biodiversity declines with increases in temperature [11–15] , extrapolating to the field is challenging because these primarily laboratory-based experiments are conducted over relatively short time scales ( e . g . , weeks to months ) . Such experiments also cannot replicate several key ecological processes that influence local biodiversity dynamics in nature . For example , species turnover occurs in communities through ceaseless immigration and local extinction , the balance of which determines local biodiversity levels [16 , 17] . Laboratory experiments are disconnected from a regional species pool and exclude dynamics of dispersal-mediated community assembly , which are important for determining metacommunity structure [16] . In addition , such experiments cannot fully capture the material recycling processes that shape natural ecosystem succession [18] . The few experiments conducted in outdoor mesocosms , and which can account for metacommunity dynamics , have not investigated the effects of warming on patterns of biodiversity and species coexistence [19 , 20] . In contrast to laboratory experiments , eco-evolutionary species distribution models of the responses of phytoplankton biodiversity to ocean warming suggest net losses in the tropics and gains at the poles [21] . These models , however , ignore trophic interactions [21] , which could play a central role in determining how warming affects competition and coexistence in phytoplankton communities [22] . These limitations , both in published experiments and in current theory , highlight important gaps in our understanding of how the biodiversity and productivity of phytoplankton communities might respond to global warming . We used a long-term , outdoor mesocosm experiment to investigate the effects of warming on the structure , biodiversity , and functioning of plankton communities . Our experimental approach differs in several fundamental ways from previous studies that have investigated the effects of warming on phytoplankton biodiversity [11–15] . First , our outdoor mesocosms are open to natural dispersal ( aerial and/or transportation on mobile vertebrates ) from the regional species pool , allowing us to explore the effects of experimental warming in the context of metacommunity dynamics . Second , because these mesocosms comprise both benthic and pelagic communities , they include fundamental material recycling processes and can recreate many of the salient biogeochemical features of natural aquatic ecosystems [23–26] . Finally , the mesocosms had been warmed for 5 yr at the time of sampling , allowing us to characterize biodiversity responses in the context of long-term successional and ecosystem dynamics over tens-of-thousands of generations for the phytoplankton . We carried out a detailed empirical survey to assess how long-term experimental warming altered the structure , diversity , and taxonomic composition of the phytoplankton communities . We then characterized how the biomass , size structure , and composition of the zooplankton ( the consumers of the phytoplankton ) were affected and determined the extent to which warming altered the functioning ( e . g . , gross primary production [GPP] and community respiration [CR] ) of the plankton communities . Finally , we use these measurements , taken across multiple levels of organization ( population , community , & ecosystem ) to explore the mechanisms that link the structure and dynamics of the plankton community to the ecosystem processes they mediate and the way in which warming altered the functioning of these ecosystems .
Average taxonomic richness of the phytoplankton communities was 67% higher in the warmed mesocosms ( Fig 1A; S1 Table ) . The Shannon Diversity Index also increased with warming ( Fig 1B; S1 Table ) , as did rarefied richness ( S3 Fig ) and total biomass ( Fig 1C; S1 Table ) , while total abundance was unaffected by warming ( Fig 1D; S1 Table ) . These results demonstrate that the higher diversity of the warmed communities cannot be attributed to greater numbers of individuals . The rank abundance distributions ( RADs ) of the phytoplankton communities for the warmed treatments were markedly flatter , with more equitable allocations of individuals among more taxa , compared to their ambient counterparts ( Fig 2A ) . Parameter estimates from the Poisson log-normal fits ( see Methods ) also demonstrated that the standard deviation of log-abundance per taxon ( σ ) was lowest in the warmed treatments ( Fig 2B; S1 Table ) , whereas the average of log-abundance per taxon ( μ ) did not differ between treatments ( Fig 2C; S1 Table ) . Overall , these findings indicate that long-term warming altered patterns of dominance and diversity in these local communities . Warming also altered the taxonomic composition of the phytoplankton . Nonmetric multidimensional scaling ( NMDS ) revealed a statistically significant separation in the taxonomic composition of the warmed and ambient treatments ( Fig 3A; PERMANOVA; F1 , 15 = 3 . 75; p = 0 . 0019 ) . These patterns could reflect deterministic [27 , 28] effects of temperature on mechanisms of community assembly , for example by altering optimal thermal niches or by changing species interactions . They could also simply reflect stochastic processes of colonisation and extinction [27 , 28] , or a combination of both stochastic and deterministic factors [29] . To investigate the effects of warming on mechanisms of community assembly , we used a null-model approach [27 , 28] to ask: do patterns of β-diversity ( i . e . , pond-to-pond differences in taxonomic composition ) deviate from those expected under a completely stochastic assembly process given the observed patterns of α- and γ-diversity ( i . e . , the “local” ( one pond ) and “regional” ( all ponds ) numbers of taxa , respectively ) ? We used the rescaled Raup-Crick metric ( βRC ) , which varies from −1 ( communities more similar than expected by chance ) to 1 ( communities more dissimilar than expected by chance ) to quantify the relative roles of stochastic versus deterministic factors in driving metacommunity assembly in our experiment [28] . Values of βRC between ambient ( z = −2 . 47 , p = 0 . 019 ) , warmed ( z = −2 . 44 , p = 0 . 013 ) and ambient-warmed ( z = −5 . 86 , p = 0 . 0009 ) comparisons were significantly lower ( e . g . , more similar ) than expected by chance alone ( Fig 3B ) , suggesting that deterministic factors dominated mechanisms of community assembly across the experiment . The trait distributions of the phytoplankton communities also differed systematically between the warmed and ambient treatments . The latter were dominated by smaller , single-celled genera of algae—e . g . , Chlamydomonas , Chlorella , Chromulina—while the former included larger , filamentous cyanobacteria and colonial algae—e . g . , Anabaena , Spirogyra , and Chlamydocapsa . Average body masses were an order-of-magnitude higher in the warmed treatments ( Fig 4B; S1 Table ) . Larger phytoplankton taxa , with diameters > 35 μm , are generally considered less susceptible to predation , because they are too large to be handled effectively and consumed by zooplankton [30 , 31] . Correspondingly , the proportional abundances of these large , inedible phytoplankton taxa were significantly higher in the warmed treatments ( S4 Fig; S1 Table ) . The effects of warming on zooplankton community structure were far weaker than for the phytoplankton: neither total biomass nor average body mass differed significantly between warmed and ambient treatments ( S1 Table; Figs 4C and 5A ) . The total biomass of cladocerans and copepods also did not differ between treatments ( S1 Table; Fig 5B ) . However , further dividing the grazers into the dominant genera revealed some treatment effects , although these were subtle , and apparently compensatory: declines in the biomass of Diaptomus were partially compensated by increases in the biomass of Chydorus in the warmed treatments , while the biomass of Daphnia , Bosmina , Alona , and ostracods were not affected by warming ( S5 Fig ) . Gross primary production ( GPP ) was higher in the warmed treatments , while rates of plankton CR did not differ significantly between treatments ( Fig 6A; S1 Table ) . We used path analysis to test hypotheses about the direct and indirect interactions between warming , the structure of the phytoplankton and zooplankton communities and rates of ecosystem metabolism . In the best fitting model , CR was directly and positively correlated with temperature ( Fig 6B ) . In contrast , warming increased GPP indirectly , because higher temperatures enhanced phytoplankton taxon richness , which in turn elevated levels of phytoplankton biomass and thus rates of GPP ( Fig 6B ) . The best-fitting model also included positive direct effects of temperature on phytoplankton taxon richness ( Fig 6 ) . These findings demonstrate that warming altered ecosystem functioning both by directly enhancing metabolic rates ( CR ) and by indirectly changing the structure of the phytoplankton communities ( impacting GPP ) .
Warming resulted in profound shifts in the organization and biodiversity of the local phytoplankton communities , increasing taxonomic richness by 67% ( Fig 1 ) . The distributions of abundance among taxa also became more even ( Fig 2 ) , and the taxonomic compositions and distributions of traits ( e . g . , body mass , edibility ) were also altered markedly by warming ( Fig 3 & Fig 4 ) . Rates of GPP and CR were elevated in the warmed treatments and path analysis revealed that warming both directly ( by stimulating metabolic rates ) and indirectly ( by increasing the biodiversity and biomass of the phytoplankton ) influenced ecosystem functioning . These results demonstrate that ecological mechanisms that determine the number of species that can coexist locally could play an important role in mediating the effects of global warming on the biodiversity and functioning of primary producers in planktonic ecosystems . Previous experiments have demonstrated that warming tends to reduce the mean body size , total biomass , and biodiversity of phytoplankton communities [14 , 32–34] . These findings have been attributed to greater competition among phytoplankton for limiting nutrients as a consequence of temperature-induced increases in rates of metabolism and resource uptake [33] . Small species , with high surface area-to-volume ratios and rapid growth rates , are at a selective advantage at high temperatures and low nutrient concentrations and consequently tend to dominate phytoplankton communities under these conditions [1] . Reductions in mean body size at the community level have been coined the “third universal response to global warming” [35 , 36] . Our results reveal the exact opposite pattern . Phytoplankton communities in the warmed treatments were more species rich , had greater evenness and standing stocks of biomass , and were dominated by larger species . So what mechanism ( s ) might explain these unexpected findings ? Experimental warming of aquatic food webs tends to enhance top-down regulation of adjacent trophic levels [37–41] by increasing temperature-dependent consumption rates [5–7 , 42] . High consumption rates can enhance species coexistence when they are frequency-dependent [9 , 10] and such active prey-switching to target the most abundant resource has been demonstrated for zooplankton [43 , 44] . It is widely thought to be a key mechanism leading to stable coexistence among resource taxa when strategies leading to higher competitive ability also increase vulnerability to consumption by selective consumers [45–47] . Warming systematically shifted the taxonomic composition of the phytoplankton towards taxa that are more resistant to grazing ( Fig 4 & S4 Fig ) , owing to larger cell size and/or colonial or filamentous growth form [30 , 48] , suggesting stronger top-down control in the warmed treatments . The strength of interactions between consumers and resources increase as a function of the ratio in their body sizes [49] and the ambient temperature [7] . Recent theory predicts that as community assembly dynamics approach immigration–extinction equilibria , consumer–resource body size ratios should converge towards those that are most stable [50] . Our experimental evidence is consistent with this expectation , because increases in the prevalence of large phytoplankton serve to decrease consumer-resource body size ratios ( because zooplankton body size was unaffected by warming , see Fig 4C ) , thereby adding weak trophic interactions to the warmed food webs and counteracting the effects of temperature on interaction strengths by stimulating metabolism . Warming did not affect the composition , body mass , or biomass structure of the zooplankton . This seems surprising given the major changes observed in the phytoplankton communities . Warming shifted the phytoplankton communities towards large taxa , presumably as a response to selection for grazer-resistant morphology [30 , 48] owing to higher temperature-dependent consumption rates . The lack of a concomitant shift in zooplankton composition and/or size structure might simply reflect the absence of appropriate traits within the regional species pool . Daphnia spp . are among the largest zooplankton grazers in freshwater ecosystems and were the dominant taxon in both the ambient and warmed treatments ( S5 Fig ) . It seems likely that larger taxa , which could track the order-of-magnitude increase in mean body size observed in the phytoplankton , were simply absent from the regional species pool . Nevertheless , the higher temperatures will have increased the metabolic demands of the zooplankton in the absence of any shifts in composition , body mass and biomass structure . Path analysis demonstrated that plankton CR was directly and positively correlated with temperature . By contrast , warming increased GPP indirectly , because higher temperatures enhanced phytoplankton taxon richness , which in turn elevated levels of phytoplankton biomass and thus rates of GPP ( Fig 6B ) . This result suggests that after controlling for other correlated variables , heterotrophic metabolism ( a proportion of CR ) was more strongly affected by temperature than rates of autotrophic metabolism ( GPP ) , in line with previous findings in both freshwater and marine ecosystems [25 , 37 , 51 , 52] . Thus , if elevated metabolic rates are correlated with higher consumption rates by heterotrophs , then this once again points towards an increase in the strength of top-down control in the warmed treatments . The results presented here for the long-term effects of warming on the structure and diversity of phytoplankton contrast sharply with those seen during the first year of this experiment [32] and provide further clues as to the mechanisms responsible for the observed patterns . At the onset of the experiment , the planktivorous fish , Rutilus rutilus was added to the mesocosms , but due to poor survivorship , they were removed from all mesocosms at the end of the first year of warming ( see Materials and Methods ) . During the first year of warming , in the presence of the planktivore , genus richness , Shannon-Diversity , and mean body mass of the phytoplankton were all significantly reduced in the warmed treatments , while total abundance increased [33 , 53] ( see S8 Fig ) . The differential response of the phytoplankton communities in the presence and absence of the fish suggest that warming interacts strongly with the trophic cascade in planktonic communities . When fish are present , their effect on zooplankton releases the phytoplankton from strong , warming-induced , top-down control , resulting in increased interspecific competition for nutrients and selection for small but edible organisms , which are good competitors for nutrients [33 , 53] . Conversely , when fish are absent , warming increases the strength of top-down control of zooplankton on the phytoplankton , which relaxes interspecific competition for nutrients and favors large and inedible taxa , which are inferior competitors for nutrients , and increases diversity . The overall weight of experimental evidence ( e . g . , flatter species abundance distributions , shifts in the taxonomic composition and trait distributions of the phytoplankton towards large colonial or filamentous algae , elevated biomass specific rates of respiration ) points towards increases in top-down control , leading to suppression of competitive exclusion and greater species coexistence as the most likely explanation for the increased biodiversity and productivity of the phytoplankton under long-term warming . This is supported by a null model analysis , which demonstrated that deterministic processes dominated mechanisms of community assembly in the phytoplankton , and a path analysis of the variables representing the structure and functioning of the plankton communities . However , without directly manipulating the grazers , we cannot isolate this as the only mechanism driving the shifts in community structure and function in the warmed treatments . Irrespective of the ultimate mechanism , however , our findings demonstrate that in open systems , where local extinctions can be counterbalanced by dispersal-mediated immigration from the regional species pool ( e . g . , metacommunity dynamics ) , warming could actually lead to increases ( as well as decreases ) in biodiversity and ecosystem functioning , in contrast to the received wisdom based on laboratory experiments in closed systems [11–15] . Taken together , our results emphasize the fundamental role temperature plays in constraining patterns of species coexistence and dominance in local communities . They also suggest that warming can alter ecosystem functioning indirectly , by shifting community structure and biodiversity , in addition to its well-known direct effects mediated by metabolism [2] . The findings we report are significant because they show that temperature can enhance the diversity of local communities through ecological mechanisms . Moreover , they mirror biodiversity patterns reported for aquatic and terrestrial taxa along broadscale gradients of temperature and latitude [54] . Latitudinal biodiversity gradients are typically attributed to long-term macroevolutionary and/or historical mechanisms [55]; however , our findings , which isolate the effects of temperature while controlling for other variables that may be confounded along latitudinal gradients ( e . g . , nutrients , productivity , disturbance regime ) , suggest that temperature may in part influence local biodiversity through its effects on ecological mechanisms of species coexistence . Through such mechanisms , future global warming could , in some cases , actually enhance species richness and primary productivity in phytoplankton communities .
The outdoor mesocosm experiment ( see S9 Fig ) is based at the Freshwater Biological Association’s River Laboratory ( 2°10`W , 50°13`N ) in East Stoke , Dorset , United Kingdom . Twenty artificial ponds , each holding 1 m3 of water , were set up to mimic–shallow standing freshwater ecosystems . The pool of species available for initial colonization was standardized at the outset by seeding all of the ponds in December 2005 with a “common garden” inoculum of organisms from surrounding freshwater habitats . Ponds were then left open to natural colonization and dispersal . Populations of an introduced planktivorous fish , R . rutilus , were maintained at constant densities ( two individuals ( age 1+ ) per mesocosm ( ~12 g C m−3 ) ) in all mesocosms until October 2007 . R . rutilus populations were removed in October 2007 by electrofishing due to ongoing poor survival . Experimental warming began in September 2006 [25] and had run for 5 yr ( including 4 yr without fish ) at of the onset of sampling in July 2011 . Ten of the twenty ponds were warmed 3–5°C above ambient temperature ( see S1 Fig ) , in accordance with the IPCC A1B global warming projections for the next 100 yr for temperate northern hemisphere regions [56] . Mesocosms were warmed by an electronic heating element connected to a thermocouple , which was used to monitor the temperature in a given pair of warmed and ambient mesocosms . Over the 5-yr experiment , temperatures were logged every 5 min , and minor adjustments were made , if required , to ensure that temperature differences between treatments were ~4°C; the mean annual temperature difference between treatments over the year of sampling was 4 . 4°C ± SE 0 . 03 ( S1 Fig ) . During the second year of the experiment , the heating elements in two of the warmed ponds malfunctioned . We therefore removed these replicates from the experiment , as well as the ambient replicates to which they were paired . All analyses presented here are on the 16 remaining replicates ( 8 warmed; 8 ambient ) . Water temperature was , as expected , significantly elevated in the warmed treatments ( S1 Table; S1 Fig ) . By contrast , total dissolved inorganic nitrogen ( S2 Fig ) and orthophosphate ( S2 Fig ) were , on average , statistically indistinguishable between the warmed and ambient treatments over the annual cycle ( S1 Table ) . Taken together , these two results show that of the variables we measured , temperature was the principal abiotic variable altered by experimental warming . The plankton community from each of the 16 mesocosms was sampled every two months between July 2011 and July 2012 ( 7 sampling occasions in total ) . The entire water column from the sediment surface to the water surface was sampled using a 0 . 8 m-length tube sampler ( Volume: 2 L ) , which was positioned at random in each mesocosm on each sampling date . Each sample was divided into two size categories ( >100 μm , <100 μm ) for preservation and subsequent analyses , via filtration through a 100 μm aperture sieve: organisms >100 μm were preserved in 4% Formalin , and a 100 mL subsample of organisms <100 μm was preserved in 1% Lugol’s iodine . Phytoplankton <100 μm were counted using a LEICA DMIRB inverted microscope at 400x magnification , following the Utermöhl method [57] . The microscope was connected to an interactive image analysis system ( LEICA EC3 camera and LAS software ) to allow for a higher magnification . For each sample , at least 400 individuals ( single cell , colony or filament ) were counted , measured , and identified . Counts were converted to volumetric estimates of abundance ( organisms mL−1 ) based on the volume of sample analyzed , which varied between 1 and 25 mL depending on the density of organisms . In total , 171 taxa were identified , 85% of which were identified to species level; the remaining 15% were identified to genus or class , or were undetermined ( see S4 Table ) . Plankton >100 μm ( typically zooplankton ) were counted , measured , and identified using a Nikon SMZ1500 dissection microscope . Of the six most dominant zooplankton taxa , which together accounted for 91% of the total zooplankton biomass , five were identified to genus level and one to class . This group included the key cladoceran and copepod genera ( e . g . , Daphnia spp . , Bosmina spp . , Chydorus spp . , Alona spp . , Diaptomus spp . ) as well as the ostracods , which are important planktonic grazers in lakes [58] . The remaining taxa were identified to the highest possible taxonomic level , typically class or family . Linear dimensions of each individual were determined using image analysis . The “size” of each organism was expressed in units of carbon mass ( μg C ) . To estimate masses of organisms >100 μm ( typically zooplankton ) , biovolumes were first determined by assigning organisms to geometric shapes that closely represented the real shape of the organism [59] . Masses were then calculated by converting biovolume to fresh weight using a factor of 1 . 1 g mL−1 and converting fresh weight to carbon content assuming a dry:wet weight ratio of 0 . 25 and a dry carbon content of 40% [59] . To estimate masses of organisms <100 μm ( typically phytoplankton ) , biovolumes were estimated by assigning organisms to matching geometric shapes [60] . For all phytoplankton individuals , biovolume was estimated by considering the linear dimensions of the whole organism , this included single cells ( e . g . , Chlorella ) , entire filaments ( e . g . , Anabaena planctonica ) , and colonies along with their mucilage ( e . g . , Aphanothece , Chlamydocapsa , Nephrocytium , Sphaerocystis ) [60] . Biovolume was then converted to carbon units following Montagnes et al . [61] . In total , >83 , 000 organisms , including both phytoplankton and zooplankton , were measured and identified . On four of the seven sampling occasions ( May , July , September , November ) spanning the main growing season , we measured in situ rates of planktonic and benthic CR and GPP in each mesocosm . At midday , on each of the sampling occasions , we incubated samples of 0 . 5 L of the pelagic plankton community of each pond in custom-made paired clear and opaque polycarbonate chambers . At the same time , benthic communities were incubated in custom-made paired clear and opaque bottomless chambers that were screwed into the sediment to a depth of approximately 5 cm [62] . Both the pelagic and benthic chambers were equipped with magnetic stirrers ( rotating at 300 rpm ) to ensure even mixing within the chamber and with a Unisense OX50 microelectrode to measure the net production or consumption of O2 over a 30-min incubation period . Net community production ( NCP ) was measured as the change in O2 concentration in the clear chamber , while CR was the change in O2 in the opaque chamber . The concentration of O2 was measured every minute , and the rate of metabolism ( NCP or CR ) was determined as the slope of a linear regression between O2 ( μmol O2 L−1 – pelagic; μmol O2 m−2 – benthic ) and time ( h ) . GPP ( μmol O2 L−1 ) was estimated as the sum of the NCP and CR terms for both the pelagic and benthic zones , taking the appropriate chamber volumes and sediment surface areas into account . Water samples for measuring dissolved inorganic nutrient concentrations were collected from mid-depth in each mesocosm at 9:00 a . m . on each sampling occasion . Samples were filtered ( Whatmann GF/F ) and stored frozen at −20°C for subsequent determination of NO3− , NO2− , NH4+ , and orthophosphate ( HPO42− + PO43− ) using a segmented flow auto-analyzer ( Skalar , San++ , Breda , Netherlands ) [63] . We used path analysis ( a variant of structural equation modeling that uses only observed variables ) to determine the extent to which warming altered ecosystem functioning directly , by stimulating rates of metabolism , or indirectly , by also altering the biomass and diversity of the plankton communities . Path analysis is a useful tool for studying complex biological systems , because it allows the user to link together component models for different response variables , enabling estimation of direct and indirect effects as well as the overall fit of a complex causal network of influence [70] . Because our experimental design resulted in hierarchical data , with multiple measurements of variables made seasonally , nested within replicate mesocosms , we used Shipley’s method of building a multilevel path model using directional separation tests [71] . This approach entailed assembling a path model as a set of hierarchical linear mixed effects models , each of which included hypothesized relationships between a response variable and a set of predictors as fixed effects and mesocosm ID as a random effect on the intercept . Each mixed effects model was fitted using the “lme” function in the “nlme” package , and the overall path model combining models for each response variable was fit using the “piecewiseSEM” package in R [72] . The predictors included temperature as an exogenous variable ( i . e . , whose variance arise outside of the model ) and GPP , CR , phytoplankton taxon richness , phytoplankton biomass , and zooplankton biomass as endogenous variables ( i . e . , those whose variance the model seeks to explain ) . All variables were natural-log transformed ( except for temperature ) and standardized using the mean to linearize relationships and reduce correlations between the coefficients . We hypothesized that warming both directly stimulated rates of GPP and CR by simply increasing metabolic rates and indirectly influenced these fluxes by altering the richness and biomass of the phytoplankton and zooplankton communities . We developed a set of candidate models , starting with a full model , which included all feasible paths between variables ( see S7 Fig ) . All paths between variables were treated as “free parameters” , and their direction and magnitude were estimated from the observed data by the model . We then removed paths to generate a set of candidate models and selected among models with two criteria . First , we used Shipley’s test for directional separation , which combines the significance of unrealized paths into a single Chi-squared distributed Fisher’s C statistic as a measure of goodness of fit [71] . For each candidate model that passed this test of adequate fit ( e . g . , p > 0 . 05 ) , we computed a small sample-size corrected AIC score ( AICc ) [73] . We then compared between models by calculating delta AICc values and AIC weights ( S2 Table ) . The relative importance of paths in the final model were compared using standardized coefficients , which are dimensionless and express a percentage change in the observed range of the response as the predictor increases across its range .
|
At the global scale , phytoplankton take up about as much carbon dioxide ( CO2 ) as the tropical rainforests . However , in spite of their importance in global carbon cycles , we understand very little about how phytoplankton communities and the critical functions they mediate , including CO2 sequestration , are likely to change as the climate warms in the coming decades . In this study , we report the results of a five-year warming study in experimental outdoor ponds , known as mesocosms . Warmed ( +4°C ) communities had 67% more species and higher rates of gross primary productivity ( CO2 fixation ) . Our results show that warming resulted in higher productivity by increasing the biodiversity and biomass of the phytoplankton . Warming also changed the species composition of the phytoplankton communities by favouring larger organisms that were more resistant to grazing from zooplankton . Our work demonstrates that future global warming is likely to have major impacts on the composition , biodiversity , and functioning of planktonic ecosystems by affecting metabolic rates and species interactions . The increases in the biodiversity and productivity of the phytoplankton seen in this study also highlights that the effects of a warming environment might not always be adverse for all ecosystems .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Five Years of Experimental Warming Increases the Biodiversity and Productivity of Phytoplankton
|
Insect flight is regulated by various sensory inputs and neuromodulatory circuits which function in synchrony to control and fine-tune the final behavioral outcome . The cellular and molecular bases of flight neuromodulatory circuits are not well defined . In Drosophila melanogaster , it is known that neuronal IP3 receptor mediated Ca2+ signaling and store-operated Ca2+ entry ( SOCE ) are required for air-puff stimulated adult flight . However , G-protein coupled receptors ( GPCRs ) that activate intracellular Ca2+ signaling in the context of flight are unknown in Drosophila . We performed a genetic RNAi screen to identify GPCRs that regulate flight by activating the IP3 receptor . Among the 108 GPCRs screened , we discovered 5 IP3/Ca2+ linked GPCRs that are necessary for maintenance of air-puff stimulated flight . Analysis of their temporal requirement established that while some GPCRs are required only during flight circuit development , others are required both in pupal development as well as during adult flight . Interestingly , our study identified the Pigment Dispersing Factor Receptor ( PdfR ) as a regulator of flight circuit development and as a modulator of acute flight . From the analysis of PdfR expressing neurons relevant for flight and its well-defined roles in other behavioral paradigms , we propose that PdfR signaling functions systemically to integrate multiple sensory inputs and modulate downstream motor behavior .
The evolution of flight in insects is linked to a number of natural behaviors including identifying food sources , mates and sites for egg-laying . The complexity of such behaviors frequently requires multiple sensory inputs that act directly and indirectly through neuromodulatory circuits , to control and fine-tune the final behavioral outcome [1] , [2] . In the context of insect flight , the cellular and molecular bases of these neuromodulatory circuits are as yet ill-defined . Our interest in the flight circuit arose from the observation that mutations in the inositol 1 , 4 , 5-trisphosphate receptor ( IP3R; itpr; [3] , [4] ) , a ligand-gated Ca2+ channel that responds to IP3 generated after GPCR stimulation , resulted in strong flight deficits in Drosophila . These results suggested that G-protein coupled receptors ( GPCRs ) linked to IP3/Ca2+ signaling may play an important role in regulating flight behavior . While receptor tyrosine kinases can also initiate IP3/Ca2+ signaling in vertebrates , genetic evidence in Drosophila does not support this mode of IP3R activation [5] . The Drosophila genome contains ∼200 GPCRs , of which ∼90 have been identified as either gustatory or olfactory receptors [6] , [7]; of the remaining GPCRs , although the ligands for most have been identified , the physiological function of only a small number is known . Some of the GPCRs are either identified or putatively assigned as receptors for neuropeptides that regulate feeding and foraging behavior , walking , modulation of visual processing and the response to stress [8]–[11] . Three neuropeptides ( SIFamide , sex peptide and NPF ) and their cognate receptors have been implicated in courtship behavior [12]–[14] . Recently , the receptors for DSK-1 , DSK-2 and CCKLR-17D1 have been shown to regulate larval locomotion [15] . However , GPCRs that are involved in regulation of flight are still being discovered . Recent pharmacological evidence has implicated various monoamines such as octopamine , dopamine , tyramine and histamine ( and presumably their receptors ) and the muscarinic acetylcholine receptor ( mAcR ) in locust flight initiation [16] . The Drosophila mAcR increases IP3 dependent intracellular Ca2+ upon activation by its agonist in transfected S2 cells [17] , [18] and in primary neuronal cultures from Drosophila [4] . Drosophila mutants that reduce octopamine levels exhibit flight initiation and maintenance defects which can be suppressed by pharmacological blocking of Tyramine receptors [19] . Signaling downstream of the Tyramine receptors suggests multiple mechanisms including cAMP [20] , [21] . Here , we describe a genetic RNAi-based screen to identify GPCRs that regulate flight through IP3 mediated Ca2+ signaling . Among the GPCRs identified , two were previously known to activate IP3/Ca2+ signaling in neurons , but were not known to regulate flight in Drosophila . Furthermore , we show that GPCR signaling is required during development of the flight circuit as well as for modulation of adult flight . One of the GPCRs identified in our screen is the receptor for the Pigment Dispersing Factor or PdfR [22] . From analysis of PdfR expression in the nervous system in the context of flight and its well-defined roles in other behavioral paradigms , we propose that PdfR signaling functions systemically to integrate multiple sensory inputs and modulate downstream motor behavior .
To identify G-protein coupled receptors ( GPCRs ) that activate Gq-Plcβ signaling leading to IP3R mediated Ca2+ release ( Figure 1A ) during flight circuit development and function , an RNAi-based screen was designed with the UAS-GAL4 system ( Figure 1B ) . A total of 224 UAS-RNAi strains specific for 108 non-olfactory and non-gustatory GPCRs were selected based on a previous bioinformatic analysis ( [6] and Table S1 ) . Each of these RNAi strains were expressed individually using the pan-neuronal ElavC155GAL4 strain which expresses in all post-mitotic neurons [23] . As indicated in the methods section , only adult female flies were tested for analysis of air-puff induced flight initiation and maintenance ( Figure 1B ) . Normal initiation and maintenance of flight was observed upon pan-neuronal knockdown of 86 GPCRs ( Figure 1C ) , while pan-neuronal knockdown of 22 GPCRs resulted in flight time of less than 80% ( Figure 1D and Table S1 ) . Flies with pan-neuronal knockdown of the 5HT1a receptor ( 16720-2 ) , neuropeptide F receptor ( 1147-2 ) , dromyosuppressin receptor 1 ( 8985-4 ) and methuselah-like 7 receptor ( 7476-3 ) showed wing posture ( expanded wings ) defects which affected their flight ability ( Figure S1 ) ; wing-posture defects however were not uniform , with a small fraction exhibiting normal wings and flight ( data not shown ) . Pan-neuronal knockdown of other methuselah-like receptors such as the methuselah-like 8 receptor ( 32475-2 ) , methuselah-like 9 receptor ( 17084-3 ) and methuselah-like 6 receptor ( 16992-3 ) showed similar expanded wing phenotypes in a fraction of the animals . Flies with normal wings also showed normal flight ability ( data not shown ) . Pan-neuronal knockdown of the SiFamide receptor ( 10823-1 , SiFaR ) resulted in lethality during pupal stages ( see later ) . Therefore , our screen at this stage yielded 22 putative GPCRs whose function appeared to be required for maintenance of air-puff induced flight in Drosophila . Previous studies have shown that pan-neuronal knockdown of the IP3R with an inducible RNAi leads to significant defects in wing posture and flight ( [24]; Figure 2A and 2B ) . To identify GPCRs that stimulate IP3 mediated Ca2+ release a secondary suppressor screen was devised and tested as follows: dgq codes for the alpha subunit of the heterotrimeric G-protein , Gq and activates phospholipase Cβ upon binding of the cognate ligand to the GPCR ( Figure 1A ) . Genetic interactions in the context of Drosophila flight have been demonstrated previously between dgq and itpr mutants [5] . dSTIM codes for the Drosophila STromal Interaction Molecule ( dSTIM ) which functions as a sensor of endoplasmic reticulum ( ER ) store Ca2+ [25] , [26] . Depletion of ER Ca2+ activates STIM followed by opening of the Orai ( dOrai ) surface channel , also referred to as the store-operated Ca2+ entry ( SOCE ) channel . Previous observations with itpr mutants support the idea that STIM and Orai function with the Sarco-Endoplasmic Reticulum Ca2+ ATPase pump ( SERCA ) to restore Ca2+ levels in the ER lumen of Drosophila neurons after GPCR activation and IP3-mediated Ca2+ release [4] , [5] , 24 . Therefore pan-neuronal expression of either a constitutively activated form of Gq ( GqQ203L or AcGq; [27] ) or dSTIM+ were first tested for their ability to suppress flight deficits in flies with pan-neuronal knockdown of the IP3R , using a previously validated itpr RNAi strain ( dsitpr; [24] ) . Pan-neuronal knockdown of the IP3R leads to a near complete flight deficit ( 4%±1 . 94 ) . While dSTIM+ over-expression could suppress this loss of flight and restore it up to 65% , in AcGq expressing animals the flight deficit was restored to 50% ( Figure 2B ) . Physiological correlates of flight , such as electrophysiological recordings from the dorsal longitudinal muscles ( DLMs ) of dsitpr; dSTIM+ and dsitpr; AcGq expressing flies showed that 11/15 flies flew normally and 4/15 flies flew for 15 sec with dSTIM+ while 3/15 flies flew normally and 12/15 flies flew for 10–15 sec with AcGq ( Figure 2C ) . Wing posture defects and spontaneous firing from the DLMs observed in flies with pan-neuronal knockdown of the IP3R were rescued in all flies by expressing either AcGq or dSTIM+ ( Figure 2A and 2D ) . Thus , reduced signaling through the IP3R in Drosophila flight circuit neurons can be restored significantly either by increasing the active form of Gq ( AcGq ) or by raising SOCE through over-expression of dSTIM+ . GPCRs linked to IP3R mediated Ca2+ signaling and required for the maintenance of flight were identified from amongst the 22 receptors shown in Figure 1D by individual pan-neuronal GPCR knockdowns in the context of over-expression of AcGq and dSTIM+ transgenes . The resulting progeny were tested in the single flight assay ( Figure 3A and S2 ) . Out of the 22 putative receptors , flight was rescued to a significant extent for 4 receptors , namely mAcR ( CG4356 ) , CCH1aR ( CG30106 ) , PdfR ( CG13758 ) and FmrfR ( CG2114 ) , by over-expression of either dSTIM+ or AcGq or both ( Figure 3A ) . Therefore , our screen identified mAcR , CCH1aR , PdfR and FmrfR as the GPCRs that are required for the maintenance of Drosophila flight through IP3 mediated Ca2+ signaling . From the remaining 18 receptors , flight defects for the frizzled-2 receptor ( CG9739 or dFz-2R ) were suppressed to a significant level by expression of dSTIM+ . Interestingly , AcGq expression did not have a significant effect in flies with pan-neuronal knockdown of dFz-2R ( Figure 3A ) . In flies with knockdown of CG43795 ( two independent RNAi constructs: 43795-1 , 43795-2 ) , rhodopsin-like receptor ( 16740-2 ) , a neuropeptide receptor ( 34411-2 ) , trapped in endoderm ( 3171-2 ) and diuretic hormone 44 receptor 2 ( 12370-1 ) , flight time reduced further upon pan-neuronal expression of either AcGq or dSTIM+ or both ( Figure S2 ) . Flight time was reduced significantly by both in 34411-2 , 3171-2 and 43795-2 ( Figure S2 and 3A ) . However , only 43795-2 was investigated further ( see discussion ) . Pupal lethality was observed upon knockdown of the SiFamide receptor ( 10823-1; SiFaR ) in neurons ( Figure 3A and 3B ) which could be rescued completely by pan-neuronal expression of dSTIM+ . Interestingly , there was no rescue of lethality by AcGq ( Figure 3B ) . Adult flies that eclosed after over-expressing dSTIM+ in background of SiFamide receptor down-regulation had normal wings , but showed significantly reduced flight time ( 40%; Figure 3A ) . While the validation by AcGq and dSTIM+ expression helped confirm signaling through the IP3R in the case of receptors shown in Figure 3A , it was of concern that in each case just one RNAi line for each GPCR showed flight deficits . We therefore tested the efficiency of GPCR knockdown for each RNAi strain , validated by rescue with either AcGq or dSTIM+ , in a qPCR analysis . Two RNAi lines were selected for each validated GPCR; one that gave a flight defect and another that did not . The transcript level for each GPCR was quantified from isolated larval brains with pan-neuronal knockdown of the GPCR in the two selected RNAi lines . In all cases RNA levels were reduced to approximately half of wild-type in RNAi strains that showed flight deficits , but not in cases where flight was maintained for normal periods ( Figure S3 ) . Thus , differential efficacy of RNAi strains appears to be responsible for the absence of flight deficits by multiple RNAi lines for a particular GPCR . Expression of itpr+ between 16 to 48 hours after puparium formation ( APF ) is sufficient for rescue of adult flight in itpr mutants , suggesting that a major role of IP3-mediated Ca2+ release in the flight circuit maybe during development [3] . Before testing if requirement for the identified GPCRs was during pupal development or in adult flight , we sought to characterize the time window ( 16–48 hr APF ) for itpr requirement more closely . For this purpose we used the TARGET ( temporal and regional gene expression targeting ) system [28] which includes a temperature sensitive GAL80 element ( GAL80ts ) that regulates GAL4 in a temperature dependent manner , with optimal repression and expression of GAL4 observed at 18°C and at 29°C respectively [29] . Experimental animals of the genotype ElavC155GAL4/+; dsitpr/+; GAL80ts/+ were shifted to the permissive temperature ( 29°C ) at specific time points after puparium formation ( APF ) . This allowed expression of the IP3R RNAi ( dsitpr ) and down-regulation of itpr transcripts from the time point of the temperature shift . Flies with a range of wing posture defects were observed upon pan-neuronal knockdown of the IP3R at 16 hours , 24 hours and 32 hours APF ( Figure 4A ) . Moreover , from the ratio of males and females obtained , there is an apparent lethality in males at the permissive temperature ( 29°C ) . The occurrence of a more severe defect in males as compared to females is very likely due to sex-specific differences in expression of the ElavC155GAL4 transgene , which is inserted on the X chromosome . Adults that emerged from these time points were quantified for the severity of wing posture defects ( Figure 4A ) . These were correlated with their ability to sustain flight ( Figure 4B ) . A strong correlation was observed between the ability to fly and the extent of wing posture defects in animals from all time points . Pan-neuronal knockdown of itpr starting at 16 hours APF lead to a complete loss of flight as evident from the single flight assay ( Figure 4B ) and air-puff induced flight patterns recorded from the DLMs ( Figure 4C ) . These animals also exhibited a significant level of spontaneous firing activity ( SPF ) from the DLMs , which is characteristic of itpr mutants ( Figure 4D and 4E; [3] ) . Animals with knockdowns at later stages showed a range of flight deficits that correlated well with their observed wing posture deficit and recordings from the DLMs ( Figure 4A–E ) , though SPF was high in all flies from the 16 hours and 24 hours APF time points , regardless of wing posture . When the temperature shift to 29°C was made 48 hours APF or later , neither wing posture nor flight deficits were observed ( Figure 4A–E; data not shown for 96 hours and 144 hours APF ) . IP3R is thus necessary from 16–32 hours APF for normal flight circuit development . Next , we investigated the temporal requirement for the identified GPCRs in the context of flight . These experiments demonstrated that pan-neuronal knockdown of either dFz-2R , mAcR or CCH1aR during pupal stages leads to flight deficits in adults , when tested in single flight assays ( Figure 5A ) . Similarly treated RNAi heterozygotes resulted in normal flight ( data not shown ) . The percentage of flight time was reduced upon pupal knockdown of dFz-2R to 53%±2 , mAcR to 66%±3 and CCH1aR to 72%±5 ( Figure 5A , colored bars within 29°C pupal , Movie S1 ) . Air puff stimulated responses recorded from DLMs were absent in a majority of non-fliers selected after the single flight assay , by pupal knockdown of dFz-2R ( 9/10 ) , mAcR ( 9/10 ) or CCH1aR ( 8/10; Figure 5B ) . Importantly , knockdown of dFz-2R , mAcR and CCH1aR during pupal stages resulted in flight deficits for each receptor that were similar to the deficits observed by knockdown throughout development ( shifted to 29°C post egg-laying ) ( Figure 5A and 5B ) , indicating that the requirement for all three GPCRs is primarily during flight circuit development . Similar experiments of SiFaR knockdown demonstrated a vital requirement during larval stages which lead to pupal lethality ( Figure 5A , green ) . However , knockdown of SiFaR during pupal stages did not affect flight duration indicating that this GPCR does not have a measurable role in either flight circuit development or in regulating flight in adults ( Figure 5A ) . Next , temporal requirements for the FmrfR and CG43795 were investigated using similar TARGET based experiments as described for the previous set of GPCRs in Figure 5 . Interestingly , knockdown of FmrfR either in adults or during pupal development resulted in flight deficits ( Figure 6A , red bars , Movie S2 ) . The extent of flight deficits by pan-neuronal knockdown of the FmrfR at the pupal stage was 45%±5 , while at the adult stage it was 60%±5 ( Figure 6A; red bars under 29°C pupal and 29°C adult ) . These deficits are comparable with post egg-laying ( PEL ) knockdown of the FmrfR , maintained all through development ( 47%±2; Figure 6A , red bar under 29°C post egg-laying ) . Similarly treated RNAi heterozygotes resulted in normal flight ( data not shown ) . Air puff stimulated responses obtained by electrophysiological recordings from the DLMs of the non-fliers selected after single flight assays were absent in 9/10 flies during pupal knockdown and in 9/10 flies during adult knockdown of the FmrfR ( Figure 6B ) . These deficits are comparable qualitatively and quantitatively with recordings from non-fliers obtained after down-regulation of FmrfR throughout development , where 8/10 animals exhibited rhythmic flight patterns for 5 sec or less ( Figure 6B , 29°C PEL ) . Thus the FmrfR receptor is required for modulation of flight both during pupal development and acute flight in adults . Unlike all other GPCRs identified in this screen , the requirement for CG43795 was only at the adult stage . Flies with adult knockdown of CG43795 showed reduced flight with a percentage flight time of 72%±2 ( Figure 6A , green bar below 29°C adult ) . Electrophysiological recordings from the DLMs of CG43795 knockdown non-fliers showed loss of flight patterns , upon air-puff stimulation , after 10–12 sec in 9/10 flies ( Figure 6B , 29°C adult ) . These flight deficits were comparable with the deficits observed upon knockdown of CG43795 throughout development ( Figure 6A , green bar in 29°C post egg-laying and 6B , 29°C PEL ) . Next , temporal requirement for the PdfR was assessed by similar TARGET based experiments . While the expression of dsPdfR during either larval or pupal stages had no significant effect on flight ( Figure 6C , blue bars within 29°C larval and 29°C pupal ) , its knockdown through both ( larval and pupal ) stages of development resulted in significant reduction in flight time ( 77%±2; Figure 6C , blue bar within 29°C larval+pupal ) and was accompanied by shorter periods of air-puff induced rhythmic action potentials recorded from the DLMs ( Figure 6D , 29°C larval+pupal ) . In addition , significant flight deficits and associated changes in flight physiology were observed upon PdfR knockdown in adults ( Figure 6C and 6D , blue bar and trace within 29°C adult ) . The flight deficit obtained by PdfR knockdown in larval and pupal development ( 77%±2 ) and by adult knockdown ( 71%±3 ) , together recapitulates the flight deficit observed when the PdfR RNAi was expressed throughout development ( shifted to 29°C post egg-laying; 54%±6; Figure 6C ) . These data suggest that signaling through the PdfR is required in separate neuronal subsets through development and in adults , and that both subsets contribute additively to the complete flight phenotype observed by PdfR knockdown through development and in adults . To identify PdfR expressing neurons which require Ca2+ release through the IP3R and SOCE for maintenance of flight , five independent GAL4 constructs that drive expression in PdfR neurons were tested [30] . These GAL4 constructs contain different regions of the PdfR regulatory domain and thus essentially drive expression in subsets of PdfR neurons [30] . The five GAL4s were used to knockdown either itpr , dSTIM or dOrai . Flies with knockdown of the IP3R using PdfR ( B ) GAL4 exhibited strong flight deficits ( 11%±3; Figure 7A ) and wing posture defects in 10% males ( data not shown ) . Moreover , air puff stimulated responses from the DLMs were found to be reduced and arrhythmic . In 8/16 animals , there was near complete loss of firing while 8/16 flies showed arrhythmic firing patterns ( Figure 7B and 7C , navy blue ) . Importantly , wing posture defects , flight defects and reduced response from the DLMs of PdfR ( B ) GAL4;dsitpr organisms could be rescued by introducing a genomic construct for the PdfR referred to as PdfR-myc ( Figure 7A–C; [30] ) . Flight deficits were also observed upon reduction of SOCE in PdfR ( B ) GAL4 expressing neurons either by knockdown of dSTIM ( 27%±3 ) or dOrai ( 18%±1; Figure 7A ) . Knockdown of dSTIM resulted in reduced firing from DLMs in 6/16 flies ( ∼5 sec ) and arrhythmic firing in 3/16 flies ( Figure 7B and 7C , light blue ) . Knockdown of dOrai , showed reduced firing in just 4/15 flies ( ∼15 sec; Figure 7B and 7C , green ) . Knockdown of dSTIM using PdfR ( B ) GAL4 also resulted in increased spontaneous firing from the DLMs ( data not shown ) ; this phenotype is characteristic of itpr mutants [3] . Importantly , all flight phenotypes including reduced electrophysiological responses from DLMs and the high spontaneous firing observed in dSTIM knockdown flies could be rescued to normal levels by over-expression of an inducible PdfR cDNA ( UAS-PdfR16L; Figure 7A–C ) . Next , we investigated the requirement for the PdfR directly in the PdfR ( B ) GAL4 expressing neurons in the context of flight . Knockdown of PdfR ( dsPdfR ) in neurons marked by the PdfR ( B ) GAL4 resulted in significant reduction in flight time ( 63%±0 . 8; Figure 7A ) and in firing responses from the DLMs in 10/15 flies ( ∼5 sec; Figure 7B and 7C ) . Further , we investigated if mutants in the cognate ligand for the PdfR , the “Pigment Dispersal Factor” ( pdf ) affected flight . We tested adults for the null allele , pdf01 for flight [31] . Homozygous pdf01 showed reduced flight time ( 78%±1; Figure 7A ) and reduced firing from DLMs in 5/15 randomly selected flies ( Figure 7B and 7C ) . However , the flight defects observed either by knockdown of PdfR or in pdf mutant flies , were not equivalent to the deficits observed by knockdown of IP3R using PdfR ( B ) GAL4 ( Figure 7A ) . These data suggest that whereas PDF activates the PdfR in the PdfR ( B ) GAL4 expressing neurons , there are probably additional roles for the IP3R in PdfR ( B ) GAL4 expressing neurons in the context of flight . Furthermore , the flight deficits observed in PDF mutant flies ( Figure 7A ) were considerably less than flight deficits observed by knockdown of PdfR using ElavC155GAL4 ( Figure 3A ) , suggesting the existence of another flight-regulating ligand acting through the PdfR . Knockdown of the IP3 receptor , dSTIM or dOrai using an independent transgenic line , the PdfR ( A ) GAL4 [30] , had no effect on normal wing posture ( data not shown ) or flight ( Figure 7A ) . However , expression of dSTIM and IP3R was reduced significantly in adult brain and thoracic ganglia upon knockdown by RNAi using both the PdfR ( B ) GAL4 and PdfR ( A ) GAL4 ( Figure 7D ) . These data suggest that the PdfR regulates flight through IP3R mediated Ca2+ signaling exclusively in the neurons marked by the PdfR ( B ) GAL4 and not the PdfR ( A ) GAL4 . Thus , to identify neuronal regions that require PdfR mediated Ca2+ signaling for regulation of flight , we compared neurons marked by expression of the PdfR ( B ) GAL4 and PdfR ( A ) GAL4 . For this purpose cells in both GAL4 strains were marked with a cytosolic form of GFP . The overall expression level of GFP in adult brains and ventral ganglia were similar in both the GAL4 strains ( Figure S4C ) . Expression patterns of each GAL4 line were visualized in the larval brain , the adult brain and the thoracic ganglion ( Figure S4A , S4B , S4E and S4F ) . Expression patterns were analyzed by searching for regions with GFP expression in PdfR ( B ) GAL4 and the absence of expression in these regions in PdfR ( A ) GAL4 . Strong GFP immunoreactivity was observed in neuronal cell bodies located near the sub-esophageal ganglion ( SOG ) , in the thoracic ganglion and the antennal mechanosensory and motor complex ( AMMC ) ( Figure S4D ) in PdfR ( B ) GAL4 ( Figure 8A , 8C , 8E , 8G and 8H ) . Expression in these regions was reduced in PdfR ( A ) GAL4 ( Figure 8B , 8D , 8F , 8I and 8J ) . Expression was also seen in other regions of the brain including the medial neurosecretary cells ( mNSCs ) , where PdfR ( B ) GAL4 and PdfR ( A ) GAL4 expressed to equivalent levels ( Figure 8K and 8L ) . A summary of the complete expression patterns of both the GAL4 lines is shown in Figure 8M . The expression analysis suggests that PdfR function in neurons of the AMMC , SOG and thoracic ganglion regulates the maintenance of flight in Drosophila .
In a genetic RNAi screen for GPCRs that regulate flight , twenty-two genes were identified amongst which eight encoded neuropeptide receptors and seven were for neurotransmitter receptors , highlighting the importance of these ligands for neuro-modulation of motor function ( Figure 9 ) . The remaining genes encoded various receptor classes with possible roles in development like the frizzled-2 receptor and the methuselah-like receptors ( 3/22 ) , putative sensory receptors ( rhodopsin-like and trehalose-sensing ) and CG43795 with no clear homology to any class of GPCRs . Despite testing multiple RNAi lines for each receptor , our screening strategy may have missed out some flight regulating GPCRs . This would be true specifically in cases where RNAi lines tested for a particular gene were not efficacious , if the pan-neuronal GAL4 strain utilized in the screen expressed weakly in the cognate neurons and due to inappropriate temporal expression of the GAL4 with respect to the temporal requirement for that receptor . In a secondary modifier screen designed to test if the signaling mechanism activated by the identified receptors was indeed intracellular Ca2+ release and store-operated Ca2+ entry , flight deficits in three out of the twenty two receptors identified ( CG34411 , CG3171 and CG43795 ) were further enhanced by expression of either AcGq or dSTIM+ suggesting that these receptors could constitute an inhibitory signaling component of the flight circuit . Inhibitory neural circuits within central pattern generators constitute an integral part of any rhythmic motor behavior [2] . When analyzed by us , the predicted protein sequence of CG43795 exhibits highest homology with the predicted sequence of CG31760 ( E = 1 . 262e-66 ) , which in turn is classified as a putative glutamate/GABA receptor . Similar to vertebrates , GABA functions as an inhibitory neurotransmitter in Drosophila [32] . The role of CG43795 , CG34411 and CG3171 as putative components of inhibitory signaling during acute flight in adults requires further study . Amongst the five receptors identified in the secondary suppressor screen , two have been linked with IP3-mediated Ca2+ release previously . The mAcR can stimulate the IP3R in transfected S2 cells [17] , [18] , [33] , [34] and by over-expression in primary neuronal cultures [4] . Similarly , the FmrfR was shown to modulate intracellular Ca2+ in type 1 nerve terminals and thus regulate light-dependant escape behavior in Drosophila larvae [35] . However , a physiological role for these receptors in the regulation of flight in adult Drosophila has not been described earlier . The temporal analysis showed a dual requirement for the FmrfR during pupal stages and in adults , which were non-additive , suggesting that the same set of neurons require FmrfR function during development and for modulating acute flight in adults . The precise neurons that require FmrfR function for maintenance of flight and the role of IP3/Ca2+ signaling in them , needs further analysis . Another recently de-orphanised receptor identified in the final screen was CCH1aR with the specific ligand , CCH1amide [36] . The ligand is found in the Drosophila mid-gut and central nervous system [37] . However , physiological functions have not been attributed to the CCH1aR so far . From the differential effect on flight obtained by knockdown of the CCH1aR , mAcR , FmrfR and PdfR as well as their differential temporal requirement , it seems likely that each receptor regulates independent aspects of either flight circuit development , function or both . This hypothesis needs further confirmation by genetic and anatomical studies for each receptor . Neuroanatomical studies for spatial localization of the identified receptors in the context of flight circuit components are required , as has been attempted here for the PdfR . From previous work we know that synaptic function of the well-characterized Giant Fibre Pathway , required for the escape response , is normal in IP3R mutants [3] , [38] , [39] . Instead , intracellular calcium signaling is required for development of the air-puff stimulated flight circuit , which is a laboratory paradigm of voluntary flight . Spontaneous calcium transients through voltage gated Ca2+ channels can affect dendritic morphology and neurotransmitter specification in developing neural circuits [40] , [41] . The developmental processes that require intracellular calcium signaling during flight circuit maturation may be similar but are not understood so far . In part , a reason for this lack of understanding is the absence of well characterized interneurons that integrate and communicate sensory information to the flight motor pathways in the context of voluntary flight . Thus , spatial localization of GPCRs found in this screen and neural connections of flight GPCR expressing cells will in future help understand and identify both neural components of the voluntary flight circuit and the role of intracellular calcium signaling in flight circuit maturation and function . The screen also identified the SiFaR as a neuronal receptor required for viability . However , since pupal lethality in SiFaR knockdown was not suppressed by expression of AcGq , the downstream signaling mechanism of this receptor remains unclear . A recent study in the Blacklegged Tick has implicated SiFaR in the regulation of feeding [42] . It is therefore possible that , lethality in SiFaR knockdown animals is a consequence of reduced feeding at the larval stages . Analysis of flight phenotypes exhibited by knockdown of dFz-2R suggests a requirement for this receptor during flight circuit development . Suppression of flight deficits in dsFz-2R expressing flies by over expression of dSTIM+ , but not AcGq indicates that this receptor does not activate the canonical GPCR/IP3/Ca2+ signaling mechanism . From previous studies , it is known that dFz-2R can signal through Wnt/βcatenin pathway [43] , [44] while recent speculations implicate a non-canonical Wnt-Ca2+ pathway as downstream of Fz-2R [45] , [46] already shown for rat ( rFz-2R ) and Xenopus ( XFz-2R ) [47] . In Drosophila , dFz-2R is thought to act via the G-protein Gαo [48] . Suppression of dsFz-2R flight deficits by dSTIM+ implicates intracellular Ca2+ signaling as downstream of dFz-2R activation for the first time in Drosophila . While the cellular correlates of dFz-2R activation need to be demonstrated directly in Drosophila flight circuit neurons , it is likely that this study will help identify other molecular components of this pathway . Interestingly , we discovered a regulatory role for the PdfR in Drosophila flight where our genetic data implicate IP3-mediated Ca2+ release as the downstream signaling mechanism . Although , PdfR stimulation increases cAMP levels in HEK293 cells transfected with Drosophila PdfR , it is also known that cellular Ca2+ levels increase moderately in response to PDF [22] . Our findings suggest that the PdfR is capable of stimulating dual G-proteins , similar to 5HT-dro2A and 5HT-dro2B , the cellular responses of which include a decrease in cAMP as well as an increase in inositol phosphates in response to serotonin [49] , [50] . The role of the neuropeptide ligand , PDF and PdfR in regulation of circadian rhythms is well documented in adults [22] , [51] , [52] and more recently during early larval development for instructing circadian circuit formation in pupae [53] . In addition PDF function in circadian neurons regulates several other processes like reproduction , arousal and geotaxis [22] , [54] , [55] . Recently , the PdfR orthologue in C . elegans was found to modulate locomotory behavior [56] . Our data support earlier published data in Drosophila suggesting that PdfR can be activated by ligands other than PDF , such as vertebrate PACAP ( pituitary adenylate cyclase activating polypeptide ) [22] . In vertebrates , signaling by PACAP modulates locomotor activity and the exploratory behavior of rats , mice , chicken and goldfish [57] . Our study shows that PdfR ( B ) GAL4 expressing neurons which rescue circadian rhythm phenotypes of PdfR mutants ( PdfR3369 , PdfR5304; [30] ) also function in flight regulation . The source of PDF and/or another ligand that activates PdfR signaling in the context of flight remains to be determined . PDF is secreted from two known sources in Drosophila; one is the lateral ventral protocerebrum ( LNvs ) and the other is neurons in the abdominal ganglion ( AbNs; [58]–[60] ) . A recent study revealed an endocrine mode of action of PDF for the regulation of ureter contractions [61] . A better understanding of the neurocircuitry underlying voluntary flight is required to distinguish between these two sources of PDF for development and function of the flight circuit , as well as to investigate whether endocrine mechanisms deliver the ligand ( s ) for activating the PdfR in the context of flight . This study adds to the growing body of evidence which suggests that signaling through PdfR could serve as a global integrator of a repertoire of behaviors important for Drosophila survival in the wild .
Drosophila strains were reared on corn flour/agar media supplemented with yeast , grown at 25°C , unless otherwise mentioned in the experimental design . The wild-type Drosophila strain used was Canton-S ( CS ) . The pan-neuronal GAL4 driver used was ElavC155GAL4 obtained from Bloomington Stock Center , Bloomington , IN . UAS-PdfR16L , Pacman PdfR-myc 70 , PdfR ( B ) GAL4 ( 2 ) and PdfR ( A ) GAL4 ( 2 ) were obtained from Paul Taghert ( Washington University , St . Louis ) [30] . G-protein coupled receptor UASRNAi lines were obtained from Vienna Drosophila RNAi center , Vienna , Austria ( VDRC ) and National Institute of Genetics Fly Stocks Centre , Kyoto , Japan ( NIG ) . The UAS-RNAi strains for dSTIM ( dsdSTIM , 47073 ) and dOrai ( dsdOrai , 12221 ) were obtained from VDRC and for itpr ( dsitpr , 1063-R2 ) from NIG [4] . The other strains used were as follows: UASdSTIM+ [24] , UASAcGq3 [27] , GAL80ts with two inserts on second chromosome ( generated by Albert Chiang , NCBS , Bangalore , India ) . UASdicer ( X ) , used in combination with dsdSTIM and dsdOrai; UASdicer ( III ) , used in combination with dsitpr and UASmCD8GFP ( II ) were obtained from BDSC . The other fly strains used were generated using standard Drosophila genetic methods . Females of the ElavC155GAL4 strain were mated with males of each RNAi strain . In the resulting progeny , male flies gave varied responses ( data not shown ) to air-puff induced flight . Therefore only adult female flies were used further for analysis . Adult females were collected soon after eclosion and aged for 3–4 days before testing for flight . Flies were anaesthetized on ice for 15 min and a thin metal wire was glued between the neck and thorax region with the help of nail polish . To test for air-puff responses , videos were recorded for 30 sec after giving a gentle mouth-blown air puff stimulus to the tethered fly . These videos were analyzed and percentage flight time was calculated . For each RNAi line , 10 flies were tethered and tested along with 10 control flies . Physiological recordings were obtained from the indirect Dorsal Longitudinal Muscles ( DLMs ) as described previously [3] . Briefly , an un-insulated 0 . 127 mm tungsten electrode , sharpened by electrolysis to attain 0 . 5 µm tip diameter , was inserted in the DLM ( fiber a ) . A similar electrode was inserted in the abdomen for reference . Spontaneous firing was recorded for 2 min and air-puff stimulated recordings were done for 30 s . All recordings were done using an ISO-DAM8A amplifier ( World Precision Instruments , Sarasota , FL ) with filter set up of 30 Hz ( low pass ) to 10 kHz ( high pass ) . Gap free mode of pClamp8 ( Molecular Devices , Union City , CA ) was used to digitize the data ( 10 kHz ) on a Pentium 5 computer equipped with Digidata 1322A ( Molecular Devices ) . Data were analyzed using Clampfit ( Molecular Devices ) and the mean and standard error ( SEM ) were plotted using Origin 7 . 5 software ( MicroCal , Origin Lab , Northampton , MA , USA ) . For isolation of RNA , the central nervous system ( CNS ) was dissected from 3rd instar wandering larvae . Each sample of RNA was extracted from five CNSs and three independent preparations were analyzed for each experiment . Total RNA was isolated using TRIzol Reagent ( Invitrogen Life Technologies , Carlsbad , CA , USA ) according to the manufacturer's specifications . Integrity of RNA was confirmed by visualization on a 1% TAE ( 40 mM Tris pH 8 . 2 , 40 mM acetate , 1 mM EDTA ) agarose gel . Total RNA ( 500 ng ) was treated with DNase in a volume of 45 . 5 µl with 1 µl ( 1 U ) DNase I ( Amplification grade , Invitrogen Life Technologies , Carlsbad , CA , USA ) with 1 mM dithiothreitol ( DTT ) ( Invitrogen Life Technologies , Carlsbad , CA , USA ) , 40 U of RNase Inhibitor ( Promega , Madison , WI , USA ) in 5× First Strand Buffer ( Invitrogen Life Technologies , Carlsbad , CA , USA ) for 30 min at 37°C and heat inactivated for 10 min at 70°C . The reverse transcription reaction was performed in a final volume of 50 µl by addition of 1 µl ( 200 U ) Moloney murine leukemia virus ( M-MLV ) reverse transcriptase ( Invitrogen Life Technologies , Carlsbad , CA , USA ) , 2 . 5 µl ( 500 ng ) random hexaprimers ( MBI Fermentas , Glen Burnie , MD , USA ) and 1 µl of a 25 mM dNTP mix ( GE Healthcare , Buckinghamshire , UK ) . Samples were incubated for 10 min at 25°C , then 60 min at 42°C and heat inactivated for 10 min at 70°C . The polymerase chain reactions ( PCRs ) were performed using 1 µl of cDNA as a template in a 25 µl reaction under appropriate conditions . Real time quantitative PCR ( qPCR ) were performed on an ABI 7500 Fast machine ( Applied Biosystems , Foster City , California , USA ) operated with ABI 7500 software version 2 ( Applied Biosystems , Foster City , California , USA ) using MESA GREEN qPCR MasterMIx Plus for SYBR Assay I dTTp ( Eurogentec , Belgium ) . qPCRs were performed with rp49 primers as internal controls and primers specific to gene of interest using dilutions of 1∶10 . Sequences of the primers used in the 5′ to 3′ directions are given below . The sequence of the forward primer is given first in each case: rp49 CGGATCGATATGCTAAGCTGT; GCGCTTGTTCGATCCGTA , Fz-2R GGTTACGGAGTGCCAGTCAT; CACAGGAAGAACTTGAGGTCC , mAcR CAAGGACGAGTGCTACATCC; CCTAAATCAGAAGGCTCCTCC , CCH1aR GACCAAAGGAATGGCGTAGTAG; CGCTCGCATCCACAGTTTAC , PdfR CAAATGCCACGGAGGTGAATC; TCAGCAGGGAAACTATAAGGGC , FmrfR GTGCGAAAGTTACCCGTCG; TAATCGTAGTCCGTGGGCG , SiFaR CAATCAGTGTGGCTGGCAG; CCTACATCGTCGTCTTCCTG . Each qPCR experiment was repeated three times with independently isolated RNA samples . The cycling parameters were 95°C for 5 min , followed by 40 cycles of 95°C for 15 s and 60°C for 1 min . The fluorescent signal produced from the amplicon was acquired at the end of each polymerization step at 60°C . A melt curve was performed after the assay to check for specificity of the reaction . The fold change of gene expression in the mutant relative to wild-type was determined by the comparative ΔΔCt method [62] . In this method the fold change = 2−ΔΔCt where ΔΔCt = ( Ct ( target gene ) −Ct ( rp49 ) ) mutant2− ( Ct ( target gene ) −Ct ( rp49 ) ) Wild type . Immunohistochemistry was performed on Drosophila adult brains expressing cytosolic GFP ( UASGFP ) with the specified GAL4 strains , after fixing the dissected tissue in 4% paraformaldehyde . The following primary antibodies were used: mouse monoclonal nc82 antibody ( 1∶20 , kindly provided by Eric Buchner ) , rabbit anti-GFP antibody ( 1∶10 , 000; #A6455 , Molecular Probes , Eugene , OR , USA ) . Fluorescent secondary antibodies were used at a dilution of 1∶400 as follows: anti-rabbit Alexa Fluor 488 ( #A1108 ) and anti-mouse Alexa Fluor 568 ( #A1104 , Molecular Probes , Eugene , OR , USA ) . Confocal analysis was performed on an Olympus Confocal FV1000 microscope . Confocal data were acquired as image stacks of separate channels and combined and visualized as three-dimensional projections using the FV10-ASW 1 . 3 viewer ( Olympus Corporation , Tokyo , Japan ) . Adult brains and thoracic ganglia were dissected from 3 to 5 day old progeny of the indicated genotypes . Protein extracts were made by homogenizing the sample in homogenizing buffer ( 40 mM Tris pH 7 . 4 , 1 mM EDTA , 1 mM EGTA , 0 . 05% Triton X-100 ) and were separated on a 6% SDS-polyacrylamide gel and transferred to nitrocellulose membrane by standard western blotting protocols . The affinity purified anti-InsP3R rabbit polyclonal antibody ( IB-9075; [34] ) was used at a dilution of 1∶300 . A mouse anti-spectrin antibody ( 3A9 ) ( 1∶50 dilution , Developmental Studies Hybridoma Bank , University of Iowa , Iowa ) was used as a loading control for the InsP3R . Two anti-dSTIM mouse antibodies ( 8G1 ) and ( 3C1 ) mixed 1∶1 ( Generated by Bioneeds , Bangalore , India ) were used at a dilution of 1∶200 . The mouse anti-GFP monoclonal antibody ( sc-9996 , Santa Cruz Biotechnology , CA ) was used at a dilution of 1∶1000 . The mouse anti-β-tubulin monoclonal antibody ( E7 , Developmental Studies Hybridoma Bank , University of Iowa , Iowa ) was used at a dilution of 1∶200 as a loading control for dSTIM and GFP . Secondary antibodies conjugated to horseradish peroxidase were used , and protein was detected in the blot by addition of a chemiluminescent substrate from Thermo Scientific ( No . 34075; Rockford , IL , USA ) .
|
A majority of behavioral patterns in flying insects depend upon their ability to modulate flight . In Drosophila melanogaster , mutations in the IP3 receptor gene lead to loss of voluntary flight in response to a natural stimulus like a gentle air-puff . From previous genetic and cellular studies it is known that the IP3R in Drosophila is activated by G-protein coupled receptors ( GPCRs ) . However , GPCRs that act upstream of the IP3R in the context of flight are not known . Therefore , we performed a genetic RNAi screen to identify GPCRs which regulate flight . This screen was followed by a secondary suppressor screen that assessed the role of each identified GPCR in activating IP3/Ca2+ signaling . We found 5 such GPCRs . Our results demonstrate that these GPCRs are required during flight circuit development and during adult flight . One flight-regulating receptor identified was the Pigment Dispersing Factor Receptor ( PdfR ) . This receptor is known to regulate behaviors such as circadian rhythms , geotaxis and reproduction . A spatio-temporal analysis of PdfR flight function indicates that it regulates both flight circuit development and acute flight through multiple neurons . We postulate that PdfR signaling could modulate and integrate multiple behavioral inputs in Drosophila and other flying insects .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
A Genetic RNAi Screen for IP3/Ca2+ Coupled GPCRs in Drosophila Identifies the PdfR as a Regulator of Insect Flight
|
Impaired mitochondrial oxidative phosphorylation ( OXPHOS ) has been proposed as an etiological mechanism underlying insulin resistance . However , the initiating organ of OXPHOS dysfunction during the development of systemic insulin resistance has yet to be identified . To determine whether adipose OXPHOS deficiency plays an etiological role in systemic insulin resistance , the metabolic phenotype of mice with OXPHOS–deficient adipose tissue was examined . Crif1 is a protein required for the intramitochondrial production of mtDNA–encoded OXPHOS subunits; therefore , Crif1 haploinsufficient deficiency in mice results in a mild , but specific , failure of OXPHOS capacity in vivo . Although adipose-specific Crif1-haploinsufficient mice showed normal growth and development , they became insulin-resistant . Crif1-silenced adipocytes showed higher expression of chemokines , the expression of which is dependent upon stress kinases and antioxidant . Accordingly , examination of adipose tissue from Crif1-haploinsufficient mice revealed increased secretion of MCP1 and TNFα , as well as marked infiltration by macrophages . These findings indicate that the OXPHOS status of adipose tissue determines its metabolic and inflammatory responses , and may cause systemic inflammation and insulin resistance .
White adipose tissue ( WAT ) determines whole-body energy metabolism by controlling lipid storage and by releasing adipokines , which may directly or indirectly affect the physiological functions of almost all cell types ( for a review , see [1] , [2] ) . These adipocyte functions are perturbed by genetic and environmental factors , which lead to adipocyte dysfunction characterized by hypertrophy , hypoxia and inflammatory process within adipose tissue [3] . Adipocyte dysfunction is further characterized by impaired insulin sensitivity , which is associated with changes in cellular composition or organelle dysfunction , particularly of the endoplasmic reticulum ( ER ) and mitochondria . An emerging concept to explain insulin resistance in obese individuals is maladaptive responses within the ER , which are prominent in adipose tissue ( for a review , see [4]–[6] ) . Besides the ER , the mitochondria in white adipocytes are linked with adipocyte differentiation and with the function of mature adipocytes . Recent studies show that drastic increases in mitochondrial biogenesis and reactive oxygen species ( ROS ) production via the OXPHOS complex play a crucial role in adipocyte differentiation . In addition , the mitochondria in differentiating adipocytes support high energy-consuming lipogenic processes to maintain mature adipocyte function [5] , [7] . Therefore , it is suggested that the contribution of adipocyte mitochondria to whole-body energy metabolism or adipocyte plasticity may depend on the mitochondrial OXPHOS capacity of the adipose tissue [6] . Consistent with this , decreased mitochondrial capacity in adipocytes may also alter their insulin sensitivity and/or function due to the high energy requirements of fatty acid storage , adipokine secretion , insulin signaling , and glucose uptake [8] , [9] . It is interesting that a marked decrease in the level of transcripts for nuclear-encoded mitochondrial genes in cells derived from the epididymal fat pads of ob/ob mice accompanies the onset of obesity [10] . In db/db and diet-induced obese mice , the expression of OXPHOS genes was markedly reduced compared with that in db/+ mice and control mice fed a standard-fat diet , respectively [11] . In humans , the mtDNA copy number is enriched in adipocytes in adipose tissue , but it decreases slightly with age and increasing BMI , and shows a strong positive correlation with basal and insulin-stimulated lipogenesis in fat cells [12] . More interestingly , suppression of OXPHOS genes is prominent in the visceral adipose tissue of humans with type 2 diabetes independent of obesity [13] . Agonists of peroxisome proliferator-activated receptor-gamma ( PPARγ ) increase the number of mitochondria and induce mitochondrial remodeling in adipocytes [10] , [11] , [14] , and significantly increase the mitochondrial copy number and expression of factors involved in mitochondrial biogenesis , including PPARγ coactivator-1alpha ( PGC1α ) and mitochondrial transcription factor A ( TFAM ) , which are required for mitochondrial transcription of OXPHOS genes in humans [15] . These observations in rodent models and human subjects suggest that the OXPHOS capacity of adipose tissue may affect the changes in adipocyte plasticity , which controls insulin sensitivity and may determine the therapeutic responsiveness to antidiabetic agents such as thiazolidinediones and CB1 receptor blockers that affect the mitochondrial content of adipocytes [10] , [16] . Here , we demonstrate that primary OXPHOS dysfunction in adipose tissue causes insulin resistance and a diabetic phenotype in mice with a Crif1 loss-of-function mutation . Crif1 is a mitochondrial protein that associates with large mitoribosomal subunits , which are located close to the polypeptide exit tunnel , and the elimination of Crif1 led to both aberrant synthesis and defective insertion of mtDNA-encoded nascent OXPHOS polypeptides into the inner membrane [17] . Targeted elimination of the Crif1 gene resulted in a phenotype characterized by organ-specific failure of OXPHOS function; therefore , we attempted to identify the adipose tissue phenotypes of adipose-specific Crif1-knockout mice using Fabp4-Cre and Adiponectin-Cre mice models . Reduced OXPHOS capacity in the WAT of Crif1-deficient mice triggered spontaneous adipose inflammation , which was characterized by macrophage infiltration and systemic insulin resistance . Therefore , the OXPHOS reserve may be the critical determinant controlling the metabolic and inflammatory responses of adipose tissue , which are closely related to systemic changes in insulin sensitivity .
Crif1 is a mitochondrial protein that specifically interacts with the protein components of the large subunit of the mitochondrial ribosome [17] . It specifically regulates the translation and insertion of the 13 polypeptide subunits that comprise mitochondrial OXPHOS complexes I , III , IV and V . Homozygous Crif1-null mouse embryonic fibroblasts ( MEFs ) showed a profound failure in translation and expression of these subunits , along with markedly low levels of basal and stimulated ( CCCP-treated ) mitochondrial oxygen consumption [17] . Disruption of the mouse Crif1 gene consistently resulted in a profound OXPHOS deficiency characterized by the loss of OXPHOS complex subunits and respiratory complexes in vivo . Crif1 mRNA is ubiquitously expressed , and it is highly expressed in brain , heart , liver kidney and skeletal muscle ( Figure S1A ) . Two types of adipose tissues , brown ( BAT ) and white ( WAT ) , contained substantial amounts of Crif1 mRNA ( Figure S1A ) . Crif1 mRNA levels were decreased in the WAT , BAT and liver of db/db and ob/ob mice compared to db/+ and ob/+ mice , respectively ( Figure S1B ) . Interestingly , Crif1 mRNA expression in WAT of C57BL/6 mice was downregulated when they were fed a high fat diet ( HFD ) for 8 weeks ( Figure S1C ) . These findings indicate that Crif1 expression correlates with the nutritional status in adipose tissue . To identify the roles of Crif1 and mitochondrial OXPHOS in adipose tissue , we tried to induce primary OXPHOS deficiency in adipose tissue in vivo using conditional Crif1 knockout mice . We crossed conditional Crif1 mice ( Crif1flox/flox ) [18] with mice expressing a Cre recombinase gene under the control of the fatty acid binding protein-4 ( Fabp4 ) promoter ( Fabp4-Cre ) and the adiponectin promoter ( Adipoq-Cre ) . The resulting pups were born healthy and viable , and showed a normal Mendelian ratio . However , these homozygous Crif1f/f , Fabp4 mice showed delayed weight gain and poor development of adipose tissue ( Figure 1A–1C ) . Unlike the control ( Crif1+/+ , Fabp4 ) and Crif1 heterozygous ( Crif1f/+Fabp4 ) mice , Crif1f/f , Fabp4 mice showed uniform lethality within 24 days of birth ( median survival = 19 . 4 days ) ( Figure 1D ) . The perirenal , subcutaneous and epididymal fat pads of Crif1f/f , Fabp4 mice comprised small adipocytes with dystrophic changes ( Figure 1E ) . To verify any mitochondrial abnormalities , the adipose tissues of Crif1f/f , Fabp4 mice were examined by transmission electron microscopy ( TEM ) . The adipocytes of these mice contained mitochondria with ultrastructural abnormalities , such as swollen and distorted cristae , but mitochondrial number was unaffected ( Figure 1F and 1G ) . In heterozygous Crif1f/+ , Fabp4 mice , hematoxylin and eosin ( H&E ) staining of adipose tissue showed no evidence of histological abnormalities compared with the controls ( Figure 1E ) . Consistent with the results of H&E staining , the mitochondria of Crif1f/+ , Fabp4 mice showed no morphological or numerical abnormalities of mitochondria in TEM ( Figure 1F and 1G ) . Collectively , this comprehensive analysis of the adipose tissues in Crif1f/f , Fabp4 mice indicated that loss of Crif1 results in a marked failure of WAT and BAT development . The Fabp4-Cre transgene is expressed and localized within the dorsal root ganglion , centrum of the vertebra and the carpals of the embryo from the mid-gestation stage [19] . Neonatal Crif1f/+ , Fabp4 or Crif1f/f , Fabp4 mice did not show developmental abnormalities when compared with control mice . Therefore , embryonic expression of the Fabp4-Cre transgene may not affect the development of Crif1f/+ , Fabp4 and Crif1f/f , Fabp4 mice , and may not be a plausible reason for observed lethality at around post-natal Week 3 . Mice are normally weaned at post-natal Week 3 , at which point the rate of lipogenesis and UCP1 expression in the BAT rises sharply and reaches maximal levels to enhance thermogenesis [20] . The Fabp4-Cre transgene was uniformly detected in BAT from the early post-natal period ( Day 7 ) , the Crif1 protein and OXPHOS complex subunits are downregulated in the BAT of 3-week-old mice ( Table S1 ) . As shown in Figure 1C , Crif1f/f , Fabp4 mice had less BAT at Day 21 , but histological examination of inter-scapular BAT showed normal histological findings ( Figure S2A ) . Crif1f/f , Fabp4 mice had fewer mitochondria than control mice , but these were larger in size and were characteristically disorganized and swollen , suggesting OXPHOS defects ( Figure S2B and S2C ) . Consistent with these findings , Crif1f/f , Fabp4 mice showed a low body temperature under ambient conditions ( 23°C ) and rapidly reached a fatally low rectal temperature within 5 minutes of immersion in cold water ( 4°C ) ( Figure S2D ) . However , although the mass of BAT was reduced , the level of UCP1 expression was not altered in Crif1f/f , Fabp4 mice ( data not shown ) . When Crif1f/f , Fabp4 mice were housed at thermoneutrality ( 30°C ) , the median survival rate was increased and mortality was reduced ( Figure S2E ) . This indicates that thermal stress caused by mitochondrial OXPHOS dysfunction in BAT following Crif1 ablation may be a critical factor in the premature death of Crif1f/f , Fabp4 mice . By contrast , the BAT of Crif1f/+ , Fabp4 mice showed normal development and histological and ultrastructural findings ( Figure S2A–S2C ) . Furthermore , the response of Crif1f/+ , Fabp4 mice ( in terms of core temperature ) to a cold environment were identical to those of control mice ( Figure S2D ) . These results showed that Fabp4-Cre driven haploinsufficiency of Crif1 may not affect the physiological function of BAT . A previous study revealed that Crif1-deficient ( −/Δ ) MEFs prepared from Crif1−/flox mice showed marked OXPHOS defects due to a profound failure of translation and insertion of the newly-synthesized OXPHOS polypeptides encoded by the mtDNA . Also , Crif1 −/Δ MEFs showed increased anaerobic glycolysis , which eventually led to accelerated cell death [17] . Similar to Crif1 −/Δ MEFs , loss of Crif1 in adipose-derived stem cells ( ADSCs ) ( Crif1M−/− ) resulted in marked impairment of differentiation and accelerated cell death , which prevented functional analysis of the mitochondria ( data not shown ) . However , control ( Crif1+/+ ) and Crif1-haploinsufficient ADSCs ( Crif1+/− ) prepared from Crif1+/+ , Fabp4 and Crif1f/+ , Fabp4 mice showed identical levels of cell viability and differentiation to those of control cells ( Figure 2A and 2B ) . Interestingly , Crif1+/− ADSCs showed lower expression of OXPHOS subunits ( ND1 , NDUFA9 , UQCRC2 and COX4 ) and assembled OXPHOS complex I on Western blot and Blue Native PAGE ( BN-PAGE ) analysis , respectively ( Figure 2C and 2D ) . Crif1+/− ADSCs consumed less oxygen under basal conditions and showed reduced maximal OXPHOS capacity ( Figure 2E ) . Taken together , Crif1 haploinsufficiency in adipocytes resulted in normal differentiation but reduced genetically-determined OXPHOS capacity . Several experimental criteria have been proposed to test whether a primary in vivo OXPHOS deficiency plays a causal role in insulin resistance [21] . One of these criteria is that perturbations in OXPHOS gene expression and function must be as modest as possible [21] . Thus , we analyzed Crif1 and OXPHOS gene expression to test whether Crif1f/+ , Fabp4 mice were suitable for our proposed experiments . Compared with Crif1+/+ , Fabp4 mice , Crif1f/+ , Fabp4 mice showed about ∼50% of the Crif1 mRNA and protein expression in epididymal WAT ( eWAT ) ( Figure 2F and 2H ) . Although basal ATP levels in eWAT were not affected by Crif1 haploinsufficiency ( Figure 2G ) , the expression levels of OXPHOS complex I , III and IV subunits were reduced in the epididymal fat pads of Crif1f/+ , Fabp4 mice ( Figure 2H ) . BN-PAGE analysis showed that the levels of Complex I and IV and supercomplex in WAT were approximately 20% , 40% and 50% lower , respectively , in Crif1f/+ , Fabp4 mice compared to control mice ( Figure 2I and 2J ) . However , normal levels of Crif1 and OXPHOS complexes were expressed in the liver and heart of Crif1f/+ , Fabp4 mice ( Figure S3A–S3C ) . In contrast to homozygous Crif1f/f , Fabp4 mice , heterozygous Crif1f/+ , Fabp4 mice exhibited normal levels of OXPHOS subunits in BAT and mitochondrial morphology was normal ( Figure 2H–2J and Figure S2B ) . Food intake and weight gain were comparable in Crif1f/+ , Fabp4 and Crif1+/+ , Fabp4 mice when fed a normal chow diet ( NCD ) ( Figure S4A and S4B ) . MR images of control and Crif1f/+ , Fabp4 mice fed a NCD or a HFD showed a similar pattern of adipose distribution ( Figure S4C ) . Triglyceride levels in the liver and plasma of Crif1f/+ , Fabp4 mice were the same as those in control mice , regardless of whether they were fed a NCD or a HFD . Serum free fatty acid ( FFA ) levels tended to be higher in Crif1f/+ , Fabp4 mice , but were not significantly different from those in control mice ( Figure S4D–S4F ) . Taken together , these results show that Crif1f/+ , Fabp4 mice have mildly reduced primary OXPHOS deficiency in adipose tissue but , unlike the lipodystrophic model , they show no defects in adipose tissue development , and no hyperlipidemia or ectopic lipid accumulation . To identify the relationship between insulin resistance and reduced OXPHOS capacity in adipocytes in vivo , control and Crif1f/+ , Fabp4 mice were subjected to glucose tolerance tests after 8 weeks or 14 weeks on a NCD or HFD . Neither control nor Crif1f/+ , Fabp4 mice fed a NCD for 8 weeks showed any differences in glucose tolerance following an intraperitoneal injection of glucose ( IPGTT , 2 g/kg body weight ) ( Figure 3A ) . However , Crif1f/+ , Fabp4 mice fed a NCD for 14 weeks developed glucose intolerance ( Figure 3B ) . More impressively , Crif1f/+ , Fabp4 mice fed a HFD for 8 weeks showed an earlier onset of glucose intolerance , which was characterized by higher peak glucose levels than those measured in control mice in the intraperitoneal glucose tolerance tests ( Figure 3C ) . Crif1f/+ , Fabp4 mice fed a HFD for 14 weeks showed more advanced glucose intolerance , with higher basal ( 168 . 8±13 . 2 mg/dL vs 131 . 3±8 mg/dL ) and peak ( 516 . 8±34 . 8 mg/dL vs 420 . 4±52 . 3 mg/dL ) plasma glucose levels ( Figure 3D ) . Therefore , regardless of the caloric state , mice with Crif1 haploinsufficiency showed reduced glucose tolerance . Crif1f/+ , Fabp4 mice fed a HFD for 14 weeks showed decreased Akt phosphorylation in the liver and muscle and a reduced glucose disposal rate after an intraperitoneal insulin challenge ( Figure 3E and 3F ) . Furthermore , suppression of hepatic glucose production ( HGP ) by insulin was not different between the two groups , but the glucose infusion rate ( GIR ) and glucose uptake rate decreased by approximately 18 . 6% and 14 . 7% , respectively , during hyperinsulinemic euglycemic clamping after 14 weeks on a HFD ( Figure 3G ) ; these data supported the insulin tolerance tests ( ITT ) results . These findings indicate that Crif1f/+ , Fabp4 mice , which have limited OXPHOS capacity in their adipose tissue , may show exacerbated diabetic mechanisms , which are characterized by insulin resistance . The levels of saturated fatty acids and ceramides in WAT , muscle and liver were not significantly altered in Crif1f/+ , Fabp4 mice ( Figure S4G and S4H ) . Thus , abnormal accumulation of ceramides and saturated fatty acids in insulin sensitive tissues does not appear to underlie the insulin resistance of Crif1f/+ , Fabp4 mice ( Figure S4G and S4H ) . To determine the molecular pathways that are dysregulated by mitochondrial OXPHOS dysfunction following Crif1 knockdown by siRNA in adipocytes , we introduced Crif1 siRNA into fully-differentiated 3T3-L1 cells . Crif1 knockdown in differentiated 3T3-L1 cells resulted in decreased expression of the OXPHOS subunits , ND1 , NDUFA9 , UQCRC2 and ATP5A1 , but did not affect the expression of Ppar-gamma , adiponectin , and Cd36 ( Figure 4A ) . A complementary DNA ( cDNA ) microarray analysis showed prominent increases in the expression levels of inflammatory cytokine and chemokine genes in adipocytes following knockdown of Crif1 ( Figure S5 ) . In particular , the chemokines monocyte chemotactic protein 1 ( Mcp1/Ccl2 ) , IFN-γ-inducible protein ( Ip10/Cxcl10 ) , Regulated upon Activation , Normal T cell Expressed and Secreted ( Rantes/Ccl5 ) and Mig/Cxcl9 , which are important for the recruitment of macrophages and T cells to WAT , were elevated in Crif1-silenced 3T3-L1 adipocytes [22] . The elevation of Mcp1 , Ip10 , and Rantes expression observed in cDNA microarrays was confirmed by real-time PCR experiments with Crif1-silenced 3T3-L1 adipocytes ( Figure 4B ) . In parallel experiments , levels of mitochondrial and cytoplasmic superoxide anions were increased in Crif1-silenced 3T3-L1 adipocytes compared to control cells ( Figure 4C ) . Treatment with the antioxidant N-acetylcysteine ( NAC ) suppressed the expression of Mcp1 , Ip10 and Rantes in Crif1-silenced 3T3-L1 adipocytes ( Figure 4D ) . Adipose inflammation links adipocyte dysfunction to insulin resistance , which are frequently observed in excessive adiposity ( for a review , see ) . The inflammatory process in adipose tissue is provoked by the activation of stress kinases , e . g . , c-Jun N-terminal kinase ( JNK ) , which inhibit insulin signaling and activate transcription factors that mediate the expression of chemokine genes [24] , [25] . Intracellular stress signals including mitochondrial ROS , FFA , ceramide and ER stress activates the stress kinases , JNK , p38 MAPK and NF-κB in adipocytes [4] , [26] , [27] . JNK mediates macrophage activation and expression of proinflammatory cytokines and inhibits insulin receptor substrate 1 ( IRS-1 ) -mediated insulin signaling pathways ( for a review , see [28] , [29] ) . To identify the roles of stress kinases in the expression of chemokines in Crif1-silenced 3T3-L1 adipocytes , p38 MAPK and JNK phosphorylation were observed by Western blot analysis . Levels of phosphorylated p38 MAPK and JNK were elevated in Crif1-silenced 3T3-L1 adipocytes compared to control cells; however , this activation was suppressed by NAC treatment ( Figure 4E ) . These results indicate that chemokine dysregulation is associated with increased ROS generation and inappropriate activation of p38 MAPK and JNK . To confirm these results , 3T3-L1 cells were treated with inhibitors of p38 MAPK and JNK ( SB203580 and SP60125 , respectively ) . Two inhibitors effectively inhibited the expression of Mcp1 in Crif1 silenced 3T3-L1 cells ( Figure 4F ) . Crif1 deficiency in MEFs results in increased ROS production [17] and induces phosphorylation of p38 ( Figure S6A ) . However , Crif1 -/Δ MEFs did not show increased Mcp1 and Ip10 expression ( Figure S6B ) . Taken together , these results suggest that limited mitochondrial OXPHOS function in fully-differentiated adipocytes triggers the expression of chemokines ( Mcp1 , Ip10 and Rantes ) in a cell-specific manner . The chemokines ( Mcp1 , Ip10 and Rantes ) upregulated in Crif1 siRNA-treated adipocytes are thought to be critical for attracting macrophages and T lymphocytes into adipose tissue in obese subjects [30] . Therefore , we wondered whether Crif1-silenced 3T3-L1 cells would trigger the migration of macrophages . As shown in Figure 4G , Crif1-silenced 3T3-L1 cells enhanced the migration of RAW 264 . 7 cells and NAC treatment inhibited the migration of RAW 264 . 7 cells . Thus , our in vitro studies show that OXPHOS deficiency induced in differentiated cultured 3T3-L1 adipocytes by Crif1 silencing results in the upregulated expression of chemokines , which then recruit or activate macrophages , ROS and stress kinase dependently . To observe ROS stress associated with abnormal chemokine responses in adipose tissues in vivo , we measured lipid peroxidation ( TBAR assays ) , stress kinase activation and cytokine expression in WAT of Crif1f/+ , Fabp4 mice fed a NCD or a HFD for 8 weeks . Consistent with the in vitro studies , lipid peroxidation in WAT and plasma was increased in Crif1f/+ , Fabp4 mice fed a HFD for 8 weeks compared to control mice ( Figure S7A ) . Levels of p38 MAPK and JNK phosphorylation were higher in WAT of Crif1f/+ , Fabp4 mice fed a HFD for 8 weeks than in control mice ( Figure 5A ) . Furthermore , the expression of Mcp1 , Ip10 and Rantes was higher in adipose tissue from Crif1f/+ , Fabp4 mice than in control mice ( Figure 5B ) . In addition , the level of secreted MCP1 , but not IP10 , were higher in the serum of Crif1f/+ , Fabp4 mice than in control mice ( Figure S7B ) . The results showing dysregulation of chemokines in the absence of Crif1 suggest that mitochondrial OXPHOS dysfunction may trigger immune cell recruitment in adipose tissue . To observe inflammation in the adipose tissue of Crif1f/+ , Fabp4 mice directly , the eWAT were stained with anti-F4/80 , an antibody that detects macrophages . Increased F4/80 reactivity was observed in the eWAT of Crif1f/+ , Fabp4 mice fed a NCD for 8 weeks . Aging and a HFD had an even more pronounced effect ( Figure 5C ) . Based on the quantitative real-time PCR results , the relative expression of proinflammatory M1 macrophage markers ( Cd11c , Cd11b and Tnfα ) increased significantly; however , the relative gene expression of an anti-inflammatory M2 macrophage marker ( arginase 1 ) did not change ( Figure 5D ) . To quantify the number of macrophages in the adipose tissue , multi-parameter flow cytometry was performed with anti-F4/80 , anti-CD11c and anti-CD206 antibodies using isolated stromal vascular fractions ( SVF ) . F4/80+/CD11c+/CD206- M1 macrophages were predominant in Crif1f/+ , Fabp4 mice compared with control mice . The proportion of F4/80+/CD11c-/CD206+ M2 macrophages tended to be higher in Crif1f/+ , Fabp4 mice , but this did not reach statistical significance ( Figure 5E ) . Taken together , the results suggested that the infiltrating macrophages were skewed towards the M1 phenotype in Crif1f/+ , Fabp4 mice . Recent studies show that B cell-mediated CD4+ and CD8+ T cell activation is required to induce inflammation and insulin resistance [31] , [32] . The present study found no difference between the numbers of CD4+ and CD8+ T cells in Crif1+/+ , Fabp4 and Crif1f/+ , Fabp4 mice ( data not shown ) . Adipocytes in adipose tissue secretes adipokines , such as adiponectin , leptin , IL-6 and TNFα , which are involved in the control of whole-body insulin sensitivity . However , proinflammatory TNFα is released by dysfunctional adipocytes and amplifies local immune responses by recruiting macrophages to WAT [33] , [34] . Serum levels of TNFα were consistently higher in Crif1f/+ , Fabp4 mice fed a HFD than in control mice ( Figure 5F ) . This indicates that TNFα may be a crucial mediator of inflammation in WAT and whole-body insulin resistance of Crif1f/+ , Fabp4 mice . The Fabp4 gene is expressed in macrophages [35] , but no Cre expression or activity was detected in macrophages isolated from Crif1f/+ , Fabp4 mice . As shown in Figure S8A , the expression levels of Crif1 and macrophage markers ( Cd11c , Tnfα , Cd11b , and Arg1 ) were not reduced in peritoneal macrophages obtained from Crif1f/+ , Fabp4 mice ( Figure S8A ) . Homologous recombination using PCR [36] identified Cre recombinase activity in WAT and BAT , but not in peritoneal macrophages in Crif1f/+ , Fabp4 mice at 20 weeks-of-age ( Figure S8B and S8C ) . To verify the adipose inflammation characterized by macrophage infiltration in Crif1-null mice , we generated another adipose tissue-specific Crif1 knockout mouse by crossing floxed Crif1 mice with Adipoq-Cre transgenic mice . Adipoq-Cre transgenic mice expressed Cre recombinase in WAT and BAT , but not in macrophages ( including adipose tissue resident macrophages , alveolar macrophages , or thioglycolate-stimulated peritoneal macrophages ) [37] . The homozygous Crif1 knockout mice ( Crif1f/f , Adipoq ) showed about ∼30% of the Crif1 expression observed in the eWAT of controls ( Figure S9A ) . They showed decreased expression of OXPHOS subunits ( ND1 , NDUFA9 , UQCRC2 and COX4 ) in eWAT and BAT , not in liver and heart ( Figure S9B ) . We compared adipocyte development in the adipocyte-specific Crif1 knockout mouse with that of Adipoq-Cre mice . H&E staining of adipose tissues indicated that the adipocytes of Crif1f/f , Adipoq mice were relatively smaller and irregularly shaped in comparison to those of Crif1+/+ , Adipoq mice ( data not shown ) . Consistent with the Crif1f/+ , Fabp4 mouse model , Crif1f/f , Adipoq mice showed higher plasma levels of MCP1 ( 1 . 9-fold higher ) , IP10 ( 2 . 5-fold higher ) and marked F4/80 immuno-reactivities in eWAT , suggesting inflammation in WAT ( Figure 6A and 6B ) . The nature of the macrophage phenotypes was further identified by flow cytometry using fluorescently-labeled anti-F4/80 , anti-CD11c , and anti-CD206 antibodies . In addition , the T cell population was also analyzed using anti-CD3 , anti-CD8 , and anti-CD4 antibodies . Crif1f/f , Adipoq mice had a higher level of M1 macrophages and a lower level of M2 macrophages in eWAT compared to Crif1+/+ , Adipoq mice ( Figure 6C ) . The level of cytotoxic CD8-positive T cells was increased 5 . 5-fold and the level of CD4-positive helper T cells was decreased 0 . 5-fold in Crif1f/f , Adipoq in comparison to control mice; however , the levels of these cells in Crif1f/+ , Fabp4 mice were not significantly different from control mice ( Figure 6D ) . This reflect differences in the severity of the defect in the mitochondrial OXPHOS complex in Crif1f/f , Adipoq and Crif1f/+ , Fabp4mice . Similar to Crif1f/+ , Fabp4 mice , Crif1f/f , Adipoq mice developed glucose intolerance even in being fed a NCD , at 8 weeks-of-age ( Figure 6E ) . Unlike Crif1f/f , Fabp4 mice , Crif1f/f , Adipoq mice were viable . This discrepancy could be due to the inherent differences in the activities of Cre-recombinase driven by Fabp4 and adiponectin promoter ( Table S1 ) . Loss of Crif1 was consistently observed in WAT in both mouse lines; however , the degree of Crif1 loss was more severe in Crif1f/f , Fabp4 mice than in Crif1f/f , Adipoq mice . Also , Crif1f/f , Fabp4 mice exhibited a severe loss of BAT and WAT mass , whereas the mass of these tissues was only mildly reduced in Crif1f/f , Adipoq mice . Consistent with these findings , homozygous Crif1f/f , Fabp4 mice rapidly reached a fatal low rectal temperature of 22 . 6+1 . 9°C ( Figure S2D ) , whereas homozygous Crif1f/f , Adipoq mice reached a milder rectal temperature of 29 . 5+0 . 7°C within 5 minutes of immersion in cold water ( 4°C ) ( Figure S9C ) . These results indicated that thermal stress caused by mitochondrial OXPHOS dysfunction in BAT following Crif1 ablation may be a causative factor of the premature death of adipocyte specific Crif1 knockout mice , and that BAT dysfunction may be partially involved in systemic glucose intolerance . To determine whether macrophages in the adipose tissue of Crif1f/+ , Fabp4 mice play a role in insulin resistance , we depleted macrophages from adipose tissue by intraperitoneal treatment with clodronate liposomes [38] . Clodronate is an apoptosis-inducing drug; therefore , injection of liposome-encapsulated clodronate into the intraperitoneal cavity can deplete phagocytic cells , such as macrophages . Control and Crif1f/+ , Fabp4 mice fed a HFD for 8 weeks were intraperitoneally administered two rounds of clodronate liposomes with an interval of 3 days . The accumulation of macrophages positively stained with an anti-F4/80 antibody was decreased in the eWAT following injection of clodronate liposomes ( Figure 7A ) . Furthermore , the level of Cd68 mRNA was significantly lower in the eWAT of mice injected with clodronate than in untreated mice ( Figure 7B ) . Administration of clodronate liposomes dramatically improved the insulin and glucose tolerance of Crif1f/+ , Fabp4 mice fed a HFD for 10 weeks ( Figure 7C and 7D ) . These findings indicate that suboptimal reserves of mitochondrial OXPHOS in the adipose tissue of Crif1-deficient mice induce macrophage recruitment , which may trigger systemic insulin resistance ( Figure 8 ) .
Mitochondrial dysfunction , characterized by reduced OXPHOS function in liver and skeletal muscle , is thought to be one of the underlying causes of insulin resistance and type 2 diabetes ( for a review , see [39] , [40] ) . In addition , reduced hepatic OXPHOS function is closely related to hepatic lipid accumulation and insulin resistance [41] . Collectively , these studies provide evidence of a role for mitochondrial OXPHOS dysfunction in the development of human insulin resistance and type 2 diabetes . However , animal models of OXPHOS dysfunction in skeletal muscle and liver do not exhibit the human insulin resistance and type 2 diabetes phenotype [21] , [42] , [43] . The absence of insulin resistance in mice with homozygous or heterozygous Crif1 deletion in the liver ( Crif1f/f , Alb , Albumin-Cre ) or skeletal muscle ( Crif1f/+ , MLC , MLC-Cre ) is in agreement with previous findings that hepatic and skeletal mitochondrial dysfunction does not cause insulin resistance ( Figure S10 ) . Therefore , whether or how mitochondrial OXPHOS contributes to the pathogenesis of insulin resistance remains to be resolved . It is reported that adipose OXPHOS capacity is controlled by both genetic and diet-induced obesity [10] , [44] , [45] , which potentially contribute to adipose tissue dysfunction and exacerbation of insulin resistance . However , whether changes in adipose OXPHOS capacity are a cause or a consequence of complications associated with insulin resistance has not been clarified in vivo . In this study , we have shown an association between limited mitochondrial OXPHOS capacity and adipose tissue inflammation and insulin resistance in a Crif1 haploinsufficiency animal model . Mitochondria play a key role in the differentiation and maturation of adipocytes . It is reported that marked mitochondrial biogenesis is observed during the adipocyte differentiation process in vitro . In fact , the concentration of mitochondrial proteins in differentiated 3T3-L1 adipocytes showed a 20- or 30-fold increase compared with that in pre-adipocytes [14] , [46] . Notably , chemical inhibition of respiratory chain function , for example by rotenone treatment , suppresses adipogenesis and induces changes in the expression levels of the key transcription factors , C/EBPα , PPARγ , and SREBP-1c [47] . However , the role played by mitochondria during adipogenesis has mostly been investigated in vitro by inhibiting or knocking down the genes encoding the OXPHOS complex . This study showed that homozygous Crif1 null mice generated by both Fabp4-Cre and Adipoq-Cre recombinase have defects in WAT and BAT development . These observations support that notion that intact OXPHOS function is critical for adipogenesis in vivo . By contrast , our own observations show that heterozygous Crif1 knockout mice do not have defects in adipogenesis and maturation under NCD and HFD conditions . This finding indicates that Crif1 haploinsufficiency and mildly reduced OXPHOS capacity do not cause the apparent failure of adipogenesis in WAT and BAT . Consistently , plasma and liver lipid levels were not increased in heterozygous Crif1 knockout mice , suggesting that the mice do not have the lipodystrophy phenotype . Furthermore , it is reported that insulin resistance in a mouse model of lipodystrophy was not relieved by controlling inflammation [48] . Therefore , insulin resistance in Crif1 haploinsufficient knockout mice may not be related to lipodystrophic changes . Similar to WAT , the development of BAT was severely perturbed in homozygous Crif1-null mice ( Crif1f/f , Fabp4 ) , which may be a critical factor in the early mortality of these mice . By contrast , BAT development was normal in heterozygous Crif1f/+ , Fabp4 mice , and the histology and ultrastructure of mitochondria were normal . Furthermore , core temperature responses to a cold environment suggest that Fabp4-Cre-driven haploinsufficiency of Crif1 may not affect the physiological function of BAT . Therefore , decreased BAT function and impaired energy expenditure may not be principal cause of the development of insulin resistance in Crif1-haploinsufficient mice . The OXPHOS capacity in adipose tissue may be controlled by tissue-specific , genetic and environmental factors . In fact , it is well known that each cell type develops and maintains a specific OXPHOS capacity to satisfy its metabolic and energetic demands . In addition , individual OXPHOS capacity is genetically determined by specific tissues [49] . The cellular and genetic factors that control adipose-specific OXPHOS capacity are not fully understood . Therefore , white adipocyte responses to marginal or limited OXPHOS capacity in vitro and in vivo remain to be elucidated . In the present study , we characterized the enhanced secretory chemokine responses in Crif1-silenced mature adipocytes . Chemokine production in WAT is physiological , but enhanced production is linked to adipose inflammation , which is usually observed in cases of excessive adiposity ( for a review , see [50] ) . Therefore , the earliest events that trigger the process of enhanced chemokine secretion are of great interest . Studies on signals that initiate adipose inflammation are mainly based on the model of ER homeostasis , lipolysis , and fatty acid signals in obese individuals ( for a review , see [4] ) . It is not known how mitochondria or OXPHOS dysfunction modify the chemokine responses in WAT under physiological and pathological conditions . We found that abnormal increases in ROS production and activation of p38 and JNK were associated with increased expression of Mcp1 , Ip10 and Rantes . The increase in ROS and the activation of p38 and JNK in adipose tissue are common denominators that respond to cellular stresses . Unexpectedly , haploinsufficient heterozygous Crif1f/+ , Fabp4 mice and control mice exhibited similar levels of adipose Akt phosphorylation in response to insulin injection . This indicates that insulin signaling in adipose tissue may not be the principal cause of the systemic glucose intolerance of Crif1f/+ , Fabp4 mice . Therefore , the relative importance of these factors ( increased ROS , p38 and JNK activation ) need to be addressed by suppressing or eliminating these events while studying abnormal chemokine responses and systemic insulin resistance . We showed that the WAT in Crif1-deficient mice is predominantly infiltrated by macrophages , regardless of excessive adiposity . An increase in the number of adipose tissue macrophages ( ATM ) is a prominent feature associated with excessive adiposity [51] , [52] . Increased expression of chemokines , especially MCP1 , is responsible for recruiting macrophages into the WAT [30] . The ATM infiltration of epididymal fat pads in Crif1-deficient mice showed several characteristic features . First , it was present in fat pads with normal adiposity . This finding suggests that macrophage recruitment to adipose tissue caused by impaired OXPHOS capacity may also develop independently of excessive adiposity , but is accentuated in cases of increased adiposity . Mitochondrial OXPHOS dysfunction in the adipose tissue of Crif1f/+ , Fabp4 mice fed a NCD for 8 weeks resulted in macrophage recruitment; however , the mice showed normal glucose tolerance . This suggests that a threshold level of macrophage recruitment or activation is required for the development of insulin resistance . Phenotypic analysis of ATM in Crif1-deficient mice demonstrated that the proportion of both M1 and M2 macrophages tended to be increased under NCD and HFD conditions . However , a phenotypic shift toward M1 macrophages was observed in the adipose tissue of Crif1-deficient mice . Thus , these features of macrophage recruitment in WAT were similar to those observed in a mouse model of diet-induced obesity [53] . Our data provide novel insights into the relationship between adipose inflammation and insulin resistance . This study supports the idea that adiposity overwhelms the genetically-determined OXPHOS capacity in adipose tissue , provoking an inflammatory response and insulin resistance . Therefore , it is possible that adipose mitochondrial OXPHOS capacity is an independent factor determining the risk of adipose inflammation and systemic insulin resistance in obese and even in non-obese subjects .
3T3-L1 cells were maintained in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% bovine calf serum ( Gibco BRL ) . Forty-eight hours post-confluence , the cells were differentiated with IBMX ( 0 . 5 mM ) , dexamethasone ( 1 µM ) , insulin ( 10 µg/ml ) and 10% fetal bovine serum ( Gibco BRL ) [54] . Crif1 siRNA ( GGA GUG CUC GCU UCC AGG AAC UAU U ) was transfected by using Lipofectamine RNAiMAX reagent ( Invitrogen ) into 3T3-L1 adipocytes on day 4 of differentiation . Migration of Raw264 . 7 cells was examined in 8 . 0 µm Transwell filters ( Corning Corp ) . Raw 264 . 7 cells were maintained on the top well , with the media from 3T3-L1 adipocytes in the bottom well . After twenty-four hours , the Raw 264 . 7 cells that had not migrated to the filter were removed , and the cells that had migrated through the filter were collected and stained with trypan-blue . ADSCs were cultured as previously described [55] . ADSCs were differentiated into adipocytes using IBMX ( 0 . 5 mM ) , dexamethasone ( 1 µM ) , insulin ( 10 µg/ml ) and rosiglitazone ( 0 . 5 µM ) in M199 medium ( Gibco BRL ) supplemented with 10% fetal bovine serum ( Gibco BRL ) . After induction of differentiation , lipid accumulation was detected with Oil red O staining . ADSCs were fixed with 10% neutralized formalin , washed with water , and then stained with freshly prepared 0 . 2% Oil red O solution . Primary antibodies against OXPHOS complex subunits ( NDUFA9 , SDHA , UQCRC2 , and ATP5A1 ) were purchased from Invitrogen . Anti-COX4 ( #4844 ) antibody was purchased from Cell Signaling . Anti-ND1 antibody ( sc-65237 ) was purchased from Santa Cruz Biotechnology . Secondary antibodies ( goat anti-mouse and goat anti-rabbit ) were obtained from Cell Signaling . Anti-p38 antibody , anti-phospho-p38 antibody , anti-JNK-antibody , anti-phospho-JNK antibody , anti-phospho-Akt and total-Akt antibodies were obtained from Cell Signaling and anti-β-actin , α-tubulin antibody was obtained from Sigma-Aldrich . Anti-UCP1 antibody was obtained from Abcam . Total RNA was isolated using Trizol ( Invitrogen ) . For Northern blot analysis , 10–20 µg of total RNA was loaded onto a 1 . 5% agarose-formaldehyde gel . A Crif1 probe was constructed using the mouse Crif1 gene digested with KpnI enzyme . The relative intensity of the Crif1/β-actin bands was normalized against that in the brain . Complementary DNA ( cDNA ) was prepared from total RNA using M-MLV Reverse Transcriptase and oligo-dT primers ( Invitrogen ) . Real-time PCR was performed using cDNA , QuantiTect SYBR Green PCR Master Mix ( QIAGEN ) , and specific primers . The primers used are described in Table S2 . Relative expressions were calculated normalized with 18s ribosomal RNA , using Rotor-Gene 6000 real-time rotary analyzer Software ( Version 1 . 7 , Corbett Life Science ) . Total RNA was prepared from fully-differentiated 3T3-L1 adipocytes transfected with control or Crif1 siRNA . RNA amplification and labeling were performed with the Low RNA Input Linear Amplification kit PLUS ( Agilent Technologies ) . Array hybridization and scanning were performed with a DNA microarray Chip and scanner ( Agilent Technologies ) . Array data was analyzed using the Feature Extraction and GeneSpring Software ( Agilent Technologies ) . Dihydroethidium ( DHE ) or MitoSOX were used to detect intracellular superoxide . Fully-differentiated 3T3-L1 cells were incubated with 10 µM DHE or 5 µM MitoSOX at 37°C for 15 min . Fully-differentiated 3T3-L1 cells were washed with Krebs-HEPES buffer ( pH 7 . 4 ) or HBSS . Images of cells stained with DHE or MitoSOX were obtained by fluorescence microscopy ( Olympus , Japan ) . Cells were trypsinized and analyzed using a FACScan flow cytometer ( BD Bioscience ) and data analysis was performed using BD FACSDiva software ( BD Bioscience ) . Before BN-PAGE , mitochondrial isolation was performed as previously described [56] with modifications . Pellets of ADSCs or tissues from mice were resuspended in buffer B ( 210 mM mannitol , 70 mM sucrose , 1 mM EGTA , and 5 mM HEPES , pH 7 . 2 ) and incubated for 5 min at 4°C . After centrifugation at 600× g for 10 min , the supernatant was re-centrifuged at 17 , 000× g for 10 min . The pellet containing the mitochondrial fraction was used in the Native PAGE Novex Bis-Tris Gel system ( Invitrogen ) to determine the content of the OXPHOS complex . A total of 20 µg of the mitochondrial fraction in Native PAGE sample buffer supplemented with 0 . 5% n-dodecyl-β-D-maltoside was loaded onto a Native PAGE Novex 3–12% Bis-Tris gel . The mitochondrial fraction was mixed with Native PAGE sample buffer containing 1% of digitonin to detect the supercomplexes . After electrophoresis , the separated proteins in the gel were transferred to a PVDF membrane , which was then incubated with an anti-OXPHOS antibody mixture ( Invitrogen ) . OCR was measured using a Seahorse XF-24 analyzer ( Seahorse Bioscience ) . Control Crif1+/+ and Crif1+/− ADSCs were prepared from the eWAT of Crif1+/+ , Fabp4 and Crif1f/+ , Fabp4 mice . After seeding ADSCs on an XF-24 plate , cells were incubated in differentiation M199 media contained with FBS , IBMX , dexamethasone , insulin and rosiglitazone . After 2 days later , Crif1+/+ and Crif1+/− ADSCs maintained M199 media with insulin for 8 days . The day before OCR measurement , the sensor cartridge was calibrated with calibration buffer ( Seahorse Bioscience ) at 37°C . Fully-differentiated ADSCs were washed and incubated with M199 ( Gibco BRL ) without sodium bicarbonate at 37°C in an incubator . Three readings were taken after each addition of mitochondrial inhibitor before injection of the subsequent inhibitors . The mitochondrial inhibitors used were oligomycin ( 2 µg/ml ) , carbonyl cyanide m-chloro phenyl hydrazine ( CCCP , 10 µM ) , and rotenone ( 1 µM ) . OCR was automatically calculated and recorded by the sensor cartridge and Seahorse XF-24 software . The plates were saved and the protein concentration was calculated to confirm that there were an approximately equal number of cells in each well . Floxed Crif1 ( Crif1flox/flox ) mice were generated as previously described [18] . Fabp4-Cre , Albumin-Cre transgenic mice ( C57BL/6J ) were purchased from the Jackson Laboratory . Adiponectin-Cre transgenic mice were kindly provided by Dr . Evan Rosen . Dr . Steven J Burden provided the MLC-Cre mouse strain . The HFD , which contained 60% fat , was purchased from Research Diets Inc . ( D12492 ) . Mice were maintained in a controlled environment ( 12 h light/12 h dark cycle; humidity 50–60%; ambient temperature 23°C±1°C ) and fed ad libitum . For the cold challenge experiments , mice were individually housed in cages pre-chilled to 4°C . Body temperature was monitored using a rectal probe attached to a digital thermometer ( TD-300 , Shibaura Denshi . Japan ) with/without cold stress . For the thermoneutrality experiments , 2-week-old mice were housed with their mothers at a temperature of 30°C±1°C . All mouse experiments were performed in the animal facility according to institutional guidelines , and the experimental protocols were approved by the institutional review board of Korean Research Institute of Biotechnology and Bioscience , and Chungnam National University . To measure the activity of Cre recombinase , PCR was performed as previously reported [36] . Briefly , after isolation of genomic DNA from WAT , BAT , and thioglycolate-induced peritoneal macrophages , PCR was performed with a combination of three primers: forward primer 1 , GGGCTGGTGA AATGTGTTG; reverse primer 2 , TCAGCTAGGG TGGGACAGA; and reverse primer 3 , TATCAGTCCG AGAAGACCTG . To ensure product specificity from PCR , the extension time was limited to 30 sec . WAT was fixed in 10% neutralized formalin for 16 h , washed , and then embedded in paraffin . Tissue sections of 5 µm thickness were deparaffinized , rehydrated , and heated in a microwave for 10 min in citrate buffer . The tissue sections were then incubated with primary antibodies ( anti-F4/80 ( diluted 1∶100; Abcam ) ) for 16 h at 4°C . Immunohistochemistry was performed using a Polink-1 HRP Rat-NM DAB Detection System ( GBI Inc ) . WAT and BAT were fixed in 1% glutaraldehyde at 4°C and then washed with 0 . 1 M cacodylate buffer at 4°C . After washing five times , the tissue was post-fixed with 1% OsO4 in an 0 . 1 M cacodylate buffer ( pH 7 . 2 ) containing 0 . 1% CaCl2 for 1 h at 4°C . Samples were dehydrated by serial ethanol and propylene oxide treatment and embedded in Embed-812 ( EMS ) . The resin was then polymerized at 60°C for 36 h . Tissue was sectioned using an EM UC6 ultramicrotome ( LEICA ) and stained with 4% uranyl acetate and citrate . Observation was performed using a Tecnai G2 Spirit Twin transmission electron microscope ( FEI Company , USA ) and a JEM ARM 1300S high-voltage electron microscope ( JEOL , Japan ) . For IPGTT , mice were fasted for 16 h and then 2 g/kg or 1 g/kg glucose was injected into the intraperitoneal cavity of each mouse . Blood glucose levels were measured at 0 , 15 , 30 , 60 , and 90 min using a glucometer ( Bayer breeze ) . ITT was performed by measuring blood glucose after 6 h of fasting followed by intraperitoneal insulin injection ( 0 . 75 U/kg; Humalog ) . Hyperinsulinemic euglycemic clamping was performed as previously described [57] . Briefly , after an overnight fast , a 2 h hyperinsulinemic euglycemic clamping was performed in Crif1f/+ , Fabp4 and control littermates ( n = 8 ) . The insulin clamp began with a primed-continuous infusion of insulin ( 0 . 3 U/kg bolus followed by 2 . 5 mU/kg/min ) . Blood samples ( 20 µl ) were collected at 10 to 20 min intervals for immediate measurement of plasma glucose concentrations , and 20% glucose was infused at variable rates to maintain glucose at basal concentrations ( ∼120 mg/dL ) . Basal and insulin-stimulated whole-body glucose uptake was estimated with a continuous infusion of 3H glucose ( Perkin Elmer Life and Analytical Sciences ) for 2 h before clamping ( 0 . 05 µCi/min ) and throughout the clamping ( 0 . 1 µCi/min ) , respectively . At 75 min after the start of the clamp , 2-deoxy-d-1-14C glucose ( PerkinElmer Life and Analytical Sciences ) was injected with a Hamilton syringe to measure insulin-stimulated glucose transport activity and metabolism in skeletal muscle . Blood samples were taken before , during , and at the end of the clamps for measurement of plasma 3H glucose and 2-deoxy-d-1-14C glucose concentrations , and/or insulin concentrations . At the end of the clamps , tissue samples ( gastrocnemius , eWAT , and liver ) were rapidly taken and stored at −70°C prior to biochemical and molecular analysis . To quantified M1 macrophages , M2 macrophages , and CD4+ and CD8+ T cell populations , the stromal vascular fractions ( SVF ) was isolated from mouse eWAT . The SVF was prepared by the lysis of eWAT with type 1 collagenase ( Gibco BRL ) in collagenase buffer at 37°C in a shaking water bath for 40 min , followed by centrifuging at 2000 rpm for 5 min . The suspended solid matter comprised adipocytes and the cell pellet comprised T cells , B cells and macrophages . The cell pellet was then incubated with RBC lysis buffer and the remaining cells were stained with specific antibodies . Anti-CD3 ( BD bioscience ) , anti-CD4 ( BD Bioscience ) and anti-CD8 ( eBioscience ) were used to stain the T cell population [58] , and F4/80 ( eBioscience ) , CD206 ( eBioscience ) and CD11c ( eBioscience ) were used to stain M1/M2 macrophages . The stained SVF cells were analyzed using a FACScan flow cytometer ( BD Bioscience ) and data analysis was performed using BD FACSDiva software ( BD Bioscience ) . The TBAR assay kit ( Cayman Chemicals ) was used to measure lipid peroxidation in the WAT and plasma of mice . WAT ( 25 mg ) suspended in RIPA buffer was sonicated , centrifuged at 1 , 600 g for 10 min at 4°C , and the supernatant was collected . The SDS solution was added to the supernatant , which was then mixed with the Color reagent according to the manufacturer's instructions . The sample was boiled for 1 h , centrifuged , and the supernatant was collected . Fluorescence at the excitation wavelength of 530 nm and emission wavelength of 550 nm was measured . The generation of liposome-encapsulated clodronate was performed as previously described [38] . Cholesterol ( 10 mg/ml; Sigma-Aldrich ) was dissolved in 100% ethanol , and 100 mg/ml phosphatidylcholine in 100% ethanol ( Sigma-Aldrich ) was made into a phospho-lipid film by drying with a low-vacuum rotary . Clodronate ( 0 . 6 M ) ( Sigma-Aldrich ) was dissolved in purified water and incubated with the phospho-lipid film by gentle rotation at room temperature and sonication in a water bath for 3 min at 55 kHz . After removing the non-encapsulated clodronate , liposome-encapsulated clodronate was resuspended in 1X PBS . Two intraperitoneal injections ( 3 days apart ) of clodronate were administered to mice fed a HFD for 8 weeks . IPGTT and ITT were performed 6 days after the first injection . Measurement of hepatic triglycerides: Liver triglycerides were extracted with chloroform and methanol , dissolved in 1× PBS , and measured in a Hitachi 7150 chemistry analyzer ( Hitachi , Japan ) . Measurement of ceramides in WAT , liver and muscle: Prior to extraction of total lipids , C17 ceramide was added as an internal standard . Ceramides were measured as previously described [59] . All liquid chromatography-mass spectrometry ( LC-MS/MS ) experiments were performed using an Agilent 1200 HPLC system ( Agilent Technologies , Santa Clara , CA , USA ) coupled to a Thermo LTQ linear ion trap mass spectrometer ( Thermo Scientific , San Jose , CA ) equipped with an electro spray ionization ( ESI ) source . Briefly , LC separation was achieved using a LunaC18 RP column ( 150 mm×2 mm I . D . , 5 µm 100 Å particles; Phenomenex , Torrance , CA ) with gradient elution . Lipid molecules separated by LC were detected by the mass spectrometer in Positive ESI mode using selected reaction monitoring ( SRM ) . The SRM channels were arranged as follows: 538→264 for C16 , 552→264 for C17 , 566→264 for C18 , 594→264 for C20 , 648→264 for C24:1 , and 650→264 for C24 . The peak area was normalized according to the internal standard and tissue weight . All values are presented as relative differences in the ratio of the extracted lipids to the internal standard . To measure the level of saturated fatty acids , tissues were homogenized in ice-cold methanol , and 1 µg of pentadecanoic acid ( C15:0 ) was added as an internal standard . Samples were incubated at 45°C overnight , then cooled to room temperature . Hexane and 1 mL of H2O were added , samples were vortexed and centrifuged , and fatty acid methyl esters were collected from the upper hexane layer . Samples were analyzed by gas chromatography–mass spectrometry ( GC-MS ) using an Agilent HP6890 GC interfaced with an HP5973N MSD . A DB-5 column was used . The GC oven temperature was initially 150°C and then increased to 280°C for 52 min . Full MS scans over a m/z range of 60 to 800 were obtained and the peaks of the characteristic ion chromatogram for each fatty acid methyl ester were used for quantification . All samples were normalized against the internal standard . Whole cardiac blood from the mice was incubated at room temperature for 2 h . The blood was centrifuged at 2 , 500 rpm for 5 min , and the supernatant was collected . TNFα and IL-4 were measured using a mouse cytokine/chemokine multiplex panel ( Millipore ) . MCP1 and IP10 were measured using an ELISA kit ( R&D Systems ) . Serum triglycerides and FFAs were measured with a Hitachi 7150 chemistry analyzer ( Hitachi , Japan ) . Data are presented as means ± or + standard deviation ( SD ) . Statistical significance for comparisons was determined using the Student's two-tailed T-test . A p value less than 0 . 05 was considered statistically significant .
|
Type 2 diabetes is one of the most challenging health problems in the 21st century . Although insulin resistance is regarded as a fundamental defect that precedes the development of type 2 diabetes , the nature and cause of insulin resistance remain unknown . Adipose tissue is an important organ that determines whole-body energy metabolism , and its dysfunction is a critical element in the development of systemic insulin resistance . Adipose mitochondrial function is suppressed in the insulin-resistant state , and increased adipose mitochondrial biogenesis is associated with the reversal of insulin resistance by a PPARγ agonist . However , despite these important observations , little is known about how mitochondrial respiratory dysfunction in white adipose tissue ( WAT ) causes insulin resistance . To determine whether adipose deficiency of mitochondrial respiratory capacity plays an etiological role in systemic insulin resistance , the metabolic phenotype of mice with mitochondrial OXPHOS ( oxidative phosphorylation ) –deficient adipose tissue was examined . Crif1 is a protein required for the translation of mtDNA–encoded OXPHOS subunits . Interestingly , mice haploinsufficient for Crif1 in adipose tissue showed reduced OXPHOS capacity and developed marked insulin resistance .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"biology"
] |
2013
|
Crif1 Deficiency Reduces Adipose OXPHOS Capacity and Triggers Inflammation and Insulin Resistance in Mice
|
Plants are continually exposed to pathogen attack but usually remain healthy because they can activate defences upon perception of microbes . However , pathogens have evolved to overcome plant immunity by delivering effectors into the plant cell to attenuate defence , resulting in disease . Recent studies suggest that some effectors may manipulate host transcription , but the specific mechanisms by which such effectors promote susceptibility remain unclear . We study the oomycete downy mildew pathogen of Arabidopsis , Hyaloperonospora arabidopsidis ( Hpa ) , and show here that the nuclear-localized effector HaRxL44 interacts with Mediator subunit 19a ( MED19a ) , resulting in the degradation of MED19a in a proteasome-dependent manner . The Mediator complex of ∼25 proteins is broadly conserved in eukaryotes and mediates the interaction between transcriptional regulators and RNA polymerase II . We found MED19a to be a positive regulator of immunity against Hpa . Expression profiling experiments reveal transcriptional changes resembling jasmonic acid/ethylene ( JA/ET ) signalling in the presence of HaRxL44 , and also 3 d after infection with Hpa . Elevated JA/ET signalling is associated with a decrease in salicylic acid ( SA ) –triggered immunity ( SATI ) in Arabidopsis plants expressing HaRxL44 and in med19a loss-of-function mutants , whereas SATI is elevated in plants overexpressing MED19a . Using a PR1::GUS reporter , we discovered that Hpa suppresses PR1 expression specifically in cells containing haustoria , into which RxLR effectors are delivered , but not in nonhaustoriated adjacent cells , which show high PR1::GUS expression levels . Thus , HaRxL44 interferes with Mediator function by degrading MED19 , shifting the balance of defence transcription from SA-responsive defence to JA/ET-signalling , and enhancing susceptibility to biotrophs by attenuating SA-dependent gene expression .
Plants and microbial pathogens co-evolve; pathogens are selected to evade host defence , and plants are selected to detect and resist pathogens [1] , [2] . Resistance mechanisms include not only pattern-triggered immunity ( PTI ) and effector-triggered immunity ( ETI ) [1] , but also local and systemic plant defence responses that are controlled through distinct , but partially interconnected pathways involving the hormones salicylic acid ( SA ) and jasmonic acid ( JA ) /ethylene ( ET ) [3] . Adapted pathogens have a substantial repertoire of effectors that can suppress PTI by various mechanisms [4] but only one effector has been shown to interfere with SA-triggered immunity ( SATI ) [5] . An important role in plant defence has been attributed to nuclear processes , since there are many reports that nuclear localisation of pathogen effectors , R proteins , and key host components , including transcription factors and regulators , is essential for plant immunity [6] . This observation suggests that effectors may manipulate host transcription or other nuclear regulatory components for the benefit of the pathogen . Although filamentous phytopathogens such as fungal rusts and powdery mildews and oomycete downy mildews and white rusts are more damaging to agriculture than bacteria , their effector functions are more poorly understood . Fungal and oomycete effectors are secreted , and then taken up by the host cell via a poorly understood mechanism that for many oomycetes involves the N-terminal RxLR motif [7] , [8] . Sequencing of several oomycete genomes including the model organism Arabidopsis downy mildew Hyaloperonospora arabidopsidis ( Hpa ) has allowed prediction of a repertoire of effector candidate genes that share N-terminal sequence motifs with known effectors [9] , [10] . To establish an inventory of the Hpa RXLR effectors ( HaRxLs ) , the draft genome of Hpa Emoy2 was scanned and HaRxL effector candidates were cloned . Because transformation of biotrophic pathogens such as Hpa is difficult , we developed heterologous systems to assess HaRxL functions [11] , [12] . We first deployed a Pseudomonas syringae pv . tomato ( Pst ) type three secretion ( T3S ) –based delivery system ( EDV ) to look for HaRxLs that enhance Pst virulence and/or that suppress host defence outputs such as callose deposition , in order to prioritize effectors for follow-up studies [13] , [14] . We next screened for the subcellular localisation of the HaRxL collection and identified 15 HaRxL effectors that localise to the plant cell nucleus when stably expressed in Arabidopsis [15] , [16] and interact in yeast with nuclear plant proteins implicated in transcription [16] , [17] . In particular , in yeast 2-hybrid ( Y2H ) assays , HaRxL44 interacts with MED19a , a subunit of the Arabidopsis Mediator complex [18] . Six other Hpa effectors interact with host Mediator or regulators of Mediator ( [18]; Figure S1 ) . Mediator is a conserved multisubunit complex that acts as a molecular bridge between transcriptional regulators at gene enhancer sequences and the activation of transcription by RNA polymerase II at the transcription start site [18] , [19] . Eight of 10 essential Mediator genes conserved between S . pombe and S . cerevisiae ( including MED19 ) also have a metazoan homologue , indicating that a Mediator core has been conserved throughout evolution and is present in all eukaryotic cells [20] . Mediator is a large complex ( >25 components ) , but different subunits are implicated in integration of specific external stimuli [21] , [22] . Mediator has numerous functions in addition to interacting directly with RNA polymerase II as it can interact with and coordinate the action of many other co-activators and co-repressors , including those acting on chromatin [20] . These interactions ultimately allow the Mediator complex to deliver outputs ranging from the maximal activation of genes , through the modulation of basal transcription , to long-term epigenetic silencing [20] . Despite the importance of Mediator , this complex has been little studied , due to the lethality of mutants in most multicellular organisms . However , null mutations of Mediator subunit genes are often not lethal in plants , making these organisms a valuable model for studying the Mediator complex . In Arabidopsis , several Mediator subunits have been shown to have a specific function in the activation of signalling pathways during plant development and in response to abiotic stress . MED12/CRP ( CRYPTIC PRECOCIOUS ) and MED13/MAB2 ( MACCHI-BOU2 ) are required for early embryo patterning , and also regulate flowering and cotyledon organogenesis , respectively [23] , [24] . MED14/SWP ( STRUWWELPETER ) is a key regulator of cell proliferation [25] . MED16/SFR6 ( SENSITIVE TO FREEZING6 ) integrates cellular and environmental cues into the circadian clock [26]–[28] and is required for cold acclimation . MED17 , MED18 , and MED20a play an important role in the production of small and long noncoding RNAs [29] . MED25/PFT1 ( PHYTOCHROME AND FLOWERING TIME1 ) was first identified as a key regulator of flowering [30] and later found to regulate final organ size and light signalling [31] , [32] . MED33a/RFR1 ( REF4-RELATED1 ) and MED33b/REF4 ( REDUCED EPIDERMAL FLUORESCENCE4 ) are required for phenylpropanoid homeostasis [33] . Mediator was recently shown to play a role in plant immunity and pest resistance . It was initially shown to be important for the activation of JA/ET-dependent defences against necrotrophic pathogens , via MED21 and MED25 [34] , [35] . Other studies reveal a role for Mediator in the activation of SATI [36] . The Mediator subunits MED14 , MED15 , and MED16 have all been reported to be required for the biological induction of systemic acquired resistance ( SAR ) [37]–[39] , suggesting that the Mediator may function in SAR activation . Both MED14 and MED15 appear to function downstream of NPR1 and do not affect the nuclear localisation or stability of NPR1 [37] , [39] , whereas MED16 makes a positive contribution to the accumulation of NPR1 protein [39] . The Mediator complex thus appears to be a “hub” for the plant immune system , but little is known about how the pathogen manipulates its function to promote disease . We report here the functional analysis of a nuclear downy mildew effector , HaRxL44 , which interacts with Mediator subunit 19a ( MED19a ) , and causes its degradation via proteasome-mediated degradation of this subunit . Expression profiling revealed an induction of JA/ET signalling in the presence of HaRxL44 , mimicking that observed after 3 d of compatible interaction . This increase in JA/ET signalling was associated with low levels of SATI in both Arabidopsis plants expressing HaRxL44 and in med19a knock-out mutants , whereas high levels of SATI were observed in plants overexpressing MED19a . Using the PR1::GUS reporter , we confirmed that Hpa abolishes PR1 expression specifically in cells containing haustoria . Thus , HaRxL44 affects via MED19a the balance between JA/ET and SA signalling and thus enhances biotroph susceptibility .
In a previous functional screen for Hpa virulence factors , we identified HaRxL44 ( Figure 1A ) as an enhancer of bacterial virulence in Arabidopsis [13] . The amino-acid sequence of HaRxL44 displays similarity to two predicted RXLR effectors from Phytophthora infestans , PITG-04266 and PITG_07586 , and avh109 from P . sojae ( Figure S2A ) . As observed for its homolog PITG_07586 from the “plastic secretome” of P . infestans [40] , HaRxL44 is found in a region of the Hpa genome enriched in retrotransposons ( Figure S2B ) and is conserved between Hpa races ( Figure S2C ) . We confirmed the effect of HaRxL44 on virulence ( Figure 1B ) by generating transgenic lines of Arabidopsis expressing HaRxL44 under the control of various promoters ( Figure S3 ) . Subcellular localization of GFP-HaRxL44 in a stably transformed Arabidopsis line ( Figure 1C ) confirmed its nuclear localization during Hpa infection , during which the nucleus is found closely associated with Hpa haustoria [15] . In an extensive Y2H screen [17] , HaRxL44 was found to interact with several nuclear proteins , including MED19a ( Figure S3A , S3B ) . We assessed the functional role of Mediator in immunity to Hpa , by studying the contribution of MED19a during Hpa infection . We first isolated med19a loss-of-function alleles ( Figure 2A , 2B ) and found that med19a mutant plants had a wild-type ( WT ) phenotype , with the exception of abnormally shaped siliques ( Figure 2C ) . In parallel , we generated Col-0 Arabidopsis transgenic lines overexpressing a construct encoding MED19a fused to a GFP tag ( OE MED19a; Figure 2D ) . Homozygous med19a-1 and med19a-2 mutants expressing GFP-MED19a were produced in order to check for complementation . We tested by Western blot the expression of GFP-MED19a in the mutant background , and selected lines with lower expression levels than observed for OE MED19a lines ( C1 , C2; Figure S4 ) . In these selected lines , GFP-MED19a rescued the phenotype observed during plant development ( Figure 2C ) . We confirmed that the fusion protein was functional , by checking that GFP-MED19a interacted with the Mediator complex . Immunoprecipitation of the GFP-MED19a protein in Arabidopsis led to the detection of both MED6 and MED7 in pull-down assays with native antibodies ( Figure 2E ) . We then analysed the subcellular localisation of MED19a in vivo in Arabidopsis by confocal microscopy . Live-cell imaging showed that GFP-MED19a and HaRxL44 were present in the same compartments: the nucleoplasm and nucleolus of the plant cell ( Figure 2F ) . Western-blot analysis of two independent transgenic lines producing GFP-MED19a ( Figure 2G ) demonstrated the presence of a GFP-MED19a protein of the expected size ( 50 kDa ) , together with additional signals at higher molecular weights ( 60 kDa and 70 kDa ) , suggesting that MED19a is modified post-translationally in planta . We then challenged the transgenic lines with Hpa and monitored pathogen growth after six days . Both the med19a-1 and med19a-2 mutants were more susceptible to Hpa than wild type , similar to a med15 mutant , which has impaired SATI ( med15 [37]; Figure 2H ) . Complemented lines displayed the same level of susceptibility as wild type plants ( Figure 2H ) , confirming the functionality of the fusion protein . By contrast , transgenic lines overproducing MED19a were more resistant to Hpa than the WT or Arabidopsis lines expressing GFP alone . Therefore , GFP-MED19a can associate with other Mediator subunits and complements med19a loss of function alleles , which suggests that the fusion protein is functional . Thus , the Mediator subunit MED19a is a positive regulator of plant immunity to Hpa . We monitored the subcellular location of RFP-MED19a and GFP-HaRxL44 using confocal microscopy . Both proteins localise to the nucleoplasm and nucleolus , whereas Bimolecular Fluorescence Complementation ( BiFC ) signals resulting from the co-expression of YFPc-MED19a and YFPn-HaRxL44 constructs are restricted to the nucleolus , following transient expression in N . benthamiana . No BiFC signal was detected in the nucleoplasm , the site of Mediator function ( blue arrow , Figure 3A ) . The destabilisation of RFP-MED19a in the presence of GFP-HaRxL44 was quantifiable by both Western blotting ( blue arrow , Figure 3B , Figure S5A ) and confocal microscopy ( blue arrow , Figure 3C ) . Furthermore , no decrease in the amount of GFP-MED19a was observed in coexpression experiments with RFP-24 and RFP-45 constructs , which encode other nuclear HaRxLs ( Figure 3C , Figure S5A ) , suggesting that MED19a is specifically targeted by the HaRxL44 effector . As MED19a transcript levels were not affected in HaRxL44 lines ( Figure S5B ) , we conclude that HaRxL44 destabilizes MED19a at the protein level . Taken together these results show that MED19a , which is found in both nucleoplasm and nucleolus , disappears in the nucleoplasm in the presence of HaRxL44 , and perhaps persists in the nucleolus because of low proteasome activity in the nucleolus . Since Mediator is known to function in the nucleoplasm , this HaRxL44-mediated degradation of MED19 likely affects Mediator activity . In the Y2H screen [17] , HaRxL44 was found to interact with two E3 ligases ( Figures 4A and S3 ) , BOTRYTIS SUSCEPTIBLE 1 ( BOI; AT4G19700 ) and MED25-BINDING RING-H2 PROTEIN-like ( MBR1-like; AT1G17970 ) . We investigated whether these E3 ligases are present in the same plant cell compartment as HaRxL44 and MED19a . We investigated the subcellular distribution of these two E3 ligases , by transiently expressing GFP-tagged versions of BOI and MBR1-like in N . benthamiana ( Figure 4B ) . GFP-BOI localises to the nucleoplasm and accumulates in foci , in four to five large bodies . Furthermore , no GFP-BOI signal was detected in the plant cell nucleolus ( Figure 4B ) . GFP-MBR1–like was also localised to the plant cell nucleus ( Figure 4B ) , in a pattern similar to that observed for proteins involved in RNA splicing [41] . GFP-MBR1–like accumulated in large amounts in the plant cell nucleolus and had a punctate distribution in the nucleoplasm ( Figure 4B ) . In order to test whether one of the two E3-ligases interacting with HaRxL44 in Y2H might be responsible for MED19a degradation , we tested the phenotype of BOI and MBR1-like loss-of-function mutants during Hpa infection . Surprisingly , both the boi RNAi line and the mbr1-like T-DNA KO line were more susceptible to Hpa ( Figure S5C ) . However , such loss-of-function experiments are difficult to interpret because BOI and MBR1-like might also affect other components of the plant immune system , leading to an increase in plant susceptibility . As HaRxL44 interacts in Y2H analysis with E3 ligases located in the plant cell nucleus ( Figures 4 and S3 ) , we hypothesised that HaRxL44 acts as an adaptor protein for E3 ligases , mediating the degradation of MED19a . Indeed , we showed that inhibition of the proteasome by the addition of 100 µM MG132 for 4 h prevented HaRxL44-induced degradation of GFP-MED19a ( Figure 4C ) . The addition of 100 µM MG132 during protein extraction prevented the degradation of GFP-MED19a in the presence of HA-HaRxL44 and made it possible to confirm the interaction of these proteins in planta , by co-immunoprecipitation ( red arrow , Figure 4D ) . We next tested if blocking the proteasome would allow the detection of the interaction between HaRxL44 and MED19a in the nucleoplasm . We showed that addition of MG132 1 h before observation with confocal microscopy allowed the detection of the interaction between YFPc-MED19a and YFPn-HaRxL44 in the nucleoplasm by BiFC ( Figure 3A ) . Thus , HaRxL44 interacts with MED19a , a positive regulator of plant immunity to Hpa , leading to its destabilisation in a proteasome-dependent manner . In order to check if the interaction between HaRxL44 with MED19a is important for its degradation , we generated a series of HaRxL44 mutants by NAAIRS-scanning mutagenesis [42] . We obtained one mutant , HaRxL44M , mutated in the nucleolus-localization signal ( Figure S6 ) , which no longer interacts with MED19a by Co-IP when transiently expressed in N . benthamiana ( Figure 5A ) . In contrast with HaRxL44 , which is visible in the nucleoplasm and the nucleolus ( Figure 5B ) , HaRxL44M presents a nuclear-cytoplasmic localisation ( Figure 5B ) . Using both cell biology ( Figure 5B , 5C ) and biochemistry ( Figure 5D ) we showed that HaRxL44M no longer degrades MED19a when transiently co-expressed in planta . Thus , the interaction between HaRxL44 and MED19a is important for proteasome-dependent MED19a degradation . First , we verified the degradation of MED19a in the presence of HaRxL44 in Arabidopsis , by generating a transgenic line expressing both GFP-MED19a and 3HA-Strep2-HaRxL44 ( or 3HA-Strep2-GUS as control ) . We showed that , as we observed in N . benthamiana , MED19a is degraded in the presence of HaRxL44 in Arabidopsis and the addition of MG132 blocks the effect of HaRxL44 on MED19a stability ( Figure S5D ) . We then investigated whether the presence of HaRxL44 affects the interaction between MED19a and the Mediator complex . We found that , even in the presence of 3HA-HaRxL44 , MED6 co-immunoprecipitates with GFP-MED19a in Arabidopsis ( Figure S5E ) , suggesting that MED6 and GFP-MED19a also associate in the nucleolus . However , overproduction of MED19a and HaRxL44 in Arabidopsis may affect the stoichiometry or nuclear/nucleolar distribution of interactions between MED19a and the Mediator complex , or Mediator subcomplexes , obscuring potential effects of HaRxL44 on the integration or stability of MED19a subunits in the Mediator complex . As MED19a is part of a major transcriptional regulatory complex , we then investigated whether and how HaRxL44 expression affects transcription . Illumina RNA-sequencing revealed a positive correlation between the genes differentially up-regulated in HaRxL44-lines and by methyl JA ( MeJA ) treatment [43] ( Hypergeometric probability <0 . 001; Figure 6 , Tables S1 and S2 ) . No correlation was observed for down-regulated genes in HaRxL44–line 1 ( Hypergeometric probability = 0 . 98 ) . This result can be explained by the lower number of genes differentially expressed in HaRxL44-line 1 compared to HaRxL44–line 2 . However , the average fold change in HaRxL44–line 1 is still correlated to what is observed in HaRxL44–line 2 ( Figure 6 , Table S2 ) . We confirmed , by QRT-PCR , that JA/ET marker genes ( PDF1 . 2 , JAZ1 , and JAR1 ) were induced in HaRxL44-lines and in med19a mutants , with respect to WT levels ( Figures 7A , 7B and S7A , S7B ) . Two of the five JA-responsive genes from the JA biosynthesis pathway [44] , OPR3 ( AT2G06050 ) and LOX2 ( AT3G45140 ) , were up-regulated ( Figure S7C , Table S2 ) , suggesting that HaRxL44 may induce JA/ET signalling . We then checked whether the induction of JA/ET-responsive genes in the presence of HaRxL44 was biologically significant . We conducted gene expression profiling over a time course of Hpa infection in Arabidopsis and found that PDF1 . 2 induction is observed 3 d after infection ( DAI ) , when HaRxL44 transcription was induced ( Figure 7C , 7D ) . Furthermore , the induction of JA/ET-responsive genes in HaRxL44 transgenic lines was similar to the induction observed during early stages of Hpa infection in susceptible accessions of Arabidopsis ( Figure 6 , Tables S1 and S2 ) . Thus , JA/ET signalling is induced in the presence of HaRxL44 , the absence of MED19a , and 3 d after Hpa infection . In Pst , the phytotoxin coronatine ( COR ) acts as an analogue of JA and contributes to bacterial invasion [45] . COR biosynthetic ( COR− ) mutants of Pst strain DC3000 exhibit reduced virulence on Arabidopsis when surface-inoculated [45] . In order to test if HaRxL44 was able to complement PstCOR− strain , we delivered HaRxL44 in planta by using EDV system ( EDV-HaRxL44 , [13] ) . When spray-inoculated in Arabidopsis Col-0 plants , PstCOR− growth was reduced by two logs ( cfu/cm2 ) compared to Pst ( Figure 7E ) . PstCOR− EDV-HaRxL44 growth was increased by one log compared to PstCOR− ( **p value<0 . 01 , Figure 7E ) . These results indicate that HaRxL44 is able to complement the deficiency of COR production in PstCOR− and supports a key role for this Hpa effector in the activation of the JA/ET pathway . Because JA/ET-induced defence is effective against necrotrophs [3] , we next challenged the transgenic lines expressing HaRxL44 with Botrytis cinerea ( Figure 7F ) . As control , we used loss-of-function alleles of HISTONE MONOUBIQUITINATION1 ( HUB1 ) shown to increase susceptibility to B . cinerea , and HUB1-OE lines that confer resistance to B . cinerea [34] . We observed that B . cinerea grew less well in HaRxL44-OE lines than in the WT , as also observed for HUB1-OE lines ( *p value<0 . 01; Figure 7F , Figure S7D to S7F ) . Altogether , these results suggest that JA/ET-dependent defence is promoted in lines that express HaRxL44 . The Mediator complex is known to be important for JA/ET signalling [46] . In particular , MED25 and MED21 are key components of Mediator that regulate JA/ET-induced gene expression [34] , [35] . We then tested if the med21 and med25 loss-of-function mutants are altered in Hpa growth . We observed that in both med21 RNAi line and med25 knock out ( KO ) mutants , Hpa growth was reduced compared with WT ( Figure 2H ) . Thus , JA/ET-responsive gene transcriptional activation via Mediator is important for Hpa virulence . As the activation of the JA/ET defence pathway can antagonise SATI [3] , we next assessed whether HaRxL44 suppresses SATI . We first observed , by QRT-PCR , that SA marker genes ( PR1 , LURP1 , WRKY70 , PR2 , PR5 genes ) are down-regulated in HaRxL44 transgenic lines ( Figures 8A–C and S8A , S8B ) . We then assessed PR1 induction after elicitation . In HaRxL44 transgenic lines , basal PR1 transcript levels are lower than those in the WT , resulting in a reduction of PR1 induction levels 8 h after SA treatment ( Figure 8D ) . Similar results were observed in med19a mutants ( Figure 8E ) , whereas MED19a OE led to stronger PR1 induction ( from 5 to 15 times higher level of PR1 expression in MED19 OE lines compared to control plants; Figure 8F ) . We then investigated whether Hpa suppresses SATI . Expression profiling in Col-0 plant infected with Hpa Waco9 revealed a 40-fold change in PR1 gene induction 3 DAI ( Figure 9A ) . We then investigated the cell-specific expression pattern of PR1 , by infecting PR1::GUS lines with Hpa . PR1::GUS staining was restricted to the plant vascular tissues in contact with Hpa 3 DAI ( Figure S8C , S8D ) , whereas strong GUS staining was observed throughout the entire leaf 6 DAI ( Figure 9B ) . An analysis of PR1 expression patterns at the cellular level showed that PR1::GUS staining was absent from Hpa-infected cells , whereas PR1::GUS staining was observed only in the cell layer surrounding the mesophyll cells into which haustoria had penetrated ( Figure 9B ) . Thus , Hpa suppresses SATI specifically in the haustoria-containing mesophyll cells to which the effector proteins are delivered . As expected , the amount of PR1 mRNA generated in response to Hpa was lower in the absence of med19a , as shown by QRT-PCR ( Figure 9C ) . We next tested whether MED19a is degraded upon infection by Hpa . In GFP-MED19a lines , we tried to image signal in an infected mesophyll cell and compare this to the signal level to the signal in the neighbouring cells . However , measurement of fluorescence by confocal microscopy in deep tissues was too difficult to allow us to obtain reliable results . Therefore , we used the med19a mutant lines complemented with GFP-MED19a in order to check by Western blot analysis the GFP-MED19a protein level in Hpa-infected tissues compared to uninfected tissues . GFP-MED19a signal in infected tissues was reduced compared to uninfected tissues ( Figure 9D ) , confirming that this positive regulator of plant immunity against Hpa is degraded after infection . We suggest that the destabilisation of MED19a by HaRxL44 results in transcriptional reprogramming , leading to changes in the balance between the JA/ET and SA pathways , promoting biotrophy .
MED19/Rox3 was originally identified in a search for mutants increasing aerobic expression of the CYC7 gene in yeast [47] . The nuclear localisation of this protein and the nonviability of null mutants suggest that the MED19/Rox3 protein is a general regulatory factor [47] . The purification of Mediator from a strain lacking the MED19 subunit [48] led to the demonstration that MED19/Rox3 regulated intermodule interactions in the S . cerevisiae Mediator complex . In Arabidopsis , MED19 is encoded by two genes , MED19a and MED19b . Only MED19a has been reported to be involved in Mediator complex formation in Arabidopsis [18] . HaRxL44 interacts with both MED19a and MED19b in Y2H screen [17] . We therefore tried to amplify both genes , but were unable to amplify the MED19b gene from cDNA or genomic DNA . Furthermore , no T-DNA insertion into the MED19b gene is available , limiting analyses of the function of this gene in response to Hpa . In this study , we focused on the role of MED19a during Hpa infection . However , it should be borne in mind that the phenotype observed for med19a KO mutants may be only partial , because MED19a and MED19b could have redundant functions . In Arabidopsis , there are other Mediator subunits encoded by several genes , such as MED10 , MED20 , MED22 , and MED33 . Transcript profiling with med20a and the RNA polymerase II subunit RPB2 mutant nrpb2-3 revealed a high degree of overlap in the lists of genes displaying down-regulation in the two mutants [29] . This suggests that even a single mutation in one of several paralogs encoding an Arabidopsis Mediator subunit can lead to a quantifiable phenotype . We first confirmed that MED19a was part of the Mediator complex , by demonstrating its interaction with MED6 and MED7 in planta . We then investigated the subcellular distribution of MED19a , which was found to be localised to the plant cell nucleoplasm , as reported for MED16 [26] . MED19a was also localised to the plant nucleolus . This is surprising , because Mediator is thought to associate with RNA polymerase II in the nucleoplasm . It has been suggested that Mediator regulates the action of other plant RNA polymerases [49] . The similarities between RNA polymerases II , IV , and V raise the possibility that Mediator may associate with another polymerase , either polymerase IV or polymerase V [49] . MED19a may even associate with the nucleolar RNA polymerase I or III . Indeed , Mediator subunits have been shown to interact with RNA polymerase I subunits in Y2H assays ( [50] , Figure S1 ) . However , as the evidence concerning the possible role ( s ) of Mediator in directing the activity of other RNA polymerases remains inconclusive , we decided to focus on the role of MED19a in the regulation of transcription by RNA polymerase II . A proteomic analysis of human nucleoli revealed the presence of a large number of proteins with no known nucleolar function [51] . Nucleolar protein composition is not static and may undergo significant modification in response to the metabolic state of the cell [52] . The regulation of protein activity by nucleolar sequestration has been reported before [53] , [54] . Indeed , this phenomenon has already been reported for human MED1 [55] . MED1 is phosphorylated by MAPK1 or MAPK3 during the G2/M phase , enhancing protein stability and promoting the entry of this molecule into the nucleolus [55] . We can speculate that MED19a is sequestered in the nucleolus to remove it from the functional pool of MED19a in the nucleoplasm . We then investigated whether the presence of HaRxL44 affected the interaction between MED19a and the Mediator complex . We showed that even in the presence of HaRxL44 , MED19a associated with MED6 in Arabidopsis . However , we cannot exclude the possibility that the overproduction of MED19a and HaRxL44 in Arabidopsis affects the stoichiometry between MED19a and the Mediator complex , obscuring potential effects of HaRxL44 on the integration of MED19a subunits into the Mediator complex . We show here that HaRxL44 interferes with Mediator function by promoting the proteasome-dependent degradation of MED19a . Effectors from plant pathogens have been reported to suppress various layers of plant defence by controlling the ubiquitination and degradation of proteins important for plant immunity via the proteasome . AvrPtoB is a well-studied Pseudomonas syringae effector that mimics a plant E3 ligase [56] and facilitates the degradation of key components of PAMP-triggered immunity [57]–[60] . The Xanthomonas effector XopL has been shown to display E3 ubiquitin ligase activity in vitro and in planta , to induce plant cell death , and to suppress plant immunity [61] . The structural fold of the E3 ubiquitin ligase domain in XopL is unique , and the lack of cysteine residues in the XL-box suggests a noncatalytic mechanism for XopL-mediated ubiquitination [61] . The P . syringae effector HopM1 mediates the degradation , by the proteasome , of AtMIN7 , a plant protein involved in the vesicular trafficking of defence components [62] , [63] . Unlike AvrPtoB and XopL , HopM1 has no E3 ligase activity , suggesting that this effector acts as an adaptor protein , connecting AtMIN7 and the proteasome [62] . Several ubiquitin proteins have been identified in the Meloidogyne incognita secretome , and a ubiquitin extension protein secreted from the dorsal pharyngeal gland of Heterodera schachtii has also been detected [64] , [65] . The Magnaporthe oryzae effector AvrPiz-t was recently reported to interact with a RING E3 ubiquitin ligase , APIP6 , abolishing its ubiquitin ligase activity [66] . In addition , the P . infestans RXLR effector AVR3a has been shown to target and stabilise the nucleolar E3 ligase CMPG1 , which is required for the programmed cell death triggered by the elicitin INF1 [67] , [68] . However , the targets for the ubiquitination of these E3 ligases have yet to be determined . We show here that HaRxL44 interacts with MED19a , destabilising this Mediator subunit in a proteasome-dependent manner . As HaRxL44 displays no sequence similarity to plant E3 ligases , we suggest that , like HopM1 , HaRxL44 acts as an adaptor , presenting MED19a to the proteasome or to an E3 ligase . However , the mechanism by which HaRxL44 induces the degradation of MED19a remains unclear . Y2H screens have shown that HaRxL44 interacts with two E3 ligases: BOI and MBR1-like [17] . BOI is encoded by a gene from a multigene family with four known members , including BOI-RELATED GENE [69] . BOI was identified in a screen for proteins interacting with BOTRYTIS SUSCEPTIBLE 1 ( BOS1 ) , which encodes an R2R3 MYB transcription factor involved in restricting necrotroph-induced necrosis [70] . BOI is an important player in plant immunity to necrotrophic pathogens [71] . BOI ubiquitinates BOS1 , leading to its rapid degradation by the proteasome [71] . In addition to its role in restricting necrosis , BOI may integrate plant responses to diverse signals [72] . Indeed , Park et al . ( 2013 ) recently showed that BOI and DELLA proteins inhibit GA responses by interacting with each other , binding to the same promoters of GA-responsive genes , and repressing these genes . In the Y2H screen carried out by Mukhtar et al . ( 2011 ) [17] , BOI was found to interact with four nuclear effectors from Hpa: HaRxL44 , HaRxL10 , ATR1 , and ATR13 . Thus , Hpa effectors may act on BOI function , to render the plant more susceptible to biotrophic pathogens . It is not clear whether the HaRxL44-mediated degradation of MED19a by the proteasome has a positive or negative impact on transcription . It is well known that one major way of regulating transcription is to couple the activity of transcription factors to their destruction by the proteasome [73] . This “transcription-coupled destruction” mechanism of activator action [74] must serve a functional purpose , such that , if blocked , repeated rounds of transcriptional activation cannot occur [73] . This “unstable when active” phenomenon is seen with many transcriptional regulators , including the Mediator subunit MED25 [75] . In Arabidopsis , MED25 is a highly unstable protein , degraded by the proteasome both in vitro and in vivo [75] . A blockade of proteasome activity prevents MED25 from inducing flowering [75] . Two E3 ubiquitin ligases , MBR1 and MBR2 , have been shown to polyubiquitinate MED25 in planta , supporting the “transcription-coupled destruction” model for the regulation of MED25 . MBR1 and MBR2 are part of a small cluster of E3 ligases in Arabidopsis [76] . HaRxL44 has been shown to interact with MBR1-Like in Y2H screens [17] . Thus , HaRxL44 may recruit different E3-ligases , to promote the destruction of MED19a , thereby promoting Hpa growth . In metazoans , Mediator complex subunits are degraded upon cell differentiation [77]–[79] . This observation is consistent with the notion that subcomplexes of Mediator may display cell type–specific activity [78] . The degradation of some subunits helps to turn off the expression of a large portion of genes , whereas the retention of other subunits is required for the expression of a smaller , highly specific subset of genes [78] , [80] . Based on our results , we hypothesise that HaRxL44 targets MED19a for degradation , to block the transcription of genes important for plant immunity ( i . e . , genes important for SA-dependent defence ) , whereas MED19a degradation allows the transcription of a small number of genes beneficial for Hpa , including JA/ET-induced genes . We showed that JA/ET signalling is induced in the presence of HaRxL44 ( or the absence of MED19a ) . Expression profiling using Illumina RNA sequencing revealed a positive correlation between the genes differentially up-regulated in HaRxL44-lines and by MeJA treatment [43] . No correlation was observed for down-regulated genes in HaRxL44–line 1 , but this result can be explained by the low number of genes differentially expressed in HaRxL44–line 1 compared to HaRxL44–line 2 . However , the average fold change in HaRxL44–line 1 is still correlated to what is observed in HaRxL44–line 2 . This result is consistent with the quantitatively different phenotypes observed in these transgenic lines , such as susceptibility to Hpa ( Figure 1B , [13] ) , induction of PDF1 . 2 ( Figure 7A ) , and suppression of SA-responsive genes ( Figure 8A-D ) . Thus , we believe that HaRxL44 affects JA/ET-regulated gene expression . Indeed , HaRxL44-expressing plants showing activation of JA/ET-responsive genes are more resistant to the necrotrophic pathogen B . cinerea for which JA/ET-dependent defence is required . Conversely , HaRxL44 expression ( or the absence of MED19a ) resulted in a loss of PR1-induction and higher rates of biotrophic pathogen growth . These results suggest that HaRxL44 affects the hormonal balance between JA/ET and SA , promoting biotrophy , by acting on the transcriptional machinery of the plant . Hpa infection also led to the expression of JA/ET-responsive genes , confirming the biological significance of the results obtained in the functional analysis of HaRxL44 . JA/ET and SA-dependent defences are known to be antagonistic . Arabidopsis mutants with impaired SA accumulation , such as eds4 , eds5 , and pad4 , display high levels of PDF1 . 2 expression in response to inducers of JA/ET-dependent gene expression [81] , [82] . Convincing evidence for such an antagonistic effect has also been reported for NON-EXPRESSOR OF PATHOGENESIS-RELATED GENES1 ( NPR1 ) [83] . NPR1 is the key regulator of SAR , an important reaction in defence against pathogens . The Arabidopsis npr1 mutant displays high levels of JA/ET-responsive gene transcript accumulation and of JA and ET accumulation in response to P . syringae infection , suggesting that NPR1 is involved in the SA-mediated suppression of JA/ET signalling [83] . Reciprocally , an mpk4 mutant has been shown to display constitutive SA-dependent gene expression and higher SA levels and enhanced resistance to biotrophic pathogens [84] . MPK4 up-regulates JA/ET-responsive genes and simultaneously suppresses SAR , placing MPK4 at the heart of the antagonistic interaction between JA/ET and SA [84] , [85] . The role of the Mediator complex in JA/ET and SA-responsive gene expression has recently been investigated . MED25 , MED21 , and MED8 have been shown to be important for the activation of JA/ET-induced gene transcription [34] , [35] . MED25 plays a major role in the JA-responsive gene transcription pathway , through its interaction with the transcription factor MYC2 , which plays a key role in the activation of JA-induced gene expression [86]–[88] . MED25 regulates JA-dependent defence responses , conferring resistance to necrotrophic pathogens , and a med25 mutant has been shown to be more susceptible than the WT to the hemibiotroph Fusarium oxysporum [89] . The effect of a med8 mutation on the JA/ET-induced expression of PDF1 . 2 is readily detectable only in med8 med25 double mutants [35] . MED21 RNA interference lines are susceptible to both B . cinerea and A . brassicicola [34] . MED21 has been shown to interact with a RING E3 ligase , HISTONE MONOUBIQUITINATION1 ( HUB1 ) , increasing resistance to necrotrophs [34] . MED14 , MED15 , and MED16 were recently reported to up-regulate SAR in Arabidopsis [37]–[39] , [90] . We show here that mutations of the gene encoding MED19a increase the basal level of JA/ET-responsive gene transcription and decrease the responsiveness of PR1 gene expression to SA . The abolition of PR1 expression or the absence of MED19a ( or the presence of HaRxL44 ) was associated with faster growth of Hpa in med19a KO mutants . Thus , HaRxL44 targets a positive regulator of plant immunity to biotroph pathogens , thereby interfering with hormonal balance and promoting biotrophy . When the first results from expression profiling host gene expression became available , a paradox emerged [91] . Even susceptible plants , in which Hpa is presumably suppressing host defences , show strong activation of a set of plant genes induced by SATI during SAR . Why does this defence activation not preclude pathogen infection ? Our cell biology analysis reported here resolves this paradox . We show that , during Hpa infection , the pathogen blocks PR1 induction in cells with haustoria , suggesting that the HaRxL effectors act at the transcriptional level , blocking PR1 expression ( and presumably other genes of the SATI regulon ) , to promote virulence . Further analysis requires methods , currently under development , to expression profile specifically from infected cells . HaRxL44 is unlikely to be the sole effector that accomplishes this shift in hormonal balance that promotes biotrophy . Indeed , other nuclear-HaRxLs have been shown to interact with the Mediator complex as well as with other regulators of JA/ET pathway , like JAZ proteins [17] . Functional analyses of these effectors should facilitate the discovery of new components of nuclear immunity and the engineering of improvements to plant defences , to strengthen disease resistance in crops .
To generate HaRxL44 constructs , primers were designed from the Hpa Emoy2 genome version 8 . 3 . HaRxL44 was amplified from the signal peptide cleavage site ( ΔSP-HaRxL44 ) until the stop codon using genomic DNA extracted from Hpa Emoy2 conidiospores , proof reading polymerase ( Accuprime Pfx , Invitrogen ) , and standard PCR conditions . The HA tag sequence was added to the Fw primer ( CACCATGTATCCGTACGACGTACCAGACTAC GCAATTGAAGTTGTCCCC ) in order to create an HA-HaRxL44–tagged version . The PCR fragment was inserted into the pENTR-D-TOPO and then in the plant expression vectors pK7WGF2 , dP2 [92] , and pBAV150 using Gateway Technology ( Invitrogen ) . The constructs were sequenced by The Genome Analysis Centre ( Norwich , UK ) and transformed into Agrobacterium tumefaciens strains GV3101 and GV3103 . For the prediction of HaRxL44 nucleolar localisation signal , NoD [93] was used http://www . compbio . dundee . ac . uk/www-nod/index . jsp . HaRxL44M NAAIRS mutant was generated by overlapping PCR using the primers Fw AATGCTGCTATA CGATCGAAACACAAGAGG and Rev CGATCGTATAGCAGCATTCTTGTGCCAGCC . MED19a ( AT5G12230 ) was amplified from Arabidopsis Col-0 cDNA obtained from flowers using the primers: MED19a F1-CACCATGGAGCCTGAACGTTTAAA and MED19a R1-TTAGCCAGCAACCCTTATTGCACC . BOI was amplified from Arabidopsis Col-0 genomic DNA using the primers F2-CACCATGGCTGTTCAAGCTCATC ACATGAACATTTTC and R2-TCAAGAAGACATGTTAACATGCACACTAGCGTTCA TGACCATATCGC and MBR1-like ( At1G17970 ) using the primers F3-CACCATGTCTTCTACAACAATCGGCGAGCACATCAG and R3-TTAAGGCTTGCC ATATGCTGCCTTCTTACAGACCG . The PCR fragment was inserted into the pENTR-D-TOPO and then in the plant expression vectors pK7WGF2 and pH7WGR2 using Gateway Technology ( Invitrogen ) . The constructs were sequenced by The Genome Analysis Centre ( Norwich , UK ) and transformed into A . tumefaciens strain GV3103 . To isolate homozygous med19a-1/med19a-1 and med19a-2/med19a-2 plants , we could not analyse the segregation of the kanamycin marker carried by the T-DNA on progenies because of the loss of kanamycin resistance in these SALK lines ( SALK_037435 . 47 . 85x and SALK_034955 . 56 . 00x ) . For mbr1-like mutant , we use a homozygous line from the SALK named SALK_025248 . 37 . 45 . x . T-DNA insertions were checked by PCR genotyping using T-DNA left border and gene-specific primers designed by the Salk Institute Genomic Analysis Laboratory ( SIGnAL ) ( http://signal . salk . edu/tdnaprimers . 2 . html ) using default conditions . Homozygote lines were identified . For protein extraction , frozen plant tissues were ground and mixed with an equal volume of cold protein isolation buffer [20 mM Tris-HCl ( pH 7 . 5 ) , 1 mM EDTA ( pH 8 . 0 ) , 5 mM DTT , 150 mM NaCl , 0 . 1% SDS , 10% glycerol , 1× Protease Inhibitor Cocktail ( Sigma ) ] . The mixture was spun down , and the supernatant was transferred to a new tube and boiled in 5× SDS loading buffer [300 mm Tris-HCl ( pH 6 . 8 ) , 8 . 7% SDS , 5% β-mercaptoethanol , 30% glycerol and 0 . 12 mg/ml bromophenol blue] . For co-immunoprecipitation experiment , frozen leaf samples were ground in liquid nitrogen . The resulting powder was transferred into prechilled SM-24 20 mL centrifuge tubes containing chilled extraction buffer ( 4–10 mL ) [1 M Tris HCl pH 7 . 5 , 5 M NaCl , 0 . 5 M EDTA , 20% glycerol , 10 mM DTT , 1× Protease inhibitor ( Sigma ) , 20% Triton X-100 , 2% PVPP] . Tubes were vortexed and equilibrated before centrifugation 20 min at 20 , 000 rpm at 5°C . After centrifugation , supernatants were filtered to remove plant debris ( Biorad Poly-Prep Chromatography columns ) . Proteins were quantified by Bradford assay . Three micrograms of total protein extracts were used for co-immunoprecipitation in protein Lo-Bind safe-lock tubes ( Eppendorf ) in which 25 µL of slurry solution of GFP beads ( Chromotek ) were added . Tubes were incubated on a rolling wheel for 2 to 4 h at 5°C . After incubation beads were washed with extraction buffer without PVPP by repeated low-speed centrifugations ( up to four washes ) . Beads were resuspended in 5× SDS loading buffer prior to flash-freezing in liquid nitrogen . Proteins were separated by SDS-PAGE , electro-blotted onto PVDF membrane ( Biorad ) , and probed with horseradish peroxidase-conjugated anti-RFP ( Abcam ) or anti-GFP ( Roche ) antibody . MED6 and MED7 primary antibodies ( from Bjorklund's lab ) were used at 1∶1 , 000 . Bands were visualized by chemiluminescence using Pico/Femto ( Thermo Scientific ) . For Hpa infection , 10-d-old plants were spray-inoculated to saturation with a spore suspension of 5 . 104 spores/ml . Plants were kept in a growth cabinet at 16°C for 3 d with a 16 h photoperiod . To evaluate conidiospore production , 10 pools of 2 plants were harvested in 1 ml of water for each line . After vortexing , the amount of liberated spores was determined with a haemocytometer as described by [94] . Statistical analyses have been performed from three independent experiments , using ANOVA . For B . cinerea infection , spores from the fungus strain B05 . 10 were obtained from Dr . Henk-Jan Schoonbeek ( John Innes Centre , Norwich , UK ) . Inoculation of Arabidopsis with B . cinerea spores was performed as described previously [95] . Briefly , 5-wk-old plants were inoculated with a suspension of 2 . 5×105 spores/mL in quarter-strength potato dextrose broth ( 6 g/L ) . Five-microliter droplets of spore suspension were deposited on six leaves per plant , with eight to 12 plants per experiment , and lesion diameters were measured at 3 d postinfection . Pst infection was performed as described by [96] . Briefly , Arabidopsis plants were sprayed with bacterial suspensions carrying the EDV construct generated by Fabro et al . ( 2011 ) [13] ( supplemented with 0 . 05% Silwet L-77 ) . Plants were then covered with a transparent lid for 48 h . Infected leaf samples were collected at 4 DAI , ground in sterile 10 mM MgCl2 , serially diluted , and spotted on NYG or low-salt LB ( Luria-Bertani ) agar medium containing appropriate antibiotics . Numbers of colonies were counted after 2 d of incubation at 28°C . For SATI assay , 5-wk-old Arabidopsis plant were used . Leave disks were equilibrated in water in the dark overnight , and the solution was changed for 200 µM SA ( Sigma ) in the morning . After 8 h of incubation with SA or mock , leaf disks were quickly dried and flash-frozen in liquid nitrogen . About 20 leaf disks per condition were used for RNA extraction . For transient assay analysis , A . tumefaciens strains GV3101 and GV3103 were used to deliver respective transgenes in N . benthamiana leaves , using methods previously described [97] . Protein stability was assessed using Western blot , as described by [98] . For stable expression in planta of selected candidates , Arabidopsis WT ( Col-0 ) plants were transformed using the dipping method [99] . Briefly , flowering Arabidopsis plants were dipped with A . tumefaciens carrying a plasmid of interest , and the seeds were harvested to select the T1 transformants on selective GM media . T1 plants were checked for expression of the construct of interest either by fluorescence microscopy and/or by Western blot analysis . T2 seeds were sown on selective GM media , and the proportion of resistant versus susceptible plants was counted in order to identify lines with single T-DNA insertion . Transformed plants were transferred to soil and seeds collected . For each construct , three independent transformed plants were analyzed . T3 homozygotes plants were used for in vivo confocal microscopy and pathotests . Frozen plant tissues were ground to a fine powder in liquid nitrogen using a precooled pestle and mortar . The powder was immediately transferred to a 1 . 5 ml tube and rapidly frozen in liquid nitrogen . Batches of 12 samples were thawed on ice , and 1 ml Tri-Reagent ( Sigma ) was added to the tubes and incubated at room temperature for 10 min . The solution was centrifuged for 20 min at 12 , 000× g , and the supernatant was transferred to a clean tube containing an equal volume of isopropanol . The tube was incubated overnight at −20°C and centrifuged for 10 min at 12 , 000× g , 4°C . Pellets were washed with 70% ethanol , air dried , and resuspended in RNase-free water . The yield and integrity of the RNAs were assessed by measuring the optical density at 260 nm and 280 nm Micro-Volume UV-Vis Spectrophotometer for Nucleic Acid and Protein Quantitation ( Nanodrop , Thermo Scientific , UK ) and agarose gel . Five micrograms of total RNAs were used for generating cDNAs in a 20 µl volume reaction according to Invitrogen Superscript II Reverse Transcriptase protocol . The obtained cDNAs were diluted five times , and 5 µl were used for 10 µl qPCR reaction , and 10 µl were used for 20 µl PCR reaction . qPCR was performed in 20 µl final volume using 10 µl SYBR Green mix ( Sigma ) , 10 µl diluted cDNAs , and primers . qPCR was run on the CFX96 Real-Time System C1000 thermal cycler ( Biorad ) using the following program: ( 1 ) 95°C , 4 min; ( 2 ) [95°C , 10 s , then 62°C , 15 s , then 72°C , 30 s]×40 , 72°C , 10 min followed by a temperature gradient from 65°C to 95°C , and then 72°C , 10 min . The relative expression values were determined using EF1α ( At5g60390 ) as reference gene and the comparative cycle threshold method ( 2−ΔΔCt ) . Primers were designed using Primer3 with the default settings . For RNA sequencing , total RNAs were extracted using TRI reagent ( Sigma ) and 1-bromo-3-chloropropane ( Sigma ) according to the procedure of the manufacturer . RNAs were precipitated with half volume of isopropanol and half volume of high salt precipitation buffer ( 0 . 8 M sodium citrate and 1 . 2 M sodium chloride ) . RNA samples were treated with DNaseI ( Roche ) and purified by RNeasy Mini Kit ( Qiagen ) according to the procedure of the manufacturers . RNA sequencing was performed as described by [100] . Briefly , total RNAs ( 3 µg ) were used to generate first strand cDNAs using an oligo ( dT ) primer comprising P7 sequence of Illumina flow cells . Double-strand cDNAs were synthesised as described previously [101] . Purified cDNAs were subjected to Covaris shearing ( parameters: intensity , 5; duty cycle , 20%; cycles/burst , 200; duration , 90 s ) . End repairing and A-tailing of sheared cDNAs were carried out as described by Illumina . Y-shaped adapters were ligated to A-tailed DNA and subjected to size selection on agarose gel . The gel-extracted library was PCR enriched and quantified using qPCR with previously sequenced similar size range Illumina libraries . The libraries were sequenced on Illumina Genome Analyzer II . Illumina libraries were quality-filtered using FASTX Toolkit 0 . 0 . 13 with parameters −q20 and −p50 ( http://hannonlab . cshl . edu/fastx_toolkit/index . html ) . Reads containing “N” were discarded , and read qualities were converted from Illumina fastq to Sanger fastq format . The libraries were separated using perfect match to the barcode . The sub-library was artefact-filtered using FASTX-toolkit . Quality-filtered libraries were aligned to the Arabidopsis Col-0 genome sequence ( TAIR10 ) using Bowtie version 0 . 12 . 8 [102] and reads with up to 10 reportable alignments were selected . Unaligned reads from previous steps were used to align to transcript sequences of Arabidopsis Col-0 ( ftp://ftp . Arabidopsis . org/home/tair/Sequences/blast_datasets/TAIR10_blastsets/TAIR10_cdna_20101214_updated ) using Bowtie version 0 . 12 . 8 . Linking of each sequenced read ( Tag ) to gene was carried out using the following considerations: reads aligning to each gene limits were assigned to that gene; reads aligning to genes with overlapping gene limits were split equally between them; and reads aligning to more than 10 genes were discarded . Differential expression analysis was performed using the R statistical language version 2 . 11 . 1 with the Bioconductor [103] package , edgeR version 1 . 6 . 15 [104] with the exact negative binomial test using tagwise dispersions . For co-localisation assays in N . benthamiana , cut leaf patches were mounted in water and analysed on a Leica DM6000B/TCS SP5 confocal microscope ( Leica Microsystems ) with the following excitation wavelengths: GFP , 488 nm; YFP , 488 nm; RFP , 561 nm . For in vivo localisation in Arabidopsis , 10-d-old Hpa-infected seedlings were mounted in water and analysed on a Leica DM6000B/TCS SP5 confocal microscope ( Leica Microsystems ) with the following excitation wavelengths: CFP , 458 nm; GFP , 488 nm; RFP , 561 nm . GUS activity was assayed histochemically with 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid ( 1 mg/ml ) in a buffer containing 100 mM Sodium Phosphate pH 7 , 0 . 5 mM Potassium Ferrocyanide , 0 . 5 mM Potassium Ferricyanide , 10 mM EDTA , 0 . 1% Triton . Arabidopsis leaves were vacuum-infiltrated with staining solution and then incubated overnight at 37°C in the dark . Destaining was performed in 100% ethanol followed by incubation in chloral hydrate solution . Sections were observed with a Zeiss Axioplan 2 microscope ( Jena , Germany ) . Aniline blue staining was used to stain callose structures in plant tissues [105] , which appeared after infection , like ring or encasements of Hpa haustoria , or like dots after Pseudomonas infection or PAMP treatment . Samples ( either Hpa-infected seedlings or leaf disks punctured from PAMP/Pseudomonas-infiltrated leaves ) were cleared in 100% methanol , washed in water , and then stained with aniline blue ( 0 . 05% w/v in 50 mM phosphate buffer pH 8 ) overnight . Samples were observed with a Leica DM6000B/TCS SP5 confocal microscope ( Leica Microsystems ) .
|
The highly conserved Mediator complex plays an essential role in transcriptional regulation by providing a molecular bridge between transcription factors and RNA polymerase II . Recent studies in Arabidopsis have revealed that it also performs an essential role in plant defence . However , it remains unknown how pathogens manipulate Mediator function in order to increase a plant's susceptibility to infection . In this article , we show that a secreted effector , HaRxL44 , from the Arabidopsis downy mildew pathogen Hyaloperonospora arabidopsidis ( Hpa ) , interacts with and degrades the Mediator subunit MED19a , resulting in the alteration of plant defence gene transcription . This effector-mediated interference with host transcriptional regulation perturbs the balance between jasmonic acid/ethylene ( JA/ET ) and salicylic acid ( SA ) –dependent defence . HaRxL44 interaction with MED19a results in reduced SA-regulated gene expression , indicating that this pathogen effector modulates host transcription to promote virulence . The resulting alteration in defence transcription patterns compromises the plant's ability to defend itself against pathogens , such as Hpa , that establish long-term parasitic interactions with living host cells via haustoria ( a pathogen structure that creates an expanded host/parasite interface to extract nutrients ) but not against necrotrophic pathogens that kill host cells . HaRxL44 is unlikely to be the sole effector that accomplishes this shift in hormonal balance , and other nuclear HaRxL proteins were reported by other researchers to interact with Mediator components , as well as with other regulators of the JA/ET signalling pathway . Functional analyses of these effectors should facilitate the discovery of new components of the plant immune system . These data show that pathogens can target fundamental mechanisms of host regulation in order to tip the balance of signalling pathways to suppress defence and favour parasitism .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2013
|
A Downy Mildew Effector Attenuates Salicylic Acid–Triggered Immunity in Arabidopsis by Interacting with the Host Mediator Complex
|
Severe fever with thrombocytopenia syndrome ( SFTS ) is an emerging tick-borne viral disease caused by the SFTS virus ( SFTSV ) from Bunyaviridae that is endemic in East Asia . However , the genetic and evolutionary characteristics shared between tick- and human-derived Korean SFTSV strains are still limited . In this study we identify , for the first time , the genome sequence of a tick ( Haemaphysalis longicornis ) -derived Korean SFTSV strain ( designated as KAGWT ) and compare this virus with recent human SFTSV isolates to identify the genetic variations and relationships among SFTSV strains . The genome of the KAGWT strain is consistent with the described genome of other members of the genus Phlebovirus with 6 , 368 nucleotides ( nt ) , 3 , 378 nt , and 1 , 746 nt in the Large ( L ) , Medium ( M ) and Small ( S ) segments , respectively . Compared with other completely sequenced human-derived Korean SFTSV strains , the KAGWT strain had highest sequence identities at the nucleotide and deduced amino acid level in each segment with the KAGWH3 strain which was isolated from SFTS patient within the same region , although there is one unique amino acid substitution in the Gn protein ( A66S ) . Phylogenetic analyses of complete genome sequences revealed that at least four different genotypes of SFTSV are co-circulating in South Korea , and that the tick- and human-derived Korean SFTSV strains ( genotype B ) are closely related to one another . Although we could not detect reassortant , which are commonly observed in segmented viruses , further large-scale surveillance and detailed genomic analysis studies are needed to better understand the molecular epidemiology , genetic diversity , and evolution of SFTSV . Full-length sequence analysis revealed a clear association between the genetic origins of tick- and human-derived SFTSV strains . While the most prevalent Korean SFTSV is genotype B , at least four different genotypes of SFTSV strains are co-circulating in South Korea . These findings provide information regarding the molecular epidemiology , genetic diversity , and evolution of SFTSV in East Asia .
Severe fever with thrombocytopenia syndrome ( SFTS ) is an emerging tick-borne viral disease characterized by fever , gastrointestinal symptoms , leukopenia , and thrombocytopenia . It was first reported in China in 2010 [1] and was subsequently identified in South Korea and Japan in 2013 [2–4] . The causative agent , SFTS virus ( SFTSV ) , belongs to the genus Phlebovirus in the family Bunyaviridae [1] . Other novel tick-borne phleboviruses , including Heartland virus ( HRTV ) , Malsoor virus ( MV ) and Hunter Island Group virus ( HIGV ) , which are genetically related to but distinctly different from SFTSV , have been isolated from humans and ticks in the United States [5 , 6] , bats in India [7] and ticks in Australia [8] , respectively . Like other members of the genus Phlebovirus , SFTSV contains a tripartite RNA genome consisting of three single-stranded RNA segments of negative polarity , designated large ( L ) , medium ( M ) , and small ( S ) . The L , M , and S segments encode the RNA-dependent RNA polymerase ( RdRp ) , the viral envelope glycoproteins ( Gn and Gc ) and both a nucleoprotein ( NP ) and a nonstructural protein ( NSs ) in an ambisense orientation , respectively [1 , 9] . Although human-to-human transmission of SFTSV through contact with blood and/or body secretions of patients has been reported [10–12] , the virus is generally transmitted to humans by tick bites . Several studies have reported SFTSV isolation or detection from tick species including Haemaphysalis longicornis , Rhipicephalus microplus , H . flava , H . concinna , Amblyomma testudinarium , and Ixodes nipponensis [1 , 13–18] . Further , this virus has also been isolated or detected from domestic animals ( e . g . , cattle , goats , dogs , chickens and cats ) , small mammals such as rodents and shrews [19–22] , and reptiles [23] . Since the first reported fatal case in 2012 in Gangwon Province in South Korea [2] , concern regarding SFTSV has grown as the numbers of SFTS patients has increased annually with 36 cases reported in 2013 , 55 cases in 2014 , 79 cases in 2015 , and 165 cases in 2016 [24] . Moreover , the mean mortality rate of these cases was approximately 21 . 8% . During our previous survey of carrier ticks from affected areas , a single SFTSV strain ( designated KAGWT ) was initially isolated from H . longicornis nymphs collected from the Samcheok-si , Gangwon Province [15] . In addition , we also isolated SFTSV from two recovered- and one fatal-human cases that presented with typical SFTS symptoms . In this study , we analyzed the whole genome sequence of the first tick-derived Korean SFTSV and human SFTSV strains to compare the genetic and evolutionary characteristics between tick- and human-derived Korean SFTSV isolates . Genetic characterization revealed that the tick-derived SFTSV is closely associated with recent KAGWH3 Korean human isolates and that at least four different genotypes of SFTSV strain are co-circulating in South Korea . Taken together , our results suggest that more intensive and continuous surveillance of SFTSV is essential for better understanding of the molecular epidemiology , genetic diversity , and evolution of this virus in East Asia .
Chungbuk National University Hospital received written consent for sample collection from each patient with SFTSV infection . All participants were adults and this study was approved by the institutional review board ( IRB ) of Chungbuk National University Hospital ( IRB no . 2015-08-009-001 ) . The KAGWT strain was isolated from H . longicornis ticks collected from the Samcheok-si , Gangwon Province in South Korea as described previously [15] . The CB1 , CB2 , and CB3 strains were isolated from the sera of SFTS patients who were hospitalized with typical SFTS symptoms at Chungbuk National University Hospital . For virus propagation , the virus was passaged five times on confluent monolayers of Vero E6 cells ( ATCC No . CRL-1586; American Type Culture Collection , Manassas , VA ) in Dulbecco’s Modified Eagle Medium ( DMEM; Gibco , Grand Island , NY ) containing 8% fetal bovine serum ( FBS; Gibco ) with penicillin ( 100IU/mL ) and streptomycin ( 100μg/mL; P/S , Gibco ) placed in 37°C incubator supplemented with 5% CO2 . Cell culture supernatant was collected after seven days and stored at -80°C as the working virus stock for whole genome sequencing . Viral RNA was extracted from 140 μL of the viral stock using QIAamp Viral RNA Mini Kits ( Qiagen , Hilden , Germany ) according to the manufacturer’s instructions . Single strand cDNA was synthesized using viral RNA and primers specific for each segment using a cDNA synthesis kit ( Cosmogenetech co , Ltd . , Seoul , Korea ) according to the manufacturer’s instructions . For the whole genome sequencing , one to six overlapping PCR fragments covering the SFTSV genome containing full-length L , M , and S segments were amplified by PCR from cDNA using SP-Taq DNA polymerase ( Cosmogenetech , Seoul , Korea ) . PCR for each segment was performed at 95°C for 5 minutes , followed by 45 cycles of amplification consisting of 95°C for 30 seconds , annealing at 53 or 60°C for 30 seconds according to the primer sets for each segment , and 72°C for 2 minutes , with a final extension at 72°C for 5 minutes . The non-coding 5′ and 3′ ends of the viral genome were determined by rapid amplification of cDNA ends ( RACE ) method as described previously [25] . PCR products were then purified using a QIAquick Gel Extraction Kit ( Qiagen ) according to the manufacturer’s instructions and direct sequenced using ABI Prism BigDye Terminator Cycle Sequencing Kits ( Applied Biosystems , Foster City , CA ) and an ABI 3730x1 sequencer ( Applied Biosystems ) at Cosmogenetech Co , Ltd . The nucleotide sequences obtained from this study were assembled using the SeqMan program in the DNASTAR software ( version 5 . 0 . 6; DNASTAR Inc . , Madison , WI ) to determine the complete genomic sequence . Genetic and phylogenetic analyses were conducted by aligning published full-length sequences of SFTSV obtained from China , Japan and Korea isolates , which are available in GenBank , together with the closely related SFTSV sequences obtained from the basic local alignment search tool results . A total of 41 full-length sequences of the L , M , and S segments , including the strains isolated in this study ( Table 1 ) , were included in the analyses . Multiple sequence alignments were performed using the Clustal W algorithm in DNASTAR version 5 . 0 . 6 or MEGA version 6 . 0 [27] . The aligned nucleotide and deduced amino acid sequences were analyzed using the MegAlign program of DNASTAR to compare the sequence homologies and amino acid substitutions . Phylogenetic analyses were performed based on the full-length L , M , and S segments of SFTSVs . Phylogenetic trees were constructed with MEGA version 6 . 0 software using the Maximum Likelihood ( ML ) method based on the Kimura 2-parameter model . The reliability of the ML tree was evaluated by the bootstrap test with 1 , 000 replications . The full-length L , M , and S segment sequences of tick- and human-derived strains determined in this study have been deposited in GenBank under the following accession numbers: KAGWT ( KY273136 to KY273138 ) , CB1 ( KY789433 , KY789436 , and KY789439 ) , CB2 ( KY789434 , KY789437 , and KY789440 ) , and CB3 ( KY789435 , KY789438 , and KY789441 ) strains .
The genomes of tick- and human-derived Korean SFTSV strains analyzed in this study were consistent with 6 , 368 nucleotide ( nt ) , 3 , 378 nt and 1 , 746 nt present in the L , M and S segments respectively , consistent with what has been reported for other members of the genus Phlebovirus [9] . The L segment encodes a 6 , 255 nt long ORF [2 , 084 amino acids ( aa ) ] for the RNA-dependent RNA polymerase gene , the M segment comprises a 3 , 222 nt precursor of the glycoprotein gene , coding for Gn ( 516 aa ) and Gc ( 511 aa ) proteins , while the S segment contains 882 nt and 738 nt long ORFs , which translate into a nonstructural protein ( 293 aa ) and a nucleoprotein ( 245 aa ) , respectively . The non-coding regions ( NCRs ) of the L , M and S segments at the 5′ termini are 16 nt , 18 nt and 43 nt , respectively; and at the 3′ termini are 97 nt , 138 nt and 29 nt , respectively . As shown in Fig 1 , complementary sequences within the 5′ and 3′ NCRs of the three segments are highly conserved between tick- and human-derived Korean SFTSV strains . Alignment and pairwise comparisons of each segment between the tick-derived KAGWT strain and other fully sequenced human-derived Korean SFTSV strains showed nucleotide ( and deduced amino acid ) sequence homology ranging from 95 . 9 ( 99 . 3 ) to 99 . 9 ( 100 ) , 94 . 0 ( 98 . 4 ) to 99 . 8 ( 99 . 8 ) , and 94 . 9 ( 98 . 3 ) to 99 . 9 ( 100 ) % for L , M and S segments , respectively ( Table 2 ) . This high homology between the tick-derived KAGWT and other human-derived SFTSV strains reflects the close level of relatedness of these strains to each other . In particular , the KAGWT strain showed the highest sequence identities at the nucleotide and deduced amino acid levels with the KAGWH3 strain isolated in 2014 from the serum of a patient from Gangwon Province , South Korea who experienced typical SFTS symptoms [25] . These results indicate that tick- and human-derived Korean SFTSV strains are most closely related with one another . Moreover , the deduced amino acid sequence of KAGWT revealed two amino acid variations ( A66S , I89V ) in the Gn protein compared with the KAGWH3 strain . In particular , one unique amino acid substitution in the Gn protein , at position 66 ( alanine to serine ) , was found only in the KAGWT strain compared with other human SFTSVs . According to previous reports , a change from phenylalanine to serine at position 330 in the M segment polyprotein ( F330S ) occurred in cell culture-adapted SFTSV and resulted in the large-focus phenotype [28] . However , all SFTSV strains analyzed in this study had a conserved amino acid sequence ( F ) at position 330 of the M segment . Compared with the other fully sequenced SFTSV strains , KAGWT showed higher identity with the genotype B SFTSV strains than with the other genotypes at each segment as shown in S1 Table . The nucleotide sequence identities between the tick-derived KAGWT and other SFTSV strains from China , Japan , and South Korea belonging to the genotype B were 96 . 4 to 99 . 9% ( L ) , 95 . 8 to 99 . 8% ( M ) , 94 . 9 to 99 . 9% ( NP ) and 95 . 8 to 99 . 8% ( NSs ) similar , respectively . However , KAGWT showed relatively low nucleotide sequence identity ranging from 95 . 7 to 96 . 3% ( L genes ) , 93 . 6 to 94 . 4% ( M genes ) , 95 . 1 to 96 . 2% ( NP genes ) and 94 . 8 to 95 . 9% ( NSs genes ) , respectively with other genotypes of SFTSV ( Table 2 and S1 Table ) . In comparison with the other fully sequenced SFTSV strains , CB2 was identified as the first genotype A SFTSV strain out of 14 full-length sequenced Korean SFTSV isolates . Genetic comparison results showed that the CB2 strain had nucleotide sequence identity ranging from 97 . 9% ( L ) , 97 . 9 to 98 . 0% ( M ) , 98% ( NP ) and 97 . 2% ( NSs ) , respectively with the other genotype A SFTSV strains circulating in China . However , the CB2 strain showed relatively low nucleotide sequence identity ranging from 96 . 1 to 96 . 7% ( L genes ) , 93 . 5 to 96 . 2% ( M genes ) , 95 . 3 to 97 . 6% ( NP genes ) and 93 . 9 to 96 . 8% ( NSs genes ) , respectively with other SFTSV strains ( Table 2 and S1 Table ) . To investigate the genetic evolutionary origins and relationship between the tick- and human-derived Korean SFTSV strains , phylogenetic analyses were conducted with full-length complete sequences of previous SFTSV isolates from China , Japan , and South Korea . Phylogenetic analyses of complete genome sequences ( L , M , and S segments ) indicated that the tick-derived Korean SFTSV strain ( KAGWT ) was clustered with the genotype B SFTSV strains circulating in humans in China , Japan , and South Korea ( Figs 2–4 ) , although KAGWT also exhibited a close relationship with the KAGWH3 human isolate [25] . Furthermore , the CB1 and CB3 human isolates also clustered with genotype B viruses although they are separated into a different sub-node from tick-driven Korean SFTSV strains ( KAGWT ) and are closely related with recent Korean SFTSV human isolates , KADGH and KACNH3 . In addition , the CB2 human isolate was clustered together with recent China genotype A viruses . Overall the phylogenetic trees revealed that Korean SFTSV strains can be classified with high branch support into four distinct genotypes ( designated A , B , D , and F ) out of the six genotypes described previously [26] . To investigate the prevalence of each genotype of SFTSV , we analyzed the SFTSV sequences available in the GenBank database . As shown in Table 3 , most SFTSV strains from South Korea belong to genotype B ( 11 out of 14 isolates ) and only one isolate each was reported from genotype A , D , and F . It is noteworthy that the CB2 is the first report case of a genotype A virus in South Korea . Isolates from China were diverse with all six genotypes being represented , although the most prevalent was genotype F followed by genotypes A and D ( Table 3 ) . In contrast , only three SFTSV strain genotypes ( B , C , and D ) observed in Japan and the most prevalent genotype was B .
SFTS is an emerging tick-borne infectious disease caused by a novel Phlebovirus , SFTSV that is highly endemic in China , Japan , and South Korea [1 , 2 , 4] . In this study , we determined the whole genome sequences of the first tick and patient-derived Korean SFTSV strains and compared them with other available whole genomic SFTSV sequences . To the best of our knowledge , this is the first report of the whole genomic sequence of an SFTSV strain isolated from ticks and of a genotype A SFTSV strain collected from South Korea . Comparison of the whole genome sequence of tick- and human-derived strains revealed that all SFTSVs consist of three segments with 6 , 368 nt in the L segment , 3 , 378 nt in the M segment , and 1 , 746 nt in the S segment . Thus , the genome organization of Korean strains are consistent with the known genome organization of SFTSV belonging to the Phlebovirus genus [9] . Complementary sequences within the 5′ and 3′ NCRs of the three segments analyzed in this study are highly conserved , as in other SFTSV strains ( Fig 1 ) . According to previous studies of Bunyaviruses , these complementary sequences form panhandle-like structures and may have important roles in transcription initiation , viral RNA replication , viral RNA encapsidation , and viral genome packaging [29 , 30] . All Korean SFTSV strains have 53 unique amino acid variations in the L , Gn , Gc , N , and NSs proteins compared with other SFTSV strains . Among them , one unique amino acid substitution in the Gn protein at position 66 ( alanine to serine ) was found only in the tick-driven KAGWT strain . Since little information is known about SFTSV , additional research regarding the association of amino acid variations in each gene product with the pathogenesis of SFTSV infection and detailed molecular studies utilizing reverse genetics will be required to further explain the functional role of these unique substitutions and in particular , the Gn protein substitution ( 66S ) . Genetic and phylogenetic analysis revealed that the tick-driven KAGWT strain isolated from H . longicornis nymphs in 2013 [15] has high sequence identity in each segment and is closely related with the KAGWH3 strain isolated in 2014 from the serum of an SFTS patient who experienced high fever , vomiting , diarrhea , and fatigue . This is not surprising given that both viruses were from the Samcheok-si , Gangwon Province [25] , and further suggests that tick- and human-derived Korean SFTSV strains are closely related to each other . Although several studies about SFTSV classification have been reported [26 , 31–34] , the uniform classification of SFTSVs has not yet to be established . Therefore , to use unified nomenclature of SFTSV genotypes , more mutual understanding and discussion are needed . In this study , we adapted the nomenclature as previously described by Fu et al . [26] which described results of most large numbers ( 205 SFTSV strains ) and complete sequences . That paper showed SFTSV strains can be divided into six distinct genotypes . Through full-length sequence analysis of human-driven SFTSV isolates we detected the first genotype A strain ( CB2 ) from a patient exhibiting severe SFTSV-like symptoms in the Chungbuk Province of South Korea in 2015 . The Korean strains belong to four of these ( genotypes A , B , D , and F ) of which genotype B strains appear to be predominant . It should be noted that in China all six genotypes of SFTSV have been reported , while only one isolate was reported for each genotype A , D and F in Korea . Thus , further detailed surveillance of tick- and human-derived SFTSVs are needed to understand the actual prevalence and possible transmission of each genotype of SFTSV strain into South Korea . Due to the segmented nature of the Phlebovirus , genetic reassortment is known to be an important process resulting in the sequence diversity necessary for viral evolution [35] . Previous studies have shown reassortment occurs in Phleboviruses including Hantavirus [36 , 37] , Rift Valley fever virus ( RVFV ) [38] , and SFTSV [26 , 31–34] . Although SPL087A and AHL/China/2011 SFTSV strains isolated in Japan and China , respectively , were identified as reassortants which contained different genotype segments in the same SFTSV strain [26 , 31–34] , no reassortant was identified in the Korean SFTSV strains tested in this study . Therefore , continuous analysis consisting of in-depth surveillance and genome sequencing will be needed in South Korea to obtain more detailed information of the molecular epidemiology , genetic diversity , and evolution of SFTSV . In conclusion , we report here the first whole genomic sequence of the tick-derived SFTSV KAGWT strain isolated from the ticks in the Gangwon province of South Korea and the comparison of the genetic characteristics of this virus with recent human SFTSV isolates . Genetic and phylogenetic analyses with full-length genome sequences revealed that tick- and human-derived Korean SFTSV strains are closely related to one another and that at least four different genotypes of SFTSV are co-circulating in South Korea . These results provide insight into the genetic origins of human of SFTSV strains as well as shed light on the molecular epidemiology , genetic diversity , and evolution of SFTSV . Furthermore , these associations will have important implications for the design of diagnostic procedures and vaccines against SFTSV .
|
Severe fever with thrombocytopenia syndrome ( SFTS ) is an emerging tick-borne viral disease caused by the SFTS virus ( SFTSV ) . During entomological surveillance of SFTSV infection in Korean ticks collected from SFTS outbreak areas , we isolated a single SFTSV strain which we designated KAGWT . In addition , we isolated three SFTSVs from human patients with typical SFTS symptoms . In this study , we report the genomic sequences of each of these isolates and compare the genetic and evolutionary characteristics between tick- and human-derived Korean SFTSV isolates . Genetic and phylogenetic analyses of these sequences revealed that the tick-derived Korean SFTSV strain is clustered into genotype B , the most prevalent genotype in South Korea , and was closely related to other SFTSV in the same group . Furthermore , our results show that at least four different genotypes of SFTSV strains are co-circulating in South Korea .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"sequencing",
"techniques",
"taxonomy",
"amino",
"acid",
"sequence",
"analysis",
"phylogenetics",
"data",
"management",
"phylogenetic",
"analysis",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"computer",
"and",
"information",
"sciences",
"sequence",
"alignment",
"bioinformatics",
"evolutionary",
"systematics",
"molecular",
"biology",
"evolutionary",
"genetics",
"nucleotide",
"sequencing",
"dna",
"sequence",
"analysis",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"human",
"genetics",
"evolutionary",
"biology"
] |
2017
|
Molecular genomic characterization of tick- and human-derived severe fever with thrombocytopenia syndrome virus isolates from South Korea
|
New approaches and tools were needed to support the strategic planning , implementation and management of a Program launched by the Brazilian Government to fund research , development and capacity building on neglected tropical diseases with strong focus on the North , Northeast and Center-West regions of the country where these diseases are prevalent . Based on demographic , epidemiological and burden of disease data , seven diseases were selected by the Ministry of Health as targets of the initiative . Publications on these diseases by Brazilian researchers were retrieved from international databases , analyzed and processed with text-mining tools in order to standardize author- and institution's names and addresses . Co-authorship networks based on these publications were assembled , visualized and analyzed with social network analysis software packages . Network visualization and analysis generated new information , allowing better design and strategic planning of the Program , enabling decision makers to characterize network components by area of work , identify institutions as well as authors playing major roles as central hubs or located at critical network cut-points and readily detect authors or institutions participating in large international scientific collaborating networks . Traditional criteria used to monitor and evaluate research proposals or R&D Programs , such as researchers' productivity and impact factor of scientific publications , are of limited value when addressing research areas of low productivity or involving institutions from endemic regions where human resources are limited . Network analysis was found to generate new and valuable information relevant to the strategic planning , implementation and monitoring of the Program . It afforded a more proactive role of the funding agencies in relation to public health and equity goals , to scientific capacity building objectives and a more consistent engagement of institutions and authors from endemic regions based on innovative criteria and parameters anchored on objective scientific data .
The World Health Organization ( WHO ) classifies diseases as Type I , Type II and Type III , which largely corresponds to Global , Neglected and Most Neglected diseases in the vocabulary of the international organization Medécins Sans Frontières ( MSF ) [1] , [2] . Type I/Global diseases know no geographic boundaries while Type II–III/Neglected-Most Neglected are predominantly or exclusively prevalent among populations of developing countries . Types II and III diseases ( from now on “neglected diseases” ) , being prevalent in poor regions , are not prioritized by pharmaceutical and biotechnological industries responsible for the manufacture of goods such as vaccines , drugs and diagnostic kits . This generates what is known as ‘market failures’ - the inefficient allocation of products and services through usual free market mechanisms . Several procedures have been suggested to cope with the three types of “health failures”: ( i ) Science failures ( insufficient knowledge prevents the development of health products such as malaria and HIV vaccines ) : Stimulate basic , fundamental research and technological development , ( ii ) Market failures ( high prices prevent access of drugs by needy populations ) : Price reduction policies ( resulting e . g . from negotiations between governments and industry ) or creating subsidizing mechanisms leading to lower prices and ( iii ) Health service failures ( inexpensive drugs do not reach the patients ) : Fighting corruption , reducing inequalities and coping with cultural , religious or infrastructure barriers , etc . that prevent access to cheap or free drugs by poor countries [3] , [4] . Several initiatives have recently been proposed to stimulate research , technological development and production of vaccines , drugs and diagnostics for neglected diseases by both Big Pharma and Small Biotech of developed countries such as “Push” mechanisms , like Public Private Partnerships ( PPPs ) or Partnerships for the Development of Products ( PDPs ) , funded in general by philanthropies or governments [5] , [6] and “Pull” mechanisms , like Advance Market Commitments ( AMCs ) , Orphan drug legislation ( e . g . the US Orphan Drug Act of 1983 ) and Priority Review Vouchers issued under the Food and Drug Administration Amendments Act of 2007 ( FDAAA ) . These mechanisms have in general been conceptualized and implemented by the developed world and either international or philanthropic organizations . They do not take full advantage of the brainpower and infrastructure existing in middle-income developing countries or in some innovative developing countries ( IDCs ) [7] such as Brazil , where considerable progress has recently been made in defining and implementing a national policy for science , technology and innovation in health [8] , [9] , [10] . Research and development on neglected diseases is one of the key strategic areas of Brazil's priority agenda for health research [8] , [11] . In 2005 the Ministry of Health together with the Ministry of Science and Technology , through their funding agencies DECIT ( Department of Science and Technology , http://dtr2001 . saude . gov . br/sctie/decit/index . htm ) and CNPq ( National Council for Scientific and Technological Development , http://www . cnpq . br/english/cnpq/index . htm ) , launched a joint Program to support research , technological development and innovation on six diseases that disproportionately hit poor and marginalized populations in Brazil: dengue , Chagas disease , leishmaniases , leprosy , malaria and tuberculosis . In 2008 schistosomiasis was added to the list and a 2nd call for applications instituted ( http://www . cnpq . br/editais/ct/2008/034 . htm ) . For additional detais on this DECIT/CNPq Program see Serruya et al [11] , [12] . As equity and capacity building were considered critical components of the Program , it was decided to invest at least 30% of the financial resources in the three Brazilian geographic Regions where these diseases are still prevalent ( North , Northeast and Center-West ) . Since the scientific productivity related to neglected diseases is less than in other areas of health sciences and several institutions located in these Regions are still maturing , traditional indicators such as number of scientific articles and impact factor of the journals where they were published would be of only limited value . We therefore decided to develop new approaches and criteria based on social network analysis [13] , [14] , [15] , [16] , to allow for a fair and efficacious allocation of resources without losing sight of scientific standards .
Publications by Brazilian authors on the seven diseases were retrieved as raw data files from the ‘Web of Knowledge’ database of the Institute for Scientific Information ( ISI ) , a database that lists the full addresses of all authors of every paper . Queries were made in ‘advanced search’ mode directed simultaneously at the country name and at words in the titles of the papers , e . g . [CU = Brazil AND TI = ( Chagas OR cruzi ) ] to retrieve papers with at least one researcher from Brazil among the authors and having “Chagas disease” or “Trypanosoma cruzi” in the title . The ISI raw data files were imported into the text-mining software VantagePoint ( http://www . thevantagepoint . com ) with the appropriate ISI filters . A process of standardization was carried out to bring together the various different names of a particular author or institution [17] and VantagePoint thesauri for names and addresses were created in order to process additional name and address lists . Co-occurrence matrices of authorship data were built into VantagePoint and exported to UCINET software for social network analysis [18] . A co-occurrence matrix shows the number of records in the dataset containing two given list items . Symmetrical , co-occurence matrices ( also called ‘adjacency matrices’ ) were created using the same set of authorship data in rows and columns in order to map co-authorships between authors ( authors×authors matrices ) or institutions ( institutions×institutions matrices ) . For additional details on the use of matrices in social network analysis see for instance Chapter 3 of Scott [19] , “Handling Relational Data” . Networks were assembled , visualized and analyzed for several parameters such as network components and cut-points with the softwares NetDraw or Pajek [20] which are embedded in the UCINET package .
The scientific environment where the Program is based and operates can be assessed analyzing the scientific productivity of Brazilian authors and institutions in peer-reviewed international journals . Table 1 and Figure 1 display that it varies widely among the diseases covered by the Program , being for instance 5-fold greater for Chagas disease and leishmaniases as compared with dengue and leprosy . Co-authorship network analyses were carried out at several stages of the two phases of the Programme: Phase I included six diseases , funding projects during the biennium 2007–2008 and the ongoing Phase II addresses seven diseases during the 2009–2010 biennium . We decided to focus our attention on network components and network cut-points , basic elements of social network analysis [19] , [21] that generate visual information readily useful for Program managers and decision-makers . In this way we emphasized the generation of graphical displays over a purely quantitative , numerical analysis . A component of a network is a portion of the network in which all actors are connected , directly or indirectly , by at least one tie ( one co-authorship in the present work ) [21] . Fig . 2 shows the component analysis of the 2001–2008 dengue co-authorship network , where 172 authors are distributed among 9 components , each one addressing in isolation its own set of specific , complementary or overlapping research topics and subjects . A cut-point of a network is an actor ( author or institution in our case ) whose removal would increase the number of components by dividing the sub-graph into two or more separate subsets between which there are no connections . Cut-points are therefore pivotal points of articulation between the elements that make up a component [19] . The role of cut-points is exemplified in Fig . 3 , which shows the 2006–2007 tuberculosis institutional co-authorship network with the cut-point nodes labeled and identified as red squares . In this network , for instance , the removal of the cut-point “Inst . Trop . Med . Prince Leopold” would disconnect FURG and IVIC from the network and the removal of the cut-point “UNICAMP” would do the same for the University of Illinois . The visualization of this network also demonstrates the power of graphic displays to rapidly detect and emphasize unique features of a given network . In this figure , the large agglomerate of nodes at the upper left immediately stands out , drawing one's attention to the presence of a publication involving a large number of coauthors and their institutions , an indicator of projects involving global networks .
Traditional scientific production indicators routinely adopted as criteria for evaluating scientific proposals and research funding programs , such as the number of publications in a given period of time and impact factor or H-index [22] , have intrinsic shortcomings [23] , [24] and are of limited value beyond ‘Mode 1’ of knowledge production ( disciplinary , primarily cognitive , context ) [25] or when the publication output of the work field or the scientific community under consideration is of small size . In fact a ‘Catch-22’ type challenge ( a no-win situation or a double bind dilemma , see http://en . wikipedia . org/wiki/Catch-22 ) arises when considering these indicators to select candidates eligible for capacity building purposes , as the researchers and institutions most in need of support are exactly those who have a modest scientific curriculum or performance . Traditional evaluations therefore become a real barrier to career progress or towards institutional development . The management of the DECIT/CNPq Program , having received the double mandate to adhere to high scientific standards and strengthen capacity in the less developed Regions of the country , as two pillars of the initiative , realized that new strategies and indicators would be needed . The 2001–2008 survey of publications shown in Tables 1 and 2 well illustrates some of the challenges the Program would face , for instance: ( i ) three out of the four most active research communities ( Chagas disease , schistosomiasis and tuberculosis ) are located in the developed South and Southeast of Brazil , far from the target Regions for capacity building and ( ii ) dengue , a disease that has caused serious problems for public health in recent years , has been addressed by one of the smallest scientific communities and needs ‘fast-track’ capacity building actions . Two social network analysis tools proved to generate particularly valuable information for the strategic management of the Program , the identification of components and cut-points of the co-authorship networks: Component analysis generates a picture of the overall network structure , revealing how fragmented it is and therefore providing valuable information on its status and opportunities for strategic management . As shown in Fig . 2 for the dengue researcher's network , the analysis of the work areas of the nine individual components , based on article keywords , suggested for instance , a collaboration between component III and VIII , as their researchers were all working on dengue vector control but did not engage in formal collaborations . The identification of network cut-points became a very important analytical tool for the management of the Program , particularly in relation to its capacity building/strengthening mandate . As the majority of institutions in the less developed Regions still need to mature , a selection based exclusively on scientific productivity would place them at a clear disadvantage in comparison with sister institutions from the developed Southeast and South . We realized that institutions acting as network cut-points were critical key players as they were responsible for keeping several institutions from these Regions in the loop and should therefore be considered as fundamental partners for training , capacity building and institutional strengthening . This reasoning is supported by work in other fields that made evident the importance of scientists who play roles as brokers for communications among others [14] , the function of nodes critically involved in connecting or bridging modular subregions of a network [26] or the cruciality of ‘creative elements’ in cells , social networks and ecosystems [27] . Table 2 shows that by adopting this cut-point criterion to help the selection of institutions worth strengthening , nine ‘cut-point institutions’ could be added to the eleven ‘top-10 institutions’ identified by classical high-productivity criteria . The Program could therefore double the number of potential investment targets in the North , Northeast and Center-West Regions with objective science-based parameters: the traditional , productivity-based indicators together with the new ones derived from the network analysis proposed in this article . The Program was shaped to operate in ‘Mode 2’ of knowledge production ( broader , transdisciplinary social and economic contexts ) [25] as its mission goes beyond academic goals to also address capacity building , institution strengthening , product development , disease control and public health . In Brazil's national health system ( SUS - Sistema Único de Saúde ) the participation of civil society and communities is assured at all levels of government - federal , state and municipal [28] . The process leading to the prioritization of R&D on neglected diseases , which made possible the launching of the DECIT/CNPq Program and set its main objectives and goals , involved strong participation of these key stakeholders e . g . at the National Health Council ( http://conselho . saude . gov . br/apresentacao/index . htm ) and at the 2nd National Conference on Science , Technology , and Innovation in Health held in 2004 which involved 15 , 000 participants [8] . Mobilizing the scientific community , disease control managers and policy/decison-makers to collaborate under the umbrella of this initiative required a sort of ‘cultural change’ from everyone involved . For this purpose the process adopted by the Program included: ( i ) Organizing priority setting workshops with equal representation by researchers , policy/decision-makers and managers interested in the seven diseases of the Program; ( ii ) Adopting guiding principles such as burden of disease and classical/network-based science indicators as the basis for workshop agendas and discussions; ( iii ) Structuring these workshops on disease-specific working groups with equal representation of policy/decision-makers , managers and scientists of high productivity and/or affiliated to network cut-point institutions; ( iv ) Mobilizing the participation of scientific communities through ‘Call for Applications’ based on the recommendations of the working groups and published in the websites of the funding agencies; ( v ) Peer reviewing of the proposals taking into account the need to allocate a minimum of 30% of the funds to projects submitted by principal investigators affiliated to institutions in the North , Northeast and Center-West Regions . Fig . 1 suggests that the DECIT/CNPq Program has been successful in stimulating scientific productivity on the six diseases in its first phase which did not include schistosomiasis as one of the targets . The future assessment of the full impact of the two phases , however , will need a thorough in-depth evaluation exercise based on input , output , outcome and impact indicators addressing scientific , technological and public health goals . Co-authorship network analysis has been employed to evaluate scientific journals [29] , [30] , institutions [31] and collaboration patterns in specific scientific fields [17] . The innovative contribution brought by this analytical approach during the shaping and implementation of the Program will be expanded and become critical when assessing the evolution , performance and robustness of the networks involved . Our results also suggest that co-authorship network analysis could become an important tool for international organizations or partnerships targeting the elimination or eradication of diseases , providing science-based information relevant to strategic analysis and planning . Lessons from past eradication campaigns demonstrated the importance of maximizing the utilization of scarce human and financial resources , functioning within existing health service structures and encouraging research at all levels [32] . Applied to today's planned efforts towards the elimination/eradication of malaria [33] , [34] or neglected tropical diseases [35] , these lessons would mean identifying and engaging health services , researchers and institutions from developed and endemic countries , an immense challenge that co-authorship network analysis could help address , providing a substantial contribution to global health .
|
The selection and prioritization of research proposals is always a challenge , particularly when addressing neglected tropical diseases , as the scientific communities are relatively small , funding is usually limited and the disparity between the science and technology capacity of different countries and regions is enormous . When the Ministry of Health and the Ministry of Science and Technology of Brazil decided to launch an R&D program on neglected diseases for which at least 30% of the Program's resources were supposed to be invested in institutions and authors from the poorest regions of Brazil , it became clear to us that new strategies and approaches would be required . Social network analysis of co-authorship networks is one of the new approaches we are exploring to develop new tools to help policy-/decision-makers and academia jointly plan , implement , monitor and evaluate investments in this area . Publications retrieved from international databases provide the starting material . After standardization of names and addresses of authors and institutions with text mining tools , networks are assembled and visualized using social network analysis software . This study enabled the development of innovative criteria and parameters , allowing better strategic planning , smooth implementation and strong support and endorsement of the Program by key stakeholders .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"science",
"policy"
] |
2009
|
Co-authorship Network Analysis: A Powerful Tool for Strategic Planning of Research, Development and Capacity Building Programs on Neglected Diseases
|
Neurons must faithfully encode signals that can vary over many orders of magnitude despite having only limited dynamic ranges . For a correlated signal , this dynamic range constraint can be relieved by subtracting away components of the signal that can be predicted from the past , a strategy known as predictive coding , that relies on learning the input statistics . However , the statistics of input natural signals can also vary over very short time scales e . g . , following saccades across a visual scene . To maintain a reduced transmission cost to signals with rapidly varying statistics , neuronal circuits implementing predictive coding must also rapidly adapt their properties . Experimentally , in different sensory modalities , sensory neurons have shown such adaptations within 100 ms of an input change . Here , we show first that linear neurons connected in a feedback inhibitory circuit can implement predictive coding . We then show that adding a rectification nonlinearity to such a feedback inhibitory circuit allows it to automatically adapt and approximate the performance of an optimal linear predictive coding network , over a wide range of inputs , while keeping its underlying temporal and synaptic properties unchanged . We demonstrate that the resulting changes to the linearized temporal filters of this nonlinear network match the fast adaptations observed experimentally in different sensory modalities , in different vertebrate species . Therefore , the nonlinear feedback inhibitory network can provide automatic adaptation to fast varying signals , maintaining the dynamic range necessary for accurate neuronal transmission of natural inputs .
Early sensory processing faces the challenge of communicating sensory inputs with large dynamic range to the rest of the brain using neurons with limited dynamic range [1–6] . In electrical engineering , such challenges are met by compressing the inputs using predictive coding [7–11] . Given this , Srinivasan et al . conjectured that the early visual systems of both vertebrates and invertebrates may also implement predictive coding [12] . More recently , predictive coding has also been conjectured to function in cortical visual and auditory processing , and attention [13–16] . A predictive coding circuit attempts to reduce the dynamic range of an input by subtracting a prediction of the current input value–based on past input values–from the actual current input value , and then transmitting only the difference , i . e . the prediction error ( Fig 1 ) . Such a strategy only works if the prediction is good , which requires the existence of stationary ( predictable ) correlations within the input , and the ability of the algorithm to adapt to them . For example , the optimal linear prediction-error filter , that minimizes the relative power of the transmission , depends on the signal-to-noise ratio ( SNR ) of the inputs [7–9] . Indeed , in the invertebrate visual system , Srinivasan et al . observed adaptations to the prediction-error filter in response to changes in the input SNR [12] . This adaptation of the response filters of a neuron to changing input properties has been explored in the literature [2 , 3 , 12 , 17–21] . This includes the predictive coding formulation introduced by Srinivasan et al . [12] , which was also extended to the vertebrate visual system [20] . Indeed , Hosoya et al . [20] demonstrated that predictive coding can explain changes to the response filters of retinal ganglion cells , for adaptations on the time scale of seconds . Other authors [22–24] alternatively postulated that sensory systems must whiten input signals . They showed that varying response filters to whiten the input signals , for different inputs , also matches the observed changes in neurons’ responses . However , the adaptation of early sensory processing circuits must also be very fast , since the input statistics often vary rapidly , e . g . across a saccade [2 , 3 , 25] , or during movements within a complex auditory environment [26] . Many of the adaptations that have been proposed in the literature vary the biophysical properties of neurons; hence , they are thought to occur at a slower time scale than the speed at which input changes occur . Nevertheless , experimental measurements of the linear filters of sensory neurons have shown that they do indeed change rapidly–often as fast as can be experimentally measured [27 , 28] . Therefore , to maintain optimality , linear response filters in models of sensory coding must adapt to changes in the input signal statistics–including the input SNR–rapidly , following each input change , yet prior to the next change . How can a linear filter change at a high rate in response to changes in its input statistics ? The addition of a time-invariant nonlinearity can allow the construction of a circuit that “instantaneously adapts” its linearized responses to input changes [28 , 29] . However , it is not clear if such a circuit exists , and how to construct it . Previous work on the effect of static nonlinearities on neurons has focused mostly on the possible computational functions of a spiking nonlinearity on the output of a single neuron [29–32] , or a multiplicative nonlinearity within a motion detecting circuit [33] . Here , we demonstrate that a network of leaky integrator neurons , with a threshold nonlinearity , is , indeed , able to achieve automatic adaptation to changes in the ratio of the predictable component of an input to its unpredictable component . Since noise is , by definition , unpredictable , one specific case of this would be sudden changes to the input SNR . We first show that , for certain stimulus ensembles , linear leaky integrator neurons can implement linear predictive coding through both a feedback and a feedforward inhibitory circuit . We find the parameters that allow these networks to minimize the output dynamic range . Comparing these implementations , we find that the structure of the adaptation in the feedback inhibitory circuit lends itself to the construction of an automatically adapting filter . Specifically , the addition of a biologically-plausible threshold nonlinearity to the feedback inhibitory circuit allows it to approximate the performance of the optimal linear filter over a range of input SNRs . We compare the responses of a nonlinear predictive coding circuit to available experimental results . The instantaneous changes to the linearized filter of the nonlinear circuit match the fast changes measured to the linear filters of neurons in different sensory modalities , in response to rapid input changes . Hence , our results support the nonlinear feedback inhibitory circuit as a circuit implementation of predictive coding that models the response of early processing in various sensory modalities , facilitating the transmission of rapidly varying , high dynamic range inputs , through slow , low dynamic range neurons .
In the field of adaptive signal processing , predictive coding algorithms have commonly been used for signal compression [7 , 10 , 34] , including in the transmission of telecommunications signals ( e . g . in the GSM standard ) [35 , 36] and in video compression [37–39] . These algorithms compute a prediction of the current input based on previous values of the input , and then subtract it away from the actual current input ( Fig 1A and 1B ) . If the signal possesses some statistical structure , an accurate prediction can be made and the output transmission will have a smaller power than the input ( Fig 1C and 1D ) [7–11] . In a general predictive coding algorithm , acting on an input time series {ft} , the prediction can be constructed in two ways: ( 1 ) from past values of the input signal , or ( 2 ) from past values of the transmitted output of the algorithm . In ( 1 ) , the algorithm is entirely feedforward ( Fig 1A ) . In contrast , in ( 2 ) , the algorithm is feedback ( Fig 1B ) . In each case , we can write the algorithm as: pt=ft−Cfeedforward ( ft−1 , ft−2 , … ) orpt=ft−Cfeedback ( pt−1 , pt−2 , … ) ( 1 ) where pt is the transmitted signal and the predictions ( computed either linearly or nonlinearly ) are C ( ⋅ ) . A crucial property of predictive coding algorithms is that they transmit information losslessly . Specifically , their function is to transmit all the input that they receive including both signal and noise . This is unlike many other algorithms commonly used in neuroscience , which separate signal from noise . Indeed , as structured in Eq ( 1 ) and assuming that there is no communication noise in the transmission process , the output of a predictive coding network can be used to reconstruct the entire input , losslessly ( Fig 1 ) [8 , 11] . Further , this is independent of the linear or nonlinear way in which the prediction is computed ( S1 Text ) . This property is important in our identification of an optimal predictive coding algorithm . To formulate this optimization problem , we define: A class of allowable predictive coding algorithms , within which to identify an optimal algorithm An input ensemble over which the algorithm is optimized . An optimization metric to measure the algorithm’s performance We start by defining the class of linear predictive coding algorithms–predictive coding algorithms in which the prediction is a linear combination of the past inputs: pt=ft−∑i=1∞wi⋅ft−i ( 2 ) We have written this in the feedforward implementation . However , since the equation is linear , this results in no loss of generality; it can be rewritten recursively to obtain the feedback case . Eq ( 2 ) describes a predictive encoding constructed by a class of linear temporal filters , which vary in the way in which they compute the prediction . Each such temporal filter is defined by its set of parameters , {wi} , causally , for all time steps , i , counting backwards through time . We now define the input ensemble , and optimization metric with which we can identify a specific temporal filter ( i . e . specific set of parameters {wi} ) that is optimal . Ideally , one would like to find the optimal filter over the space of natural images . However , given the complexity of this space , we chose to use a subset of such inputs . Natural image amplitudes are distributed , over time , with a power law distribution over temporal frequencies [40] . It is possible to decompose such an input ( up to some high frequency roll-off ) into a sum of several exponentially correlated components , each with a different time constant [41 , 42] . Therefore , we chose an input composed of one such exponentially correlated signal ( with a single time constant , τs ) , combined with uncorrelated noise , combined at a particular SNR , σ ( Methods ) : ft=σ1+σst+11+σεt ( 3 ) This input provides the ensemble over which we can identify an optimal linear predictive coding filter . We believe that this subset of inputs is naturalistic , since it should be possible to combine several input subsets ( constructed with different time constants ) to generalize back to the space of natural images . The final part of the formulation of the optimization problem is a performance metric against which to optimize the filter . Since the goal of applying predictive coding is to reduce the dynamic range required to transmit a signal , a natural measure of performance would be the degree of reduction in the power of the transmitted signal , relative to the input power . We term this the network gain , defined as: Network Gain≡Transmitted PowerInput Power=1t∑i=1t ( pi ) 21t∑i=1t ( fi ) 2 ( 4 ) for pi as defined in Eq ( 2 ) , and fi as defined in Eq ( 3 ) . Ideally , any metric of performance , for a compression algorithm , would include both the degree of compression , and a measure of the information lost due to the compression , e . g . reconstruction error . However , as introduced above , predictive coding algorithms encode inputs losslessly . Hence , any reconstruction error is necessarily zero ( Fig 1C ) . Therefore , we can measure the performance of different predictive coding algorithms just by using the network gain , and we now solve for the optimal linear temporal filter that minimizes this gain over the input ensemble . Finding the linear filter that minimizes the network gain is a specific example of a common optimization in the adaptive signal processing literature [7–9] , but the specific derivation that we utilize is detailed in S1 Text . Briefly , we compute the power of the filter by transforming Eq ( 2 ) into the frequency domain , and then compute the output power by applying the transformed transfer function to the autocorrelation function of the input . We then solve for the optimal values over the set of wi by differentiation . The solution is a function of the two parameters that characterize the input ensemble , the time constant ( τs ) and SNR ( σ ) : wi*=Λ*1−Λ*⋅ ( β⋅ ( 1−Λ* ) ) i ( 5 ) where β=e−1τsandΛ* ( β , σ ) = ( β2−1 ) ( 1+σ ) + ( β2−1 ) β2 ( σ−1 ) 2− ( β2−1 ) ( 1+σ ) 22β2 ( 6 ) It is important to note that: β is dependent only on the correlation time constant of the signal component Λ* is dependent both on the signal , and the SNR Plotting each of these variables , for some values of the input parameters , provides their qualitative structure . First , β varies from 0 to 1 , as the correlation time constant of the signal increases ( Fig 2A ) . Second , for each fixed β , Λ* varies from 0 to 1 , as the SNR increases ( Fig 2B ) . This quantitative value of Λ* varies with β , but the increase from 0 to 1 always holds , qualitatively . Substituting Eq ( 5 ) back into Eq ( 2 ) , we can write down the linear predictive coding filter that minimizes the power of the transmitted signal: pt=ft−Λ*1−Λ*∑i=1∞ ( β⋅ ( 1−Λ* ) ) i⋅ft−i ( 7 ) An interesting property of the optimal linear predictive coding algorithm is its structure in the high noise , low signal regime ( i . e . low SNR in Fig 2B ) . In that regime , Λ* → 0 . Hence , the prediction approaches 0 , and pt → ft . Initially , this might seem counterintuitive , since the predictive coding network is transmitting the entire , noisy input to downstream circuits . However , any prediction of an uncorrelated input will , on average , increase the power of the transmitted output . Therefore , sending out no prediction does indeed minimize the network gain . Further intuition about the values of the parameters is most useful when applied to specific implementations of this algorithm . Therefore , we first show that it is possible to implement Eq ( 7 ) using a circuit of linear leaky integrator neurons . A simple model of a biological neuron is a leaky integrator ( S1 Fig ) [43 , 44] , whose voltage response ( v ) is modeled by an exponential low pass filter ( Methods ) : vt=gsτm∑i=0∞e−iτmvt−iinput ( 8 ) where the time constant of the filter , τm , is the membrane time constant , and gs is the synaptic conductance ( measured as a fraction of the membrane conductance ) . Comparing Eq ( 8 ) to Eq ( 7 ) , we see that the optimal linear predictive coding filter can be implemented by a combination of two leaky integrator neurons , with different time constants . First , in the limit of τm → 0 , Eq ( 8 ) implies that vt∝vtinput . Therefore , the first term in Eq ( 7 ) can be implemented by a neuron with a short time constant . Second , the subtracted prediction in Eq ( 7 ) is simply an exponential weight on the past inputs and , hence , can also be implemented by a neuron with a specific time constant . This structure can be implemented with different circuits , and we explore both feedforward and feedback two-neuron circuit implementations of the predictive coding algorithm . The feedforward and feedback implementations of predictive coding ( Fig 3 ) are each characterized by two parameters: the time constant of the interneuron ( feedforward: α^; feedback: α ) and the loop gain ( feedforward: Γ^; feedback: Γ ) , i . e . the gain of input through the interneuron relative to direct input . For each network , the two parameters may take different values to implement the optimal linear predictive filter ( that minimizes network gain for the given input statistics ) . To derive these values , we first construct a recursive analytical form for each network’s dynamics , following the flow of information around the circuit , and adding a time step delay at the synapses leading into the interneuron ( as diagrammed in Fig 3 ) ( Methods ) . The recursive dynamics of the feedforward circuit ( Fig 3A ) are: {nt=α^ ( nt−1+Γ^ft−1 ) pt=ft−nt ( 9 ) where the interneuron’s time constant is characterized by a discounting factor , α^=e−1/τ<1 , that discounts the voltage from the past time step by a fixed multiple , and the feedforward gain through the loop is Γ^ ( Fig 3A ) . Solving this recursion , we get ( S1 Text ) : pt=ft−Γ^⋅∑i=1∞α^ift−i ( 10 ) Comparing Eq ( 10 ) to Eq ( 7 ) , the feedback inhibitory network will implement the optimal linear prediction filter if α^=β ( 1−Λ* ) andΓ^=Λ*/ ( 1−Λ* ) ( 11 ) For the feedback inhibitory network ( Fig 3B ) , we can identify and solve a similar pair of recursive equations ( Methods ) . This gives us: pt=ft−Γ1−Γ⋅∑i=1∞ ( α⋅ ( 1−Γ ) ) ift−i ( 12 ) where the interneuron’s discounting factor is α and the feedback gain is Γ . Comparing Eq ( 12 ) to Eq ( 7 ) , the feedforward inhibitory network will implement the optimal linear prediction filter if: α=βandΓ=Λ* ( 13 ) Summarizing the dependence of the optimal network parameters on the input statistics ( from Eqs ( 11 ) and ( 13 ) ) makes explicit the differences between the feedforward and feedback models in their adaptation to input changes: Although the resulting linear prediction-error filter changes in the same way for both circuits , the mechanistic difference between the circuits places different demands on the interneurons . For example , consider the changes in response to increasing input SNR ( σ ) , for a fixed correlation of the signal component ( τs ) . For the feedforward network , Γ^ increases without bound , while α^ approaches 0 . In contrast , for the feedback network , Γ approaches 1 ( Fig 2B ) , while α remains fixed to β . Therefore , whereas in the feedforward network both the gain and the interneuron time constant must vary to maintain optimality , in the feedback network , only the feedback gain must vary . Focusing on the interneuron discounting factor , in the feedforward case , the interneuron gets progressively faster as noise increases , as if the interneuron is reducing the time over which it averages the signal to obtain a prediction , given cleaner inputs ( with less noise ) . However , in the feedback case , the interneuron averages over the same time scale , perhaps to provide a matched filter to select the specific correlated signal . This emphasizes the different roles for the interneuron within each circuit . Further , this difference suggests that the feedback network may lend itself more readily to the construction of an automatically adapting nonlinear network–to respond to rapidly varying input SNR . One can imagine changing the output from one component of a circuit using a nonlinearity , as is necessary for adaptation of the feedback network . However , it would seem to be quite difficult to vary the time constant of a neuron , a cellular property , using a nonlinearity , as is necessary for adaptation of the feedforward network . Therefore , we now explore the construction of such a nonlinear feedback circuit . As introduced earlier , a nonlinearity can allow an invariant circuit to automatically change its linearized response to varying inputs [29–33] . Here , we consider the automatic adaptation of a feedback circuit to rapid changes of the input SNR by introducing such a nonlinearity . In the following sections , we demonstrate that its performance is close to optimal . Our analysis of the optimal linear feedback circuit shows that as σ varies from 0 ( pure noise ) to ∞ ( pure signal ) , the feedback gain , Γ , must increase from 0 to 1 . Therefore , to automatically match the optimal filter over a range of input SNRs , the nonlinearity in the feedback inhibitory circuit must increase the strength of the output from the interneuron onto the principal neuron as the noise decreases ( and vice versa ) . Since inputs of different SNR are integrated differentially by the interneuron , we define the shape necessary for the static nonlinearity . Integrating uncorrelated noise is equivalent to a random walk . In contrast , integrating a correlated signal is equivalent to a biased random walk . Hence , on average , the output of a leaky integrator neuron , from an input with greater correlated component ( i . e . higher SNR ) , will be larger in amplitude . Therefore , any automatic adapting nonlinearity–applied to the output of the feedback interneuron–must push the gain towards 0 for small output amplitudes , and pull the gain towards 1 for large output amplitudes . One simple piecewise linear nonlinearity satisfies this requirement: the threshold or rectilinear nonlinearity , which increases linearly , from a fixed threshold ( Fig 4A and 4B; Methods ) . In response to increasing input amplitudes , the linearized feedback gain of the nonlinear circuit increases from a gain of 0 , for small inputs ( below the threshold ) , to a gain of 1 , for large inputs ( colored lines in Fig 4A ) . This precisely matches the range over which Γ must vary , as the SNR of the input changes to minimize the network gain . Also termed a dead-zone nonlinearity by engineers [45] , the rectification nonlinearity is biologically plausible . The nonlinearity ( Fig 4B ) needs to respond symmetrically around 0 , i . e . to both positive and negative inputs . This is biologically unreasonable for any one neuron . However , such a nonlinearity can be constructed by using a pair of neurons , with each receiving half the inputs ( a single sign ) , and having oppositely signed output connections . The rectification can then be implemented through half-wave rectification of each of the neurons’ outputs , either through the neurons’ spiking thresholds , or through the minimum voltage required to release a vesicle [43 , 44] . Therefore , a feedback inhibitory circuit with a rectification nonlinearity applied to the feedback interneuron ( Fig 4B ) seems to be a plausible candidate to perform automatic adaptation , and approximate the minimal network gain , across a range of input SNRs in real neural circuits . To confirm this , we first demonstrate that the nonlinear feedback inhibitory network does , indeed , change its responses as the input properties change , and that the resulting network gain has the qualitative structure necessary for automatic adaptation . To understand the operation of the nonlinear feedback circuit ( Fig 4B ) , independent of the specific value of the threshold , it is convenient to consider its network gain in the frequency domain ( Fig 4C ) . This will provide intuition for the circuits’ response to inputs with different degrees of predictability . Indeed , since low frequency inputs change slowly , they are predictable . In contrast , high frequency inputs ( near the Nyquist frequency ) are unpredictable and , therefore , for the purposes of the feedback circuit equivalent to noise . The optimal linear feedback circuit for each of these input regimes differs . For low frequency ( predictable ) inputs , the optimal linear feedback circuit would set the feedback gain to 1 . In contrast , for high frequency ( unpredictable ) inputs , the optimal linear feedback circuit should set the feedback gain to 0 . We observe that , without changing any parameters , each nonlinear network ( with a specific threshold ) shows network gains that approach those of the Γ = 1 linear network for low frequency inputs , and those of the Γ = 0 linear network for high frequency inputs ( Fig 4C , dotted lines show the performance of the two linear networks ) . This suggests that the rectilinear feedback inhibitory circuit approaches the optimal performance in different regimes of activity , without internal adaptation , i . e . performance of the form necessary for automatic adaptation . Further , the performance is qualitatively independent of the precise value of the threshold , suggesting–as introduced above–the dynamically changing performance is a general property of the shape of the rectilinear nonlinearity . We now demonstrate that this qualitative understanding is also supported quantitatively , for fast varying input statistics: comparing the performance of the nonlinear feedback network against that of optimal linear networks . To compare the quantitative performance of the nonlinear and the linear circuits in the regime where input properties change too fast to allow for parameter adaptation , we define a class of non-stationary inputs . Each such input–termed a mixture–is composed of two components with different SNRs , mixed in time , such that there is one component for a fixed amount of time , and then the second component for the same amount of time . In this way , we are modelling the response of the circuit to an input with a rapid change from one SNR to another , as opposed to a single input , with a fixed SNR . Within this input regime , we compare the nonlinear network to two different types of linear networks: Type 1: The linear network that obtains the minimal network gain over the specific mixture of two SNR inputs , i . e . the minimal network gain for a non-adapting linear predictive coding network . Type 2: The linear network that has sufficient time to adapt separately to optimally transmit each component of the mixture , i . e . this network has the minimal network gain for any linear predictive coding network ( over this specific input mixture ) . We demonstrate ( Fig 5C and 5D ) that the nonlinear network ( red curves ) both: ( i ) outperforms the type 1 linear network ( blue curves ) , and ( ii ) approximates the performance of the type 2 linear network ( dashed black curves ) . In more detail , the response of both the linear and nonlinear networks to a mixture of two input components ( Fig 5A ) is simulated , and the network gain computed ( summarized in Fig 5B ) . By varying the network parameters , we find the network which minimizes the gain for the specific mixture ( Methods ) . To robustly test the performance of the nonlinear circuit , we simulated its response to a mixture composed of components that are as distinct as possible . Therefore , we chose the first component of the mixture to be pure , predictable , correlated signal , and the second component to be unpredictable . As defined earlier , the correlated component is exponentially correlated with a fixed time constant . For the unpredictable component of the mixture , we utilize one of two inputs: ( 1 ) input at the Nyquist frequency , or ( 2 ) Gaussian white noise . Both these inputs are–for the purposes of a nonlinear predictive coding circuit with a non-zero time constant in the feedback neuron–unpredictable . Since the input mixture transitions from one extreme SNR to another , it should provide a strong test of the ability of a fixed nonlinear circuit to respond to a range of input SNRs . We first show that the best linear network of type 1 is outperformed by the nonlinear network ( Fig 5C , and green region in Fig 5E ) . Indeed , when the unpredictable component of the mixture was composed of input at the Nyquist frequency , the improvement of the nonlinear network over the type 1 linear network was particularly large ( 30–40% ) ( Fig 5D ) . When the unpredictable component of the mixture was modelled by the more biologically realistic white noise stimulus , the improvement was smaller , but it was still approx . 20% ( Fig 5F ) . It is important to note that the linear network of type 1 against which we compare the nonlinear network’s performance has the minimal network gain of any such network . We could have used a linear network adapted to the first component of the mixture , and then measured its performance over both components . This would be a natural model for the case where a network was adapted to some input statistic , which changed rapidly , and the network had had insufficient time to adapt to the new statistic . However , the type 1 linear network outperforms any such linear network . Therefore , it provides a strong baseline against which to compare the performance of the nonlinear network . Our results show that the nonlinear network’s improvement over the type 1 linear network persists even if ( a ) the unpredictable component has larger average amplitude than the predictable ( correlated ) component ( Fig 5C–5F ) , or if ( b ) the fraction of non-stationarity within the input is low ( S2 Fig ) . In both cases , the greatest relative improvement occurs when the uncorrelated component is comparatively smaller ( either in amplitude or time ) than the correlated component . However , the improvement persists over a wide range of input mixtures . Therefore , mixtures of noisy inputs with cleaner ones can be dealt with automatically–and effectively–by the nonlinear network , independent of either the amplitudes of each component , or the amount of each component within the mixture ( modelling the sampling distribution over the environment ) . Continuing beyond the improvement over type 1 networks , we next observe that the performance of the nonlinear network approximates the performance of the type 2 linear network ( Fig 5C and diagonal hashed region in Fig 5E ) . The type 2 linear network is allowed to adapt independently to each component of the mixture . Therefore , its performance is the lowest possible network gain , for any linear predictive coding algorithm–assuming sufficient time to adapt to each new input SNR . The observation that the nonlinear network is able to approximate the type 2 linear performance–despite having a fixed set of network parameters–for a range of input mixtures ( with different input signal distributions ) , is precisely the desired , quantitative , demonstration of automatic adaptation . Given this , we explore the potential role of nonlinear feedback inhibitory networks in real sensory systems . Classical experiments , such as the seminal work of J . D . Victor in cat retinal ganglion cells [17] , have shown that the adaptation of neural responses to varying input stimuli occurs nonlinearly . However , these observations do not directly assess the presence of an automatically adapting circuit since , classically , the neuronal responses were measured after the input had been introduced stably , for some time . Therefore , any observed nonlinear adaptation can be explained by the presence of a second , nonlinear estimate of the input statistics , whose output is then , secondarily , used to adapt the measured linear neural response properties . Such a mechanism would have a delay but , given the experimental time scale , such a delay was not constrained . Nevertheless , simulating the responses of the nonlinear feedback inhibitory network shows that its responses do agree qualitatively with the classically observed nonlinear experiments ( S3 Fig ) . To test specifically for the presence of automatic adaptation , experiments must constrain the speed of the change of the neuronal response function: an automatically adapting circuit will respond to a change in the input structure with an adaptive change , on the timescales of neuronal dynamics . Experimental evidence demonstrating these fast changes to neuronal responses has been found , recently , both in the salamander visual system [21 , 27] ( Fig 6A and 6B ) and in the auditory centers of the avian brain [28] ( Fig 7A–7C ) . Therefore , these changes appear to be a general property of sensory systems . We demonstrate that these experimental changes match the expected changes in the response filters , within the nonlinear network . Baccus and Meister [27] found that the linear filters of retinal ganglion cells ( RGCs ) changed as fast as could be experimentally measured–within 200ms of a sudden input change . Further , as the mean input contrast is increased , the filter shifts its maximum weights onto inputs from the more recent past ( Fig 6A and 6B ) [27] . To match the qualitative structure of the observed temporal filter , with a smooth increase from zero to the first peak , we make two biologically reasonable changes to the nonlinear model ( Methods , S4 Fig ) . This modified network produces a smoothly varying temporal filter ( with zero weight at t = 0 ) that can be compared to experiment ( Methods ) . These two ( biologically reasonable ) changes are actually necessary; subsets of this model , with fewer neurons , or fewer time constants , won’t result in a smooth temporal filter ( Methods ) . Given this model , we demonstrate that the resulting response filter for the nonlinear feedback network shifts in the same direction as that measured by Baccus and Meister [27] , in response to sudden input mean changes ( Fig 5C ) . For example , the filter shifts towards more recent inputs ( i . e . speeds up ) in response to a sudden increase in input amplitude . This qualitative agreement between the change in the response filter introduced automatically by the nonlinear feedback network , and the responses of the RGCs , suggests that a nonlinear circuit could underlie the observed fast adaptation . Mante et al . [21] also showed a rapid shift of the linear impulse response function for neurons in the lateral geniculate nucleus ( LGN ) . Their observed change ( Figures 3 and 4 in [21] ) has the same qualitative structure as that observed by Baccus and Meister [27] ( Fig 6B ) , but is observed in response to both changing input contrast and changing input luminance . Since the shift in the filter due to the nonlinear feedback inhibitory network is independent of the specific input , the nonlinear network is able to model the observed filter changes in the LGN . This independence on the precise input also allows the automatic adaptation of the nonlinear feedback inhibitory network to generalize to non-visual sensory modalities . Nagel and Doupe [28] measured changes to the linear filters of auditory neurons in the zebra finch forebrain , which occurred rapidly , within 100ms of changes to the input auditory stimuli ( Fig 7A ) . They quantitatively characterized these changes: In response to increasing input amplitude , the first peak of the temporal filter decreased in amplitude , and the first valley increased in amplitude . Therefore , there was a decrease in the ratio of total positive response of the filter to total negative response , when the input amplitude changes from high to low ( points below the diagonal in Fig 7B ) . The shift in the location of the peaks can be characterized by the change in the peak frequency response of the filter ( best mean frequency , BMF ) . Therefore , comparing high to low amplitude inputs , the authors identified an increase in the BMF ( points above the diagonal in Fig 7C ) . We demonstrate , through simulating the responses of the nonlinear feedback circuit ( Fig 7D ) that the shift in the filter responses agrees qualitatively with the shift in the measured linear responses ( Fig 7A ) . Indeed , as the amplitude increases , one observes not only a shift in the locations of the extrema , but also the addition of a second peak , following the valley , at the highest amplitudes . Further , each of the changes to the filters measured by Nagel and Doupe [28] are a natural property of the nonlinear circuit: The positive/negative ratio of the simulated filter of the nonlinear network decreases as the input amplitude increases ( Fig 7E ) . Further , the simulated ratio at high and low input amplitudes agrees quantitatively with Nagel and Doupe’s measurements ( colored circle in Fig 7B ) . The BMF of the simulated temporal filters of the nonlinear network increases for increasing input amplitudes ( Fig 7F ) . Further , the value of the BMF for the simulated filters at high vs . low input amplitudes agrees quantitatively with the experimental measurements ( colored circle in Fig 7C ) . This suggests that a nonlinear feedback circuit could underlie the observed fast adaptation in the zebra finch auditory forebrain . Importantly , the changes to the response filters of the nonlinear predictive coding network are a general property of the network , and not a function of the specific parameters chosen . Indeed , it is possible to demonstrate analytically that a nonlinear model of neurons with just two time constants , assuming only that the time constant of the interneuron is longer than that of the principal neuron , already shows a shift of its single extremum towards the more recent past ( as input amplitudes increase ) ( S1 Text ) . Therefore , the rectified feedback inhibitory circuit design automatically lends itself to fast changes of the response filters of the output neuron , matching the observed fast adaptations in many different neural circuits .
Neuronal circuits must transmit input signals which vary , rapidly , by multiple orders of magnitude [2 , 3 , 25 , 26] , through neuron channels that have limited dynamic ranges and slow response times [1–5] . This results in two computational challenges: transmitting signals with minimal power , and responding to rapid changes in the input statistics . We demonstrated that an optimal predictive coding algorithm , that reduces the transmission power of a correlated signal ( and thereby ameliorates the first challenge ) , can be implemented with linear leaky integrator neurons using either feedback or feedforward inhibition . Inclusion of a static nonlinearity in the feedback inhibition circuit allows it to approximate the performance of the optimal linear predictive filter , for a range of input SNRs , while keeping its circuit parameters unchanged . Such a circuit , therefore , helps to address the second challenge . Indeed , we showed that the nonlinear feedback inhibitory network’s responses are in agreement with experimentally measured rapid changes to neuronal response functions in different sensory modalities [21 , 27 , 28] . Our analysis distinguishing the feedforward and feedback implementations of the algorithm demonstrated the importance of the neural implementation in developing intuition about an algorithm . For example , in the feedforward inhibitory circuit , it was possible for predictive coding to be implemented by a neuron with a short time constant , for large SNR inputs . However , in the feedback circuit , the predictive neuron is matched to the properties of the signal component , and independent of the SNR . Therefore , each implementation may have differing properties , which may each prove differently useful . This work also demonstrates the necessity of studying inputs with rapidly varying statistics , to understand the different constraints they place on circuit implementations of an algorithm . Both the feedforward and feedback circuits can implement optimal linear predictive coding–when the input is statistically stationary . However , by analyzing the responses of the two different implementations to non-stationary inputs , we found that the feedback implementation provides a natural way to approximate an optimal response , through the addition of a circuit nonlinearity . In contrast , the alternative , feedforward , implementation requires adaptation of its underlying cellular properties , which would be difficult to vary through a circuit nonlinearity . In this report , we introduced intuition on the shape of a nonlinearity necessary to perform automatic adaptation . Given its mathematical convenience , and biological plausibility , we focused on the rectilinear nonlinearity . However , it is important to note that the rectification nonlinearity is not the only nonlinearity that can satisfy the necessary structure . Indeed , any nonlinearity with the necessary inflection point should also be able to perform an automatic adaptation . This provides an avenue for further analysis . Another avenue for further exploration is how to generalize our results on nonlinear predictive coding networks to more complex stimuli . We found the optimal linear predictive coding algorithm for an exponentially correlated signal with a single time constant . However , naturalistic stimuli can be modeled as a combination of several exponentially correlated signals , with different time constants . This suggests that to respond to a naturalistic stimulus , there should be several predictive coding circuits , each adapted to one of the correlations within the signal . However , how could these different predictive coding circuits be combined to respond optimally overall ? One possible solution may be the addition of mutually inhibitory connections between the parallel predictive coding circuits . This circuit design should allow each neuron to respond maximally to the input component that it was adapted for , while simultaneously removing that input component from the remaining neurons . Hence , it might allow the net response of the larger circuit to remain close to optimal . A similar network design has been shown to implement predictive coding across a spatial scene ( for a non time-varying stimulus ) [46] . Exploring circuits of this form , built using simpler building blocks , should prove useful in understanding the design of more complex circuits . Another direction to explore is the computational function of the nonlinear circuit , beyond its linearized responses . In this work , we demonstrated that the network gain of the nonlinear feedback network approximates that of the optimal linear network . However , this does not imply the two networks have identical responses to stimuli . For example , the linear algorithm amplifies high frequency inputs ( flattening the output frequency distribution , for an exponentially correlated signal with noise ) ( S5 Fig ) . In contrast , the nonlinear network reduces the transmission of high frequency inputs ( since they are less likely to cross the threshold ) . This difference , where the nonlinear network may partially denoise the input , hints at additional , computational functions for nonlinear networks , which should be explored further . In general , adaptation of network dynamics , causing them to respond faster when inputs are more salient , has been observed in different experiments [2 , 12 , 17 , 27 , 28] , often as a shifting of the linear kernels . Conceptually , this shift has been understood as the system responding more quickly to salient stimuli , since the relevant information can be extracted sooner [47] . Our results suggest that nonlinear feedback inhibitory circuits may provide a general neural mechanism with which to implement such a shift , automatically . In addition , the presence of fast adaptation in different species suggest that this mechanism may have been conserved over evolutionary history . Finally , one long standing goal of computational neuroscience has been to develop circuit motifs , in a manner similar to electrical engineering . We believe that the nonlinear feedback inhibitory network could be one such neuronal circuit motif . It performs a specific computational function without losing information , and is stable with respect to internal disturbances ( S1 Text ) . Hence , this nonlinear motif could be inserted into a larger computational circuit without affecting its function , while still maintaining stability , and lowering the internal dynamic range requirements . Further , the circuit’s structure is simple and biologically plausible . Hence , it can reasonably be implemented in many areas of the central nervous system . Therefore , we believe that the identification of this motif should provide a useful tool in the analysis of larger circuits in many neural systems .
The input used in optimizing the linear predictive coding circuit ( as in Eq ( 3 ) ) was composed of an exponentially correlated signal and uncorrelated noise , combined with an SNR ( of the power of the input ) of σ . In detail , the signal component of the input , st , was defined as: 〈st〉=0〈stst+i〉=e−iτs≡βi ( 14 ) and the noise , εt , as: 〈εt〉=0〈εtεt+i〉=δ ( i ) ={10i=0i≠0 ( 15 ) Our derivation of the optimal linear predictive coding filter did not require any constraint on the distributions ( for either the signal or noise components of the input ) . Hence , for maximal generality , we left them unconstrained . The predictive coding circuits were constructed with linear leaky integrator neurons ( S1 Fig ) . The voltage dynamics of a single such neuron is described by: v˙ ( t ) =−1τmv ( t ) +gsτmvinput ( t ) ( 16 ) where the synaptic conductance , gs , is measured as a fraction of the cell's membrane conductance . Discretizing Eq ( 16 ) in time , and solving for v , we find that the resulting neuronal building block is an exponential , low pass filter , with time constant τm , as described in the text . The steps to derive the recursive equation governing the dynamics of the feedback circuit Eq ( 12 ) are shown . The corresponding derivation for the feedforward circuit can be found in S1 Text . The discretized feedback circuit is described by a pair of linear , recursive equations . As introduced in the text , we derive this pair of equations by computing the input to each cell in the circuit ( Fig 3A ) , and adding a delay for all inputs to the interneuron . We can ignore the corresponding delay onto inputs to the principal cell , because adding an additional delay to the inputs to pt , i . e . having pt = ft−1 − nt−1 , would not change the structure of the recursion in Eq ( 9 ) . It would simply add an additional time step of delay to everything . This process gives us: pt=ft−nt ( 18 ) nt=α ( nt−1+Γpt−1 ) ( 19 ) Substituting Eq ( 18 ) into Eq ( 19 ) ( at time point , t-1 ) : nt=α⋅nt−1+α⋅Γ⋅ ( ft−1−nt−1 ) =α ( 1−Γ ) ⋅nt−1+α⋅Γ⋅ft−1 ( 20 ) Letting η = α ( 1−Γ ) , and then substituting Eq ( 20 ) , at an earlier time point , back into itself , we get: nt=η⋅ ( η⋅nt−2+α⋅Γ⋅ft−2 ) +α⋅Γ⋅ft−1=η2⋅nt−2+α⋅Γ⋅ ( ft−1+η⋅ft−2 ) ( 21 ) Repeating this process , we have: nt=α⋅Γ⋅∑i=1∞ηi−1⋅ft−i=Γ1−Γ⋅∑i=1∞ ( α ( 1−Γ ) ) i⋅ft−i ( 22 ) Finally , substituting back into Eq ( 18 ) gives the expanded recursion in the text , Eq ( 12 ) . The equation for the rectilinear nonlinearity ( also known as a dead zone nonlinearity in the engineering literature ) is as follows: R ( x ) ={x+δx<−δ0−δ<x<δx−δδ<s ( 23 ) In Fig 3B , we present the Bode plots of both the linear and the nonlinear networks . However , the Bode plot for the nonlinear network is , necessarily , an approximation ( since the response of any nonlinear network is dependent on its input history ) . The approximation used in this plot is obtained through Describing Function analysis [45] , and is commonly used in control theory . In this analysis , we assume that the response of the network to a single frequency of input ( at a single amplitude ) is linear for each such input . This linear model is allowed to vary with each different frequency ( and amplitudes ) . We compute a look-up table for the effect of the nonlinearity by balancing the input across the nonlinear loop , for each frequency ( and input amplitude ) . For every single frequency input to the network , however , it is necessary to make the assumption that the output of the network only produces a single frequency output ( the primary component of the Fourier transform ) . This means that describing function analysis automatically discards any spread of the initial frequency into higher Fourier harmonics . The resulting Bode plots are , therefore , not quantitatively correct . However , it has been well established that describing function analysis does provide a reasonable , qualitatively correct result . Response of optimal linear and nonlinear networks to varying mixtures of stimuli was simulated . To construct these plots , a 1:1 mixture of two input components was used . The first component was pure exponentially correlated signal , and the second was pure noise . Further , the amplitude of the noise component was varied ( values on the x-axis of Fig 5C–5F ) , while the signal amplitude was kept constant at 1 . The response of three different networks , two linear and one nonlinear , was simulated for each input mixture , and the network gains computed . For all three networks , the discounting factor was matched to the time constant of the input within the signal component of the mixture . Parameter variation was then used to find the optimal value of Γ ( for the linear networks ) and both Γ and the threshold δ ( for the nonlinear network ) that allowed the corresponding network to minimize the cost function . For the type 1 linear network , Γ was only optimized once , over the entire mixture . However , for the type 2 linear network , Γ was optimized twice , for each component of the mixture , separately . Finally , for the nonlinear network , the parameters were again only optimized once , over the entire mixture . The resulting network gains , and relative % improvements are plotted in Fig 5C–5F . It was necessary to modify the analytically-derived nonlinear feedback circuit to obtain simulations that can be directly compared to the experimentally measured response filters of different sensory neurons . The experimentally measured filters place a low weight on inputs at t = 0 , with the weight increasing to a peak , followed by a reducing oscillation between peaks and troughs ( Figs 5B and 6D ) . However , with the structure of the network introduced analytically ( diagrammed in S4A Fig ) , it is straightforward to show that the response filter of the corresponding linear circuit applies a maximal weight at t = 0 . Since real neurons must have a non-zero time constant , the first change that we considered to the model , was to add a non-zero time constant to the principal neuron . However , the linearized response filter of this modified model still has maximal weight at t = 0 . An alternative change is to add a time constant to a neuron providing input to the two-neuron network ( S4B Fig ) . Since real neurons are embedded in a larger network of neurons , this is reasonable . However , again , the linearized response filter of this three-neuron model , with two non-zero time constants , has maximal weight at t = 0 . In contrast , combining both the above changes into a single model with three neurons , all with non-zero time constants ( S4C Fig ) , provides a model that shows the same behavior as experimentally measured , with a low weight on inputs from the immediate past , increasing to a peak , followed by a decrease to a trough . This model is a simple , biologically reasonable modification to the original predictive coding model that provides a response filter that is qualitatively similar to experiment . Therefore , we used this model for the in silico comparisons with experiment in Figs 5 and 6 . The response of the nonlinear network , with three neurons with non-zero time constants ( S4 Fig ) , was simulated for >2000 stimulus patterns , each containing 1000 time points of uncorrelated white noise . The stimulus amplitude was varied by varying the standard deviation of the sampling distribution . The linearized filter response of the nonlinear network to the inputs was then estimated by picking a point in time , and weighting the input to the network on each trial by the output of the network at that time point . This method is an analog of the spike-triggered averaging ( STA ) algorithm utilized by experimentalists ( and is only modified by the use of the graded output values , in the absence of an output spiking neuron ) . The filter responses to inputs of different amplitudes could then be compared , and appropriate parameters extracted from the simulated curves .
|
An animal exploring a natural scene receives sensory inputs that vary , rapidly , over many orders of magnitude . Neurons must transmit these inputs faithfully despite both their limited dynamic range and relatively slow adaptation time scales . One well-accepted strategy for transmitting signals through limited dynamic range channels–predictive coding–transmits only components of the signal that cannot be predicted from the past . Predictive coding algorithms respond maximally to unexpected inputs , making them appealing in describing sensory transmission . However , recent experimental evidence has shown that neuronal circuits adapt quickly , to respond optimally following rapid input changes . Here , we reconcile the predictive coding algorithm with this automatic adaptation , by introducing a fixed nonlinearity into a predictive coding circuit . The resulting network automatically “adapts” its linearized response to different inputs . Indeed , it approximates the performance of an optimal linear circuit implementing predictive coding , without having to vary its internal parameters . Further , adding this nonlinearity to the predictive coding circuit still allows the input to be compressed losslessly , allowing for additional downstream manipulations . Finally , we demonstrate that the nonlinear circuit dynamics match responses in both auditory and visual neurons . Therefore , we believe that this nonlinear circuit may be a general circuit motif that can be applied in different neural circuits , whenever it is necessary to provide an automatic improvement in the quality of the transmitted signal , for a fast varying input distribution .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
Automatic Adaptation to Fast Input Changes in a Time-Invariant Neural Circuit
|
Typhoid is an important public health challenge for India , especially with the spread of antimicrobial resistance . The decision about whether to introduce a public vaccination programme needs to be based on an understanding of disease burden and the age-groups and geographic areas at risk . We searched Medline and Web of Science databases for studies reporting the incidence or prevalence of typhoid and paratyphoid fever confirmed by culture and/or serology , conducted in India and published between 1950 and 2015 . We used binomial and Poisson mixed-effects meta-regression models to estimate prevalence and incidence from hospital and community studies , and to identify risk-factors . We identified 791 titles and abstracts , and included 37 studies of typhoid and 18 studies of paratyphoid in the systematic review and meta-analysis . The estimated prevalence of laboratory-confirmed typhoid and paratyphoid among individuals with fever across all hospital studies was 9 . 7% ( 95% CI: 5 . 7–16 . 0% ) and 0 . 9% ( 0 . 5–1 . 7% ) respectively . There was significant heterogeneity among studies ( p-values<0 . 001 ) . Typhoid was more likely to be detected among clinically suspected cases or during outbreaks and showed a significant decline in prevalence over time ( odds ratio for each yearly increase in study date was 0 . 96 ( 0 . 92–0 . 99 ) in the multivariate meta-regression model ) . Paratyphoid did not show any trend over time and there was no clear association with risk-factors . Incidence of typhoid and paratyphoid was reported in 3 and 2 community cohort studies respectively ( in Kolkata and Delhi , or Kolkata alone ) . Pooled estimates of incidence were 377 ( 178–801 ) and 105 ( 74–148 ) per 100 , 000 person years respectively , with significant heterogeneity between locations for typhoid ( p<0 . 001 ) . Children 2–4 years old had the highest incidence . Typhoid remains a significant burden in India , particularly among young children , despite apparent declines in prevalence . Infant immunisation with newly-licensed conjugate vaccines could address this challenge .
Typhoid ( enteric ) fever caused by Salmonella enterica serovar Typhi ( S . Typhi ) is an important cause of morbidity and mortality . The global annual burden was estimated at approximately 12 million cases for 2010 [1 , 2] . Most of these were effectively treated with antibiotics , although the case fatality rate remains at about 1% such that about 130 , 000 typhoid deaths occur annually . Antibiotic resistance is a challenge for effective treatment of typhoid and is likely to become increasingly problematic with the spread of multi-drug resistant strains [3] . The situation is further complicated by increased incidence in some countries of S . Paratyphi A as a cause of enteric fever [4] . This serovar is not prevented by currently available typhoid vaccines and represents an increasing threat to human health . The incidence of typhoid and paratyphoid varies geographically , with south-central and south-east Asia having the highest incidence—typically exceeding 100 cases per 100 , 000 person-years for typhoid and with lower , variable rates for paratyphoid . In one multicenter study , the annual incidence of typhoid per 100 , 000 children aged 5–15 years was 180 in North Jakarta , Indonesia , 413 in Karachi , Pakistan and 494 in Kolkata , India [5] . In the same settings , the annual incidence of paratyphoid was considerably lower , with the highest annual incidence reported from Pakistan of 72 per 100 , 000 children aged 2–16 years [6] . The burden of typhoid fever shows substantial variation within as well as between countries . Commonly identified risk-factors include a lack of clean drinking water , poor sanitation , inadequate hygiene practices and low socio-economic status [2 , 7] . Outbreaks may occur following a defined event of food or water contamination with the bacterium , in which case locally-specific risk factors or exposures may be identified e . g . eating milk products from a sweet shop , [8] . In some instances the originating infection may be a chronic carrier who persistently sheds the bacterium as a result of infection of the gall bladder . Chronic carriage occurs following primary infection in approximately 2–5% of cases in the absence of antibiotic treatment and is strongly dependent on age and sex [9] . However , the contribution of chronic carriers to typhoid transmission in endemic regions is unknown [10] . Several safe and effective vaccines that could help reduce disease burden are licensed and available in India . Three or four doses of orally-administered , live-attenuated Ty21a provide about 50–70% protection for at least 7 years and are licensed in capsule form from 5 years of age or as a liquid formulation from 2 years of age [11 , 12 , 13] . The single-dose injectable Vi polysaccharide vaccine provides similar levels of protection for at least 3 years and is licensed from 2 years of age [11 , 14 , 15] . A Vi polysaccharide conjugated to Pseudomonas aeruginosa exotoxin A ( rEPA ) as a carrier protein and administered to 2–5 year old children gave approximately 90% protective efficacy against typhoid over 4 years [16 , 17] . More recently , two Vi polysaccharide vaccines conjugated to tetanus toxoid have been licensed in India from 3–6 months of age based on their encouraging immunogenicity [18] . The immunogenicity of conjugate typhoid vaccines in children under 2 years of age ( cf . Vi polysaccharide vaccines ) is an important advance , given the significant burden of disease in young children and infants [19] . The WHO recommends the programmatic use of typhoid vaccines for controlling endemic disease , although in most countries vaccinating only high risk populations is recommended [20] . In India , routine typhoid vaccination is not implemented and decision-making has been hampered by the lack of reliable disease burden data with very few prospective surveillance studies in the past two decades . The one exception we are aware of is in Delhi where each year approximately 300 , 000 children aged 2–5 years are vaccinated with Vi polysaccharide vaccine . With the recent development of conjugate vaccines that can be administered to children under 2 years old , the case for more widespread immunization is stronger and in 2014 the Indian Academy of Pediatrics ( IAP ) Advisory Committee on Vaccines and Immunization Practices ( ACVIP ) strongly urged the Government of India ( GoI ) “to include universal typhoid vaccination in its UIP [Universal Immunisation Programme] all over the country . ” [21] The GoI decisions about whether to introduce a public typhoid vaccination program , its extent and the immunization schedule , need to be based on a firm understanding of the disease burden and the age-groups and geographic areas at risk . We therefore carried out a systematic review to estimate the burden of typhoid and paratyphoid in India and to identify knowledge gaps that need further evaluation . We searched for hospital and community-based studies that reported the incidence or prevalence of typhoid and paratyphoid fever and used meta-analysis and meta-regression models to summarize our findings and identify risk factors for disease .
We searched Medline and the Web of Science literature databases for articles published between 1950 and May 2015 for studies on the burden of typhoid or paratyphoid fever in India , with no language restriction . The search consisted of terms related to typhoid or paratyphoid fever ( typhoid OR Salmonella Typhi OR enteric fever OR Salmonella enterica OR paratyphoid OR Paratyphi ) , combined with terms for Indian geography ( including a list of state names ) and terms for measures of incidence and prevalence ( burden OR incidence OR prevalence OR mortality , etc . ) . The complete search term is given in the S1 Appendix . Titles and abstracts of articles were read to identify potentially relevant articles . Studies were considered eligible for further examination in full text if they reported incidence , prevalence , number of reported cases , mortality or the burden of typhoid or paratyphoid in India . Studies were also examined in full text if only a title was returned by the initial search . Full text articles were obtained through online publisher websites , the British Library and the Christian Medical College library in India . We excluded papers reporting a small number of cases ( n < 10 ) , no information about the number of S . Typhi or Paratyphi infections , no laboratory confirmation of infection ( based on culture or serology ) , no distinction between Salmonella serovars , vaccine trials ( unless cluster-randomized with a control arm ) , unknown geographical areas or outside India , or a review of the literature only . If the typhoid burden was reported multiple times for the same region , study population and time period , the study with the longest follow up time was selected . Two reviewers ( CVA and NCG ) independently extracted data from the included studies and entered these into independent Excel databases . Disagreements between the two databases were resolved by consensus among all authors . Year of publication , study design , setting ( hospital or community based study ) , study location , inclusion and exclusion criteria for study participants , start and end date of recruitment , type of samples , laboratory tests , whether the study was an outbreak report , number of participants , number of cases , age distribution of cases and sex of cases were collected . Longitude and latitude information of the study location was obtained from the US National Geospatial Intelligence Agency [22] . The outcome measures of interest were the prevalence of S . Typhi or Paratyphi among patients tested for infection in hospital settings or the incidence of typhoid and paratyphoid fever recorded in community studies . We did not publish our study protocol prior to completing the systematic review . We calculated the proportion of patients with laboratory confirmed typhoid or paratyphoid fever together with Wald 95% confidence intervals calculated on a logit scale for each study reporting data from hospitals [23] . Pooled estimates of prevalence were obtained by combining studies in a binomial regression model with a normally distributed random effect on the intercept . Heterogeneity between the studies was assessed using a likelihood ratio test ( LRT ) comparing a saturated mixed-effects model ( with dummy variables for the random-effects ) with a fixed-effects only model . A stratified analysis was performed due to anticipated heterogeneity between studies , based on the characteristics of the patients included in the studies ( either fever described in the publication as suspected typhoid fever or fever where clinical suspicion of typhoid is either not present or not reported ( hereafter just termed ‘fever’ ) ) . Independent variables potentially associated with the prevalence of typhoid or paratyphoid were included as fixed-effects in univariate and multivariate binomial meta-regression models . The incidence of typhoid and paratyphoid was calculated per 100 , 000 person-years of observation for prospective community-based studies and Wald 95% confidence intervals calculated on a log scale . Pooled estimates of incidence were obtained by combining studies in a Poisson regression model with a normally distributed random effect on the intercept and heterogeneity assessed as above . Independent variables potentially associated with the incidence of typhoid or paratyphoid were included as fixed effects in univariate and multivariate Poisson meta-regression models . Analyses were all performed in the R statistical programming language using the ‘metafor’ package [24 , 25] .
The search strategy initially yielded 1 , 152 records of which 361 were duplicates ( Fig 1 ) . Six hundred and eleven records were excluded after screening the title and abstract . Full text copies were retrieved for 160 of 180 potential relevant records . After excluding non-eligible articles and duplicates , we included 37 studies that reported on typhoid and 18 that reported on paratyphoid . The characteristics of the included studies are given as Table A in the S1 Appendix . Three studies of typhoid were community cohorts while the other 34 were conducted in hospitals , with all but 3 conducted in urban settings . Among the studies conducted in hospitals , 13 included participants with fever and 21 with suspected typhoid fever while all the community based studies included participants with fever . Thirty of the hospital studies and both the community studies reported culture confirmed typhoid , while four studies reported either a combination of culture and serology or serology alone . Studies reporting typhoid based on serology were included only if the serologic diagnostic criteria was clearly described . All of the 18 studies reporting the prevalence of paratyphoid included S . Paratyphi A , and one described both S . Paratyphi A and B . Incidence of paratyphoid was described in two community cohort studies that reported from the same location ( Kolkata ) . The estimated prevalence of S . Typhi detected through culture or serology across all hospital-based studies in the random effects model was 9 . 7% ( 95% confidence interval ( CI ) : 5 . 7–16 . 0% ) ( Fig 2 ) . There was significant heterogeneity in prevalence among studies ( LRT p<0 . 001 ) . Prevalence was higher among participants with suspected typhoid fever ( estimated prevalence in separate random-effects model was 14 . 5% , 95% CI: 8 . 4–23 . 9% ) compared with fever ( estimated prevalence 4 . 9% , 95% CI: 1 . 9–12% ) . This was confirmed in the univariate mixed-effects , meta-regression model ( Odds Ratio ( OR ) of laboratory confirmation for suspected typhoid fever compared with fever was 3 . 34 , 95% CI: 1 . 11–10 . 1; p = 0 . 032 ) ( Table 1 ) . In the same analysis , study year ( or midpoint for multiannual studies ) was significantly associated with the odds of laboratory confirmation of typhoid . The OR was 0 . 95 ( 95% CI: 0 . 92–0 . 99 ) for each unit increment in the study year , although this decline is apparent in the forest plot only for studies from 1991 onwards ( Fig 2 ) . Typhoid was also more likely to be confirmed for studies that reported during an outbreak , although this was only of borderline significance in the univariate analysis ( OR 3 . 66 , 95% CI: 0 . 95–14 . 1; p = 0 . 060 ) . Other study characteristics , including location ( urban vs . rural , latitude ) and type of laboratory assay ( culture , serology or both ) were not significantly associated with the odds of confirmation of typhoid . In the multivariate meta-regression model including all covariates , the year of the study and whether it reported during an outbreak were significantly associated with the odds of laboratory confirmation of typhoid ( Table 2 ) . The duration of fever among patients eligible for testing and their age distribution were available in only 9 of the 37 including studies and therefore subgroup and meta-regression analysis based on these variables were not carried out . The estimated prevalence of laboratory confirmed paratyphoid across the hospital-based studies in the random effects model was 0 . 9% ( 95% CI: 0 . 5–1 . 7% ) ( Fig 3 ) . There was significant heterogeneity among studies ( LRT p<0 . 001 ) . Prevalence was not significantly different according to whether studies included patients with fever or suspected typhoid fever . In the univariate and multivariate meta-regression models only location ( urban vs . rural ) was significantly associated with the prevalence of paratyphoid , although this was driven by a single rural study with high prevalence [26] ( Table 2 ) . In the multivariate meta-regression model , reporting during a typhoid outbreak was associated with an increased odds of laboratory confirmed paratyphoid of borderline statistical significance ( OR 4 . 16 , 95% CI: 0 . 91–19 . 0; p = 0 . 067 ) . Funnel plots of typhoid and paratyphoid prevalence against study size were strongly suggestive of publication bias , such that studies with high prevalence were more likely to be published ( Figs A and B in S1 Appendix ) . The incidence of laboratory confirmed typhoid fever varied between the two locations where community cohort studies were carried out , with a more than four times higher incidence in Kalkaji ( Delhi ) of 976 per 100 , 000 person-years ( 95% CI: 736–1250 ) compared with Kolkata ( pooled estimate 235 , 95% CI: 203–271 ) ( Fig 4A ) . Although the former study reported for individuals aged 0–40 years and the latter for all ages ( under 2s were excluded in [27] ) , this does not explain this large difference in incidence , since individuals over 40 years old made up only 24% of the population in 2000 [28] . The pooled incidence across all studies was 377 ( 95% CI: 178–801 ) per 100 , 000 person-years although with significant heterogeneity among studies ( LRT p<0 . 001 ) . It was difficult to compare the age-distribution of typhoid incidence between studies because of differences in reporting of age-categories , although incidence was typically highest in the 2–4 year age-group ( Fig 4B ) . The incidence of paratyphoid was only reported for two studies in Kolkata that met our inclusion criteria , which gave a pooled estimate of incidence of 105 per 100 , 000 person years ( 95% CI: 74–148 ) for all ages ( although [27] only reported from 2 years of age ) . Note that the incidence of typhoid and paratyphoid in 2004 in Kolkata was estimated using number of individuals in the relevant study area and age-group at baseline because the number of person-years of observation was not reported [5 , 29] .
There have , surprisingly , been very few epidemiological investigations of the incidence of typhoid in India . The three community cohort studies that we identified in Kolkata and Delhi , the last of which reported data nearly a decade back , suggest a variable incidence of typhoid both over time as well as across regions . The variable burden of typhoid was also apparent in the meta-analysis of hospital-based studies , which showed significant heterogeneity in the reported prevalence of laboratory confirmed typhoid among patients with fever or suspected typhoid fever . In the meta-regression of hospital studies , testing of patients with suspected typhoid fever or during a typhoid outbreak was more likely to lead to laboratory confirmation of typhoid fever . The meta-regression also revealed a significant decline in laboratory confirmed typhoid among patients with fever or suspected typhoid fever over time , apparent since the early 1990s ( Fig 2 ) . The odds of detecting typhoid decreased by approximately 5% each year and this remained significant in the multivariate model accounting for differences in study location , laboratory assay , case definition and whether reporting was during an outbreak . Grouping the studies by decade shows that this significant decline is largely the result of the high prevalence of typhoid in hospital studies during 1980–2000 compared with more recent studies ( Table 1 ) . The cause of the high rate of typhoid isolation at this time is not clear . Moreover , inference of a trend from hospital-based studies in different locations and with variable health-seeking behaviours must be tentative . In particular , it is possible that increased use of effective antibiotics before blood collection could have contributed to this decline . The prevalence of typhoid fever was not significantly associated with any other covariates , including study location ( latitude , urban vs . rural ) , although the number of studies in rural areas was small , limiting the power of this analysis . The incidence of paratyphoid was reported in only two studies that met our inclusion criteria , both in Kolkata . The estimated incidence of paratyphoid in this setting was 105 per 100 , 000 person years , which compared with 235 per 100 , 000 person years for typhoid . In Kalkaji ( Delhi ) the incidence of paratyphoid was not originally reported , although in a companion publication [30] the number of paratyphoid cases recorded during a slightly longer follow-up ( 18 months ) compared with the original study ( 12 months ) [31] was 31 compared with 98 for typhoid over the same period . These estimates suggest an incidence rate for paratyphoid in these settings that is about 30–50% of the rate estimated for typhoid . The lower incidence of paratyphoid was confirmed in the meta-regression of hospital-based studies , which found a prevalence for paratyphoid that was approximately 10-fold lower than for typhoid ( estimated pooled prevalence of 0 . 9% vs . 10 . 7% ) . The significantly lower prevalence of paratyphoid among patients tested in hospital may also reflect the shorter duration of fever and more mild clinical characteristics of paratyphoid compared with typhoid [32] . Although significant heterogeneity in the prevalence of paratyphoid was identified among the hospital-based studies , this was not associated with case definition , laboratory assay , whether an outbreak was reported or any other study covariates , with the exception of study location . Prevalence was significantly higher in rural compared with urban locations , but this result was driven by a single study of paratyphoid in a rural area that had high prevalence . Consistent across the community cohort studies was the finding of a high incidence of typhoid in children under five years of age , suggestive of a substantial burden in a group that would benefit from infant rather than school-based immunization . This is consistent with recent recommendations from the IAP on the creation of an immunisation slot at 9–12 months of age for typhoid vaccination [21] . Risk-factors that would allow targeting of infant immunization to high-risk groups were not identified in this systematic review and meta-analysis . Significant heterogeneity was observed among studies , but this is likely in part to reflect differences in patient inclusion criteria , laboratory methods and changes in antimicrobial use and resistance patterns . Vaccine introduction is likely to be more sustainable , equitable and to provide indirect herd effects when it is done through the universal immunisation programme . The age-distribution of paratyphoid incidence was not reported , although the mean age in 2004 in Kolkata was reported as being significantly higher compared with typhoid ( 17 . 1 vs . 14 . 7 years ) [29] . Paratyphoid vaccines are not yet available and currently licensed typhoid vaccines offer limited or no protective immunity against paratyphoid A and B , the predominant serotypes [33] . However , vaccines are in the development pipeline , including bivalent conjugate vaccines that could offer protection against both typhoid and paratyphoid . There were limitations common to published studies that hampered our systematic review of the burden and risk factors for typhoid and paratyphoid fever in India . There have been only three community cohort studies of incidence , in just two locations and with highly variable findings . Although far more numerous , the hospital-based studies have a number of limitations . Firstly , hospital studies provide no information about incidence without a detailed understanding of local health-seeking behaviour—something missing from published studies . Secondly , they had varying inclusion criteria for patients , sample collection and laboratory methods , making interpretation of these data challenging . We focussed on laboratory confirmed typhoid or paratyphoid fever , mostly blood culture . However , blood culture has a poor sensitivity of about 50% and is strongly influenced by the quantity of blood , prior administration of antibiotics and culture techniques including quality of media [34] . We excluded cross-sectional community studies using serology , since these were likely to be highly non-specific for typhoid . Thirdly , detailed data on the inclusion criteria for patients including the duration of fever and age-distribution were usually missing , limiting the number of covariates we could include in the meta-analysis . In some studies , a failure to clearly define inclusion criteria along with ambiguous reporting forced us to exclude them because the denominator population was unclear . Fourthly , most studies did not report clinical outcomes and therefore we were unable to evaluate trends in severe disease or mortality . Finally , there was evidence from the funnel plots for publication bias , such that studies finding a high burden of typhoid or paratyphoid were more likely to be reported and published . Therefore , while there appears to be a declining trend in typhoid isolation in hospitals , drawing inference about the underlying burden of disease from hospital based data needs to be approached cautiously . The limitations of the community cohort and hospital-based studies make it difficult to estimate the total burden of typhoid and paratyphoid in India . Extrapolating the estimates of typhoid incidence from Kolkata and Delhi to the rest of India is clearly problematic . A naïve approach applying the pooled estimate of the incidence rate to the 2011 census population of 1 . 2 billion would give an estimated annual incidence of 4 . 6 million cases . This could be revised upwards by approximately twofold based on the poor sensitivity of culture-based confirmation of typhoid [2] . However , the community cohort studies were deliberately planned in densely populated urban areas with poor sanitation , likely to have a high incidence of typhoid . Correcting the national estimate for access to improved water following the approach used for regional estimates by Mogasale et al . 2014 [2] would give 2 . 1 million cases annually , or approximately 3 . 4 million correcting for imperfect culture sensitivity . Correction for other risk factors , such as population density , would likely reduce this estimate further . Strengthening surveillance across geographically representative sentinel sites is key to better disease burden estimates . Inclusion of other data sources such as large healthcare facilities , and the National Disease Surveillance Project is likely to further understanding of disease burden . Well defined surveillance criteria combined with standardized laboratory methods will greatly enhance comparability of estimates from diverse sites . Since blood cultures are highly dependent on volume of inoculum , prior antibiotic exposure and laboratory methods , a combination of conventional , molecular and serologic diagnostics modalities would probably be optimal . Information about time trends and antimicrobial resistance patterns that arise from such a systematic surveillance will enhance our understanding of typhoid and paratyphoid in India and strengthen public health decision making .
|
Typhoid fever is an important cause of avoidable mortality in regions without adequate access to safe water and sanitation . Highly immunogenic typhoid conjugate vaccines are now licensed and under consideration as a public health intervention in India . The decision about whether and how to introduce a public vaccination programme needs to be based on an understanding of disease burden , and the age-groups and geographic areas at risk . We performed a systematic review and meta-analysis of published studies reporting typhoid and paratyphoid incidence and prevalence in India between 1950 and 2015 . The estimated prevalence of laboratory-confirmed typhoid and paratyphoid among individuals with fever across all hospital studies was 9 . 7% ( 95% CI: 5 . 7–16 . 0% ) and 0 . 9% ( 0 . 5–1 . 7% ) respectively , with a significant decline in prevalence of the former over time . We found only three population-based studies that reported incidence . Pooled estimates were 377 ( 178–801 ) and 105 ( 74–148 ) per 100 , 000 person years for typhoid and paratyphoid respectively , with incidence being highest in in children between 2 and 4 years . Despite an apparent decline in prevalence , typhoid remains a significant burden in India , particularly among young children . Studies are required to evaluate the effectiveness of infant immunisation with conjugate typhoid vaccines .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"immunology",
"geographical",
"locations",
"india",
"vaccines",
"preventive",
"medicine",
"bacterial",
"diseases",
"signs",
"and",
"symptoms",
"vaccination",
"and",
"immunization",
"research",
"and",
"analysis",
"methods",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"serology",
"research",
"assessment",
"typhoid",
"people",
"and",
"places",
"diagnostic",
"medicine",
"asia",
"fevers",
"systematic",
"reviews",
"biology",
"and",
"life",
"sciences"
] |
2016
|
The Burden of Typhoid and Paratyphoid in India: Systematic Review and Meta-analysis
|
Environmental signals induce diverse cellular differentiation programs . In certain systems , cells defer differentiation for extended time periods after the signal appears , proliferating through multiple rounds of cell division before committing to a new fate . How can cells set a deferral time much longer than the cell cycle ? Here we study Bacillus subtilis cells that respond to sudden nutrient limitation with multiple rounds of growth and division before differentiating into spores . A well-characterized genetic circuit controls the concentration and phosphorylation of the master regulator Spo0A , which rises to a critical concentration to initiate sporulation . However , it remains unclear how this circuit enables cells to defer sporulation for multiple cell cycles . Using quantitative time-lapse fluorescence microscopy of Spo0A dynamics in individual cells , we observed pulses of Spo0A phosphorylation at a characteristic cell cycle phase . Pulse amplitudes grew systematically and cell-autonomously over multiple cell cycles leading up to sporulation . This pulse growth required a key positive feedback loop involving the sporulation kinases , without which the deferral of sporulation became ultrasensitive to kinase expression . Thus , deferral is controlled by a pulsed positive feedback loop in which kinase expression is activated by pulses of Spo0A phosphorylation . This pulsed positive feedback architecture provides a more robust mechanism for setting deferral times than constitutive kinase expression . Finally , using mathematical modeling , we show how pulsing and time delays together enable “polyphasic” positive feedback , in which different parts of a feedback loop are active at different times . Polyphasic feedback can enable more accurate tuning of long deferral times . Together , these results suggest that Bacillus subtilis uses a pulsed positive feedback loop to implement a “timer” that operates over timescales much longer than a cell cycle .
Cells are capable of responding to stimuli extremely rapidly , on timescales of seconds or less [1] . In some situations , however , cells respond to stimuli only after extended delays of multiple cell cycles . A classic example occurs in the developing mammalian nervous system , where , in the presence of appropriate signaling molecules , precursor cells will proliferate for up to eight cell generations before differentiating into oligodendrocytes [2] . Although many aspects of the system remain unclear , oligodendrocyte differentiation is similarly delayed in vivo and in cell culture , suggesting a cell-autonomous “timer” mechanism . Another example is the mid-blastula transition in developing Xenopus embryos , which occurs after 12 cell cycles of proliferation [3] , [4] . In both cases , the deferral of differentiation enables a period of proliferation preceding commitment to new fates . In bacteria , non-cell-autonomous strategies for deferring responses are well known . For example , in the marine bioluminescent bacterium Vibrio fischeri , cells use quorum sensing mechanisms to defer light production until the population reaches a critical density [5] . Similarly , Bacillus subtilis can defer sporulation through cannibalism [6] , [7] , a response triggered by cell-cell signaling at high cell density , in which one subpopulation of cells lyses another , releasing nutrients that sustain growth . Although there has been much work on circuit architectures that speed response times [8] , fewer studies have addressed cell-autonomous deferral mechanisms . Cell autonomous deferral requires the cell to keep track of the total time or number of division events since the appearance of the stimulus . It has remained unclear whether and how individual bacterial cells can achieve this functionality using genetic circuit components . The key problem is that as the cell grows and divides , its components dilute out . This dilution process sets an effective upper limit to the typical timescale over which the concentration of a protein responds to a step change in its production rate [8] . For example , a step change in the rate of production of a stable protein causes the concentration of that protein to exponentially approach its new steady-state value with a timescale of one cell cycle [9] . Thus , most gene circuits tend to relax to new steady states over timescales close to , or faster than , that of the cell cycle . Alternatively , genetic circuits can give rise to long deferral times in some cells through occasional stochastic switching between metastable states . While such systems can be tuned to generate long mean intervals between switching events , without cascades of multiple states , these mechanisms cannot generate well-defined , unimodal distributions of deferral times across a population [10] , [11] . B . subtilis sporulation provides an ideal model system to address this problem . Sporulation is a canonical microbial stress response behavior , during which cells respond to stress by differentiating into an environmentally resistant spore . Sporulation is a terminal differentiation decision , and its initiation is regulated by a well-characterized genetic circuit whose dynamics can be analyzed in individual cell lineages [12] . This circuit , in response to diverse environmental and metabolic signals [13] , controls the activation of the master regulator Spo0A through transcriptional regulation and phosphorylation [14] . High levels of phosphorylated Spo0A ( Spo0AP ) are sufficient to induce sporulation [15] . However , under some conditions , Spo0AP levels increase gradually over multiple cell cycles , allowing cells to proliferate prior to differentiation . The ability to defer sporulation while proliferating could provide a fitness advantage to cells by increasing their numbers relative to immediate sporulators ( Figure 1A ) , although it could also impose a cost to cells that do not sporulate in time to survive extreme conditions . During the deferral period , cells may also explore other fates , such as biofilm formation , which are known to occur at intermediate levels of Spo0AP [16] . The genetic circuitry controlling Spo0A activation includes multiple types of interactions ( Figure 1C ) . Histidine kinases such as KinA , KinB , and others autophosphorylate and transfer phosphates through a phosphorelay consisting of Spo0F and Spo0B to Spo0A [17] . Phosphatases reduce the total level of Spo0AP . For example , Spo0E directly dephosphorylates Spo0AP [18] , while rap phosphatases drain phosphates from the phosphorelay through Spo0F [19] . The system also includes extensive transcriptional regulation . Spo0AP regulates its own transcription as well as that of spo0F . It also regulates many other genes , including global regulators such as AbrB [20] . Finally , Spo0AP also indirectly regulates its own activity by activating kinase expression [21] . These transcriptional interactions typically occur at much longer timescales than the fast phosphotransfer reactions of the phosphorelay . Nevertheless , it remains unclear whether and how this circuit facilitates deferred differentiation . Here , using time-lapse fluorescence microscopy of individual cells , we show that under some conditions B . subtilis cells defer sporulation for multiple cell cycles through a predominantly cell-autonomous mechanism . We observed a progressively increasing series of pulses of Spo0A phosphorylation during deferral . Manipulation of circuit interactions revealed that pulse growth and regulated deferral both required positive feedback on kinase expression . These results suggest that B . subtilis uses a pulsed positive feedback loop to gradually “ratchet up” Spo0AP activity pulses over multiple cell cycles in order to defer sporulation . Finally , mathematical modeling of this mechanism further suggests that pulsing could enable a “polyphasic” feedback mechanism , in which different parts of the overall positive feedback loop are active at different times , facilitating regulation of deferral . This may be a general strategy that cells can use to enable regulation of timescales much longer than the cell cycle .
In order to understand how deferral is achieved , we set out to observe phosphorelay circuit dynamics during the deferral period . To read out Spo0AP activity we chromosomally integrated a Pspo0F-yfp reporter construct . The Pspo0F promoter exhibits a high affinity for Spo0AP and is therefore classified as a low-threshold activated gene [21] . To quantify Pspo0F activity over time , we computed its YFP production rate ( promoter activity ) in single cells . Promoter activity takes into account measurements of the change in total cellular fluorescence between time-points , the instantaneous cellular growth rate ( which varies considerably , even within a single cell lineage , Figure 1E , bottom panel ) , and other cellular parameters ( Materials and Methods ) . Compared to mean cellular fluorescence , whose interpretation is complicated by the stability of fluorescent proteins , promoter activity better reflects production from Pspo0F and thus Spo0AP dynamics . We also inserted a constitutively expressed red fluorescent expression construct , PtrpE-mCherry , which we used to aid in the automatic segmentation of cells in images . Pspo0F promoter activity could be observed in discrete pulses in individual cells , similar to those reported previously ( Figure 1D , E ) [26] . These pulses began after transfer to nutrient-limited conditions and continued until sporulation . In contrast , cells in rich media exhibited no measurable production from the Pspo0F promoter , or sporulation associated genes generally . Pulses were not specific to the Pspo0F reporter , but were observed across a range of Spo0A target genes ( Figure S1 ) , affecting many processes in the cell , including the expression of the global regulator abrB [27] , [28] and sdp , a component of the “cannibalism” pathway [6] , [21] . However , the phasing of pulses relative to the cell cycle differed between promoters , reflecting their different regulation modes ( Figure S1 ) . For example , Spo0AP-activated and Spo0AP-repressed promoters showed opposite phasing with respect to the cell cycle ( Figure S1B ) . Each cell cycle typically contained one pulse ( Figure S1C ) . Promoters not regulated by Spo0A , such as the σA-dependent PtrpE promoter , sometimes fluctuated in expression but exhibited a much smaller dynamic range , and no characteristic cell cycle phasing , and were thus qualitatively different from Spo0A-dependent pulsing ( Figure S3 ) . In principle , pulses could be caused by a change in either the abundance or the phosphorylation state of Spo0A . To eliminate both transcriptional and phosphorylation control of Spo0A activity , we replaced spo0A with the well-characterized , constitutively active variant spo0Asad67 [29] , under the control of the IPTG-regulated hyperspank ( Phyp ) promoter . Although Pspo0F was activated in response to spo0Asad67 induction , this strain showed no pulsing ( Figure S4 ) , consistent with the fact that spo0Asad67 does not require phosphorylation to be active . In addition , very few cells formed phase bright spores . The potential for pleiotropic effects of spo0Asad67 expression prevents us from concluding that successful sporulation requires pulsing . On the other hand , a strain in which spo0A was under the control of Phyp retained similar pulse dynamics as wild-type ( Figure S5 ) and consistently formed phase bright spores . These results strongly indicate that phosphorylation of wild type Spo0A is required for pulsation . What molecular mechanism could be responsible for pulse generation ? The sporulation kinase inhibitor Sda is regulated in a cell-cycle-dependent fashion , suggesting that it might be involved in pulse generation [30] . A null sda mutant exhibited increased mean Spo0A activity , and therefore strongly reduced the dynamic promoter activity of the sensitive Spo0A-repressed PabrB promoter [21] , as observed previously ( Figure S6 ) [26] . However , in the Δsda mutant , Pspo0F continued to pulse similarly to wild-type , showing that while Sda modulates the dynamic range of Spo0A activity , it is not required for pulsing . Intrinsic network dynamics involving negative feedback loops provide another possible pulse generation mechanism [31]–[33] . Together , spo0A , abrB , and spo0E form such a feedback loop ( Figure 1B ) . However , deletion of spo0E did not eliminate pulsing ( Figure S6 ) . Other potential negative feedback loops involve Spo0A-dependent up-regulation of rap phosphatase expression . But deleting rapA and rapB individually and in combination similarly failed to abolish pulsing ( Figure S6 ) . Finally , we asked whether pulsing might be driven specifically by one of the phosphorelay kinases . In nutrient-limited conditions , KinA and KinB are the dominant phosphodonors [34] , [35] . Strains lacking either kinA or kinB exhibited pulsed dynamics ( Figure S7 ) , suggesting that pulsation does not specifically require kinA or kinB individually . Together , these results show that pulsing is robust to deletion of a variety of different circuit components . While further elucidation of the mechanism of pulse generation will be important , we focus below on the consequences of pulsing for the deferral of sporulation . In principle , the extended multi-cell-cycle timescale for activation of Spo0A could be achieved in three different ways ( Figure 2A ) : Internal genetic circuitry could generate a slow rise in a critical regulator ( CIRCUIT cartoon ) . Alternatively , an inhibitor of sporulation could gradually dilute out over multiple cell cycles during the proliferation phase ( DILUTION cartoon ) . Either of these two mechanisms would function cell-autonomously . Finally , cells could defer sporulation through a non-cell-autonomous mechanism involving the build-up of extracellular signaling molecules that modulate the phosphorelay ( quorum sensing ) [6] , [36]–[39] or through degradation of the local micro-environment ( QS/ENV cartoon ) . We performed two experiments that distinguish between these possibilities and , together , support a cell-autonomous mechanism that does not involve dilution for gradual build-up of Spo0A activity . In the first experiment , we sought to distinguish between cell-autonomous and non-cell-autonomous deferral mechanisms . Low initial cell densities on resuspension media pads did not permit the growth and sporulation of cells , suggesting that at least some cell-generated factors were required for proliferation and possibly sporulation in these conditions . However , it was not clear whether these signals were responsible for deferring sporulation . To address this question , we developed a co-culture assay where unlabeled cells were mixed with red mCherry-labeled cells on the same pad ( Figure 2B–D ) . The unlabeled wild-type cells were introduced ∼10 h before the labeled cells , allowing them to condition the pad as they proliferated and eventually sporulated ( Figure 2B–D ) . If deferral were controlled by cell-extrinsic factors , then the red cells should sporulate earlier with the unlabeled cells than without them ( Figure 2B lower cartoon ) . On the other hand , if the deferral of sporulation were cell-autonomous , then the red cells would proliferate for an equal amount of time in the presence or absence of the unlabeled cells ( Figure 2B , upper cartoon ) . In order to quantify this effect , we counted the number of cell cycles required for 50% of cells in a microcolony , starting from a single labeled cell , to initiate sporulation , as measured by the formation of a phase-bright forespore . Because the actual distribution of deferral times has a tail ( Figure 1B ) , this measure , denoted as T50 , approximates but slightly underestimates the actual mean deferral time as measured in individual cell lineages ( Materials and Methods ) . We found that sporulation of labeled cells was only modestly accelerated by unlabeled sporulating microcolonies ( Figure 2C and Figure 2D ) , reducing T50 by 25% , from ∼4 to ∼3 cell cycles . This measurement provides an upper limit to cell-extrinsic effects in our conditions . Although cell-extrinsic factors do play some role , deferral appears to be controlled in a predominantly cell-autonomous fashion . We next sought to determine whether cell autonomous deferral in our conditions was caused by slow depletion of internal factors following the switch to resuspension media ( Figure 2A , middle panel ) . One specific molecular candidate for the dilution mechanism is a slow depletion of intracellular GTP levels , which control repression of stationary phase genes by CodY through the alarmone ( p ) ppGpp [40] . However , in our experimental conditions , a ΔcodY strain showed similar deferral behavior as wild-type cells ( Figure S8 ) , demonstrating that codY is not necessary for deferral . Because the dilution mechanism need not work through codY , we designed an experiment to rule out the dilution model more generally ( Figure 2E–G ) . In this experiment , a strain with IPTG-inducible spo0A ( Δspo0A Phyp-spo0A ) is first grown on one nutrient-limited pad lacking IPTG and then transferred to a second , similar , nutrient-limited pad containing 100 µM IPTG ( Figure 2E ) . The first pad allows cells to grow for multiple cell cycles without inducing sporulation ( Figure S2A ) . This growth dilutes out putative internal factors not produced in nutrient-limited conditions . On the second pad , IPTG is present , enabling immediate constitutive transcription of Spo0A . After one cell cycle , Spo0A concentration reaches a steady state expression level at or exceeding that in sporulating wild-type cells ( Figure S2B ) . In the dilution model , dilution of sporulation inhibitors during growth on the first pad would cause cells to sporulate immediately on the second pad . On the other hand , if deferral were due to cell-autonomous Spo0A circuit dynamics , growth on the first pad would have no effect on deferral on the second pad . In fact , the T50 distribution on the second pad was not substantially affected by 3–4 cell cycles of growth on Pad 1 , with T50 = 4 . 1 ± 0 . 2 versus 4 . 4 ± 0 . 1 ( mean ± SD ) with and without Pad 1 , respectively ( Figure 2F , G ) . These results rule out the dilution-driven deferral model . Together , these results strongly suggest that multi-cell-cycle deferral is controlled by an extended cell-autonomous accumulation of Spo0AP . To better understand how the sporulation initiation circuit controls deferral time , we consider two classes of genes . The first class consists of the phosphorelay genes Spo0A , Spo0F , and Spo0B and the sporulation kinases KinA–KinE , whose products directly contribute to the phosphorylation of Spo0A . Limited expression of these genes could potentially defer sporulation by slowing the phosphorylation of Spo0A . The second class consists of phosphorelay phosphatases , whose expression could potentially defer sporulation by draining phosphates from the relay , slowing the accumulation of phosphorylated Spo0A . We investigated the impact of these phosphorelay components on multi-cell-cycle deferral , distinguishing between two qualitatively different regimes , similar to an approach used previously ( Figure 3A ) [41]: In a relay-limited regime , phosphorelay protein concentrations ( e . g . Spo0F and/or Spo0B and/or Spo0A itself ) limit the rate of phosphotransfer and thus the level of Spo0AP . In contrast , in a flux-limited regime , the level of Spo0AP is principally controlled by the rate at which phosphates are injected into the circuit by kinases and/or removed by phosphatases . To experimentally distinguish between the two regimes , we analyzed the behavior of unlabeled wild type cells alongside ( cocultured with ) mCherry-labeled cells engineered to overexpress different phosphorelay components . Overexpression of limiting components , but not non-limiting components , will accelerate sporulation relative to wild type as shown schematically in Figure 3B . Thus , overexpression of spo0A or an operon of phosphorelay components ( spo0A , spo0B , and spo0F ) should accelerate both Spo0AP buildup and sporulation in the relay-limited regime , while having little to no effect in the flux-limited regime . Conversely , in the kinase-limited regime , kinase overexpression should accelerate both Spo0AP buildup and sporulation in the flux-limited regime but have little or no effect in the relay-limited regime . We note that previous related work by Fujita and Losick has established the strong effects of kinA overexpression in a different context , showing that it is sufficient to induce immediate sporulation in rich media conditions , which strongly suppress sporulation altogether [15] . We observed little to no acceleration in the onset of sporulation when we expressed spo0A or the spo0A-spo0B-spo0F operon in the labeled cells , despite the ability of these constructs to complement corresponding mutants ( Figure 3C ) . These cells sporulated with a T50 = 3 . 7 ± 0 . 2 ( mean ± SD ) , similar to 4 . 0 ± 0 . 2 in wild-type cells . On the other hand , induced expression of kinA strongly accelerated both the activation of Spo0A , as measured by Pspo0F expression , and the onset of sporulation ( Figure 3D ) , resulting in T50 = 0 . 2 ± 0 . 1 . These results suggest that the deferral of sporulation is flux-limited , but not relay-limited . To further test whether kinases or phosphatases were responsible for flux limitation , we constructed strains lacking phosphorelay phosphatases individually and in combinations . Simultaneous deletion of spo0E , rapA , and rapB reduced deferral by about one cell cycle , but did not abolish the multi-cell-cycle deferral . Deletion of other phosphatase genes , including the Spo0A phosphatases yisI and ynzD , and the Spo0F phosphatases rapE , rapH , and rapJ , had no discernible effect ( Figure S8 ) . Evidently , phosphatases alone cannot explain the flux limitation underlying multi-cell-cycle deferral , whereas kinase over-expression is sufficient to abolish multi-cell-cycle deferral . Together , these results implicate the slow buildup of kinase as the predominant deferral mechanism . This hypothesis is supported by analysis of a PkinA-yfp reporter , which confirmed that KinA concentration indeed builds up gradually in the cell cycles preceding sporulation , and does so to an extent that cannot be explained by the less than 2-fold slowing of growth rate during the experiment ( Figure S9 ) . Similarly , while cells on Pad 2 in the dilution experiment ( Figure 2F ) exhibited systematically slower growth rates than wild type cells ( Figure S11 ) , they still sporulated with a similar deferral period . Evidently , regulation of kinA expression leads to a progressive increase over multiple cell cycles . One of the most prominent activators of the principle sporulation kinases kinA and kinB is Spo0A itself . Spo0AP inhibition of the transcriptional repressor AbrB leads to up-regulation of kinA through σH [42] and de-repression of kinB [43] . Thus , increased kinase activity could be driven by the engagement of a positive feedback loop , in which Spo0A activity pulses activate kinase transcription , increasing the amplitude of subsequent Spo0A pulses , and thus ratcheting up kinase levels once per cell cycle . A comparison of Spo0AP levels ( inferred from Pspo0F promoter activity ) with KinA levels ( measured with PkinA-yfp fluorescence ) demonstrated that kinA expression correlates with Spo0AP pulse amplitudes ( Figure S9C ) . Imaging of a kinA-gfp protein fusion confirmed that KinA protein levels increase during the deferral period ( Figure S10A ) [44] . Furthermore , ectopic expression of a constitutively active spo0A mutant , spo0Asad67 , in a Δspo0A background , led to full up-regulation of a PkinA-yfp reporter ( Figure S10B ) , and no reporter expression was observed in this strain without induction of spo0Asad67 . Together , these results indicate that active spo0A is necessary and sufficient for full kinA expression . To investigate the potential role of this positive feedback loop , we developed a method to quantify pulse growth in individual cells . First , we characterized each Pspo0F promoter activity pulse by its peak amplitude ( Figure 4A ) . This allows promoter activity time traces to be represented by a discrete sequence of pulse amplitudes , one per cell cycle . We then plotted these pulse sequences on a “return map , ” where the amplitude of each pulse ( labeled pN+1 ) is plotted against the amplitude of the pulse immediately preceding it ( labeled pN ) ( Figure 4B ) . Pulse growth causes points on the return map to lie above the diagonal line pN = pN+1 . In wild type cells , pulse amplitudes , though variable , tended to grow with successive cell cycles . Thus , on the return map , over two-thirds of data points lie above the diagonal , with the strongest growth at low and intermediate pulse amplitudes ( Figure 4C ) . At high pulse amplitudes the trend saturates , so that the amplitude of a pulse eventually becomes independent of its predecessor . These results are consistent with the existence of a saturating positive feedback on kinase expression . By contrast , if kinase expression were constitutive ( Figure 4D ) , then induced kinase expression , and thus Spo0AP pulse amplitude , would relax to a steady state with a timescale of about one cell cycle ( similar to Figure S2B ) , eliminating systematic pulse growth ( Figure 4E ) . To test this prediction experimentally , we constructed a “feedback bypass” strain , combining a ΔkinA ΔkinB ΔkinC triple deletion with IPTG-inducible kinA expression . In this strain , modest levels of IPTG ( 2 µM ) allowed cells to grow and divide multiple times while activating Spo0A . Like wild-type cells , these cells exhibited variable amplitude Spo0AP pulses correlated with the cell cycle ( Figure S7 ) . However , lacking transcriptional feedback on kinase expression , the pulse amplitudes showed no systematic growth over successive cell cycles ( Figure 4F ) . These results suggest that feedback on kinase transcription is required for pulse growth . Furthermore , the lack of pulse growth in the feedback bypass strain predicts an extremely sensitive dependence of sporulation timing on kinase expression levels . In this strain , IPTG concentration controls the steady-state kinase concentration , but not the timescale to reach it , which is set by the cell division time ( Figure 5A ) . At low IPTG levels , Spo0AP can never grow high enough to induce sporulation . Conversely , at high IPTG levels , sporulation would be induced almost immediately . Between these two extremes , sporulation would be deferred for multiple cell cycles only in a narrow window of kinase expression levels . To test this prediction , we compared sporulation timing in our feedback bypass cells , labeled with mCherry , to that of wild-type cells co-cultured on the same agarose pad ( Figure 5B ) . At low IPTG induction , these cells largely failed to sporulate , while at high induction levels , cells sporulated within one or two cell cycles ( Figure 5C ) . The fraction of sporulated cells at 30 h showed a sharp dependence on kinA induction level , equivalent to a Hill coefficient of 4 . 1 ± 1 . 8 ( 95% confidence interval ) ( Figure 5D ) . Together , these results show that positive feedback on kinase expression is necessary for regulated deferral . How does positive feedback enable cells to set long deferral times , and what role can the pulsatile activation of Spo0A play ? To explore these questions we constructed a mathematical model of the sporulation initiation circuit . We used a simplified model ( Text S1 ) in order to gain insights into qualitative differences between different circuit architectures , but not to reproduce all known molecular interactions in the circuit . We modeled pulsatile Spo0A phosphorylation by activating kinase autophosphorylation for a fixed fraction of each cell cycle . We also simplified the phosphorelay into a two-component phosphotransfer from kinase to Spo0A . Although they are likely to be important for some aspects of the natural system , inclusion of Spo0F and Spo0B does not qualitatively affect the conclusions below . Finally , based on the insensitivity of deferral time to phosphatase deletions , we modeled phosphatase activity with a constant level of Spo0E . Sporulation initiation is believed to require a threshold level of phosphorylated Spo0A . Indeed , we found that maximal Pspo0F-yfp promoter activity in sporulating cells was systematically higher than in vegetative cells ( Figure S12 ) . Recently published experiments in bulk cultures have also demonstrated that cells sporulate at a threshold level of KinA [45] . Therefore , to analyze deferral , we quantified the number of cell cycles required for phosphorylated Spo0A to grow from a low initial level to a high threshold level . We performed this analysis for three distinct circuit architectures , which differ in how kinase expression is controlled ( Figure 6 ) : In the first circuit , kinase is produced constitutively ( open loop , Figure 6A ) . In the second , kinase production is instantaneously activated by Spo0AP ( pulsed instantaneous positive feedback , Figure 6B ) . In the third circuit , kinase is indirectly activated by Spo0AP , leading to an effective time delay ( τ ) between Spo0A phosphorylation and consequent up-regulation of kinase expression . If the Spo0AP pulse terminates before kinase expression initiates , the pulsing and time delay together effectively divide the deferral period into distinct phases where either Spo0AP pulsing or kinase transcription ( or neither ) is active , but never both; we call this type of feedback “polyphasic feedback” ( Figure 6C ) . In the polyphasic mechanism , τ could represent a number of possible intermediate processes including indirect regulation as well as transcriptional and translational time delays . Our current methods cannot firmly establish nor rule out such time delays , due to the relatively low time resolution inherent in promoter activity measurements made with fluorescent protein reporters . For example , we could not detect a consistent time difference between pulses of PkinA-yfp and Pspo0F-cfp promoter activity ( Figure S9B ) . Higher time resolution and methods to track protein phosphorylation in individual cells could help to constrain the exact magnitude of such effects . For each circuit , we systematically modulated kinase or phosphatase production rates , both of which directly control phosphate flux to Spo0A . For each production rate , we monitored the time required for Spo0AP to exceed a fixed threshold ( deferral time , Figure 6D ) , and computed the sensitivity of this time to parameters such as the kinase production rate , ( Figure 6E ) . The open loop circuit showed an extremely sensitive dependence of deferral time on phosphorylation parameters ( Figure 6D&E , blue ) , consistent with the sensitive dependence on kinase production observed experimentally ( Figure 5D ) . The positive feedback loop reduced this sensitivity ( Figure 6D&E , green ) , and the polyphasic feedback loop reduced it still further ( Figure 6D&E , red ) . Models were tuned so the steady-state Spo0AP levels in the polyphasic circuit exceeded those in the positive feedback circuit , ensuring that longer deferral times were not caused by lower steady states . Positive feedback , especially in the polyphasic regime , evidently could make it easier for cells to regulate multi-cell-cycle deferral times by reducing the sensitivity of deferral time to key control parameters [46] . How can we explain the relative sensitivities of the three circuits ? In the open loop circuit , protein dilution due to cell growth determines the kinase concentration kinetics . The dilution rate is determined by the cell cycle time , making it difficult to achieve deferrals longer than a single cell cycle . In the positive feedback circuit , protein production and dilution compete with each other [47] to set the timescale of kinase accumulation . Parameters that affect net feedback strength ( e . g . , kinase or phosphatase promoter strengths ) directly tune this timescale , and thereby modulate deferral time . Finally , the polyphasic positive feedback circuit includes the benefits of positive feedback . In addition , however , the combination of Spo0AP pulsing and a time delay in its feedback onto kinase production together cause most of the new kinase produced by a pulse to appear only after the pulse terminates . Consequently , kinase cannot instantaneously feed back to amplify the pulse that produced it . Since feedback occurs from pulse to pulse , rather than compounding continuously as in standard positive feedback , pulse growth is much less sensitive to changes in feedback strength . Qualitative insights into the three circuit architectures can be obtained by analytically solving a set of corresponding simplified one-dimensional models ( Text S1 and Figure S13 ) . In these simplified models , although protein concentration grows exponentially in both positive feedback and polyphasic circuits , the time constant of the polyphasic circuit is exponentially less sensitive to feedback strength .
To respond properly to the challenges posed by environmental and developmental constraints , cells respond to stimuli across widely varying timescales . In some systems , the challenge is to achieve extremely rapid responses [1] . In other cases , however , cells may face the opposite challenge of deferring a response for relatively long times . Sporulation initiation represents an ideal example , where a sudden change in environment leads to a particular response—sporulation—only after many cell cycles . Although sporulation is deferred , it is clear that cells respond to the change in conditions throughout the deferral period , for example through continual increases in Spo0A activity . In principle , several different mechanisms can produce a deferred response . Quorum sensing mechanisms can defer activation of a response until a critical cell density is reached , as occurs in the V . harveyii light production circuit [5] . Our data do not rule out a role for quorum sensing , but show that it cannot explain most of the multi-cell-cycle delay observed here ( Figure 2B–D ) . Furthermore , since quorum sensing is a response to absolute cell density , rather than to rounds of cell division , it may be better suited to measuring population size as opposed to time intervals . A second potential mechanism for deferral is dilution of an internal molecule that represses sporulation . Dilution failed to explain sporulation deferral in our experiments ( Figure 2E–G ) . A dilution mechanism requires the regulator to be produced continually before nutrient limitation at a level tuned to provide the appropriate deferral time during nutrient limitation . Thus , this strategy might be better adapted to a more deterministic environment , such as multicellular development [2] , rather than the more unpredictable environments that microbes experience [48]–[50] . In contrast , the cell-autonomous feedback-dependent mechanism analyzed here allows deferral time to be quickly tuned from immediate to multiple cell cycles under different conditions . Indeed , previous studies of sporulation have used conditions optimized to induce immediate , rather than deferred , sporulation with the same circuit [24] , [41] . The relative advantages of each type of mechanism may become clearer as additional examples of deferred differentiation are identified and elucidated . Feedback loops are known to affect the response times of gene circuits . Negative feedback has been previously shown to accelerate responses [8] . Here we demonstrate the complementary role of positive feedback in extending timescales . This latter function is particularly important when proliferating cells need to postpone responses beyond a single cell cycle , the longest fundamental timescale of protein turnover in growing cells . Positive feedback extends timescales by competing with protein dilution to set the net relaxation time for protein concentrations . In B . subtilis sporulation , our results reduce the overall circuit to a core two-element positive feedback loop involving the master regulator Spo0AP and the sporulation histidine kinase KinA . This feedback loop progressively ratchets up Spo0AP levels , approaching the threshold level required for sporulation only after multiple cell cycles , and thereby enabling multi-cell-cycle deferral . A striking aspect of the system analyzed here is pulsatile phosphorylation of Spo0A . Pulsing imposes additional temporal structure on circuit dynamics that can lead to novel regulatory capabilities . For example , the yeast transcription factor Crz1 undergoes discrete pulses of nuclear localization at a frequency set by extracellular calcium concentration [51] . The relative fraction of time Crz1 spends in and out of the nucleus is determined by the pulse frequency . This “FM” regulation sets the fraction of time that all of Crz1's targets are activated , leading to proportionally coordinated expression of the entire regulon . In addition to quantizing responses , pulsing can also dictate the relative timings of different interactions in a circuit . Here , the pulsed buildup of Spo0AP defers sporulation for multiple cell cycles through a Spo0AP-KinA positive feedback loop . When time delays are present in this feedback loop , increased KinA production occurs after the Spo0AP pulse ends . As a result , Spo0AP production and KinA production are temporally separated . In this “polyphasic” regime , pulsing and time delay work together to prevent instantaneous feedback , making the buildup rate significantly less sensitive to parameter values than it would be in a conventional positive feedback loop . It will be interesting to develop techniques that can access these dynamics with higher time resolution , and to see if this polyphasic strategy provides a general design principle for regulation of multi-cell-cycle deferral times in other systems . Finally , one can ask whether deferring sporulation might have other benefits in addition to enabling proliferation . B . subtilis cells could explore alternative cell fates such as competence , biofilm formation , or cannibalism , during the deferral period . In this way , deferred progression to sporulation , implemented by a simple cell-autonomous pulsed positive feedback circuit , provides a critical foundation upon which multifaceted developmental programs can unfold .
Custom MATLAB software , similar to that described in Rosenfeld et al . [52] , was used to extract time lapse fluorescence values for individual cells and lineages in a microcolony . All subsequent data analysis was also done in MATLAB using customized software .
|
How long should a cell wait to respond to an environmental change ? While many pathways such as those affecting chemotaxis respond to environmental signals quickly , in other contexts a cell may want to defer its response until long after the signal's onset—sometimes waiting multiple cell cycles . How can cells create “timers” to regulate these long deferrals ? We study this question in the bacterium Bacillus subtilis , which responds to stress by transforming into a dormant spore . We show that B . subtilis can defer sporulation for extended time periods by first undergoing multiple rounds of growth and proliferation , and only then sporulating . The timer for this deferral is a pulsed positive feedback loop , which ratchets up the concentration of the sporulation master-regulator Spo0A to a critical level over multiple cell cycles . Finally , using mathematical modeling , we illustrate how a novel dynamic feedback mechanism , “polyphasic positive feedback , ” lets cells defer sporulation more robustly than with other circuit strategies . Developing techniques that can access pulsing and time-delay dynamics with higher time resolution will enable us to determine if this polyphasic strategy provides a general design principle for the regulation of multi-cell-cycle deferral times seen in other systems .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"systems",
"biology",
"developmental",
"biology",
"prokaryotic",
"models",
"model",
"organisms",
"biology",
"computational",
"biology",
"microbiology",
"genetics",
"and",
"genomics"
] |
2012
|
Pulsed Feedback Defers Cellular Differentiation
|
Since the 1990s , Singapore has experienced periodic dengue epidemics of increasing frequency and magnitude . In the aftermath of the 2004–2005 dengue epidemic , hospitals refined their admission criteria for dengue cases to right-site dengue case management and reduce the burden of healthcare utilization and negative outcomes . In this study , we describe the national trends of hospital admissions for dengue and disease severity in terms of length of stay ( LOS ) , admission to the intensive care unit ( ICU ) and death in hospital , and case fatality rate ( CFR ) in Singapore . We conducted a retrospective study of notified cases and laboratory confirmed dengue patients admitted to all public and private hospitals between 2003 and 2017 . Case notifications for dengue and hospitalization records were extracted from national databases . The proportion of dengue cases hospitalized was lower in recent years; 28 . 9% in the 2013–2014 epidemic , compared to 93 . 2% in the 2004–2005 epidemic , and 58 . 1% in the 2007 epidemic . Median LOS remained stable over the years; overall LOS was 3 to 4 days and ICU stay was 2 to 3 days . Less than 2% of hospitalized patients were admitted to the ICU . Overall CFR was low and remained below 0 . 5% . The proportions of dengue cases hospitalized and patients admitted to the ICU were highest in the elderly aged 65 years and older . While the proportion of dengue cases hospitalized saw a drastic decline due to more selective admission criteria , there was no concomitant increase in adverse outcomes , suggesting that admission criteria were appropriate to focus on severe dengue cases . Further studies are needed to optimize dengue management in older adults who are more likely to be hospitalized with greater disease severity , given the higher proportions of hospitalizations and severe disease among older adults .
Dengue was ranked by the World Health Organisation ( WHO ) as the “most important mosquito-borne viral disease in the world” in 2012 , in view of its increasing spread into previously unaffected areas and its high disease burden [1] . The global incidence of dengue has increased 30-fold over the past 50 years , and an estimated 2 . 5 billion people are at risk of infection , with dengue virus ( DENV ) endemic in over 100 countries spanning the Americas , Caribbean Basin , Asia and Africa [2] . The WHO Southeast Asia Region and Western Pacific Region are the most seriously affected , and together they contribute about three-quarters of the global dengue disease burden [3] . Singapore , a globally-connected island city-state , is one of several countries with high disease burden of dengue [4] . It is widely recognized that dengue epidemics impose a substantial burden on public health and health services , and incur considerable economic , societal and personal costs . In a local study , the annual average disease burden of dengue in Singapore was estimated to be 9–14 disability-adjusted life-years per 100 , 000 population during the ten-year period from 2000 to 2009 , while the average economic impact of dengue illness ranged from $0 . 85 billion to $1 . 15 billion in 2010 US dollars [5] . To tackle the public health impact of dengue in Singapore , a comprehensive nationwide Aedes prevention and control programme incorporating environmental management and source reduction , health education and law enforcement was launched in 1969 and successfully implemented since 1973 , as evidenced by a sharp reduction in the Aedes house index ( percentage of residential premises found to be breeding Aedes mosquitoes ) and low disease incidence [6 , 7] . Despite these efforts , dengue epidemics of increasing magnitude and an elevated level of endemicity have occurred since the late 1980s in a five- to six-year cycle [8–10] . The WHO , in laying out the global strategy for dengue prevention and control , has cited the particular challenges arising from unexpected surges in dengue cases , as well as the strain on health services arising from over-admission because of the limitations of triage in reliably predicting which severe cases will require hospital care [1] . The hospitalization rates of persons suspected of dengue viral infection are high , as doctors tend to err on the side of caution and admit them for monitoring . This leads to higher bed occupancy rates , although majority of dengue cases are unlikely to require or benefit from medical care in hospital for their mild form of illness . In the aftermath of the 2004–2005 dengue epidemic in Singapore , hospitals reviewed and refined their admission criteria for dengue cases [11–13] . The Singapore Ministry of Health ( MOH ) sent out circulars to hospitals , government primary care clinics and medical practitioners to apprise them of the dengue situation during epidemic periods , and to provide periodic updates on guidelines for the management of dengue [14–19] . In 2007 , a medical expert committee established the criteria for immediate referral to hospitals , which comprised objective criteria such as significant bleeding , fall in blood pressure , dehydration and/or postural hypotension , a rise in the haematocrit ≥ 20% above the baseline and platelet count < 80 000 cells / mm3 , as well as subjective criteria such as severe vomiting or diarrhoea , severe abdominal pain , and elderly patients with medical co-morbidities who are unwell [15] . Another comprehensive set of hospital referral and admission criteria was also instituted during the 2013 dengue epidemic , and a key change to outpatient management of dengue was the platelet threshold of <60 , 000 / mm3 in adults and <80 , 000 / mm3 in children [17] . Besides the warning signs and symptoms listed in WHO’s dengue guidelines for diagnosis , treatment , prevention and control in 2009 [20] when considering referral to the hospital , MOH recommended to consider additional factors such as persistent fever , dizziness and platelet thresholds [17] . In 2015 , the platelet threshold was lowered to <50 , 000 / mm3 in adults with no warning signs [18] . In this study , we describe the national trend of hospital admissions for dengue and disease severity in terms of length of stay ( LOS ) , admission to the intensive care unit ( ICU ) , death in hospital , and case fatality rate ( CFR ) over a 15-year period from 2003 to 2017 . This will provide a baseline for future in-depth analyses of risk factors associated with adverse outcome to inform local guidelines for referral and hospitalization of dengue cases .
MOH provides clinical criteria for the diagnosis of dengue , and recommends appropriate laboratory tests and clinical management [21] . Under the Infectious Diseases Act in Singapore , it is mandatory for all medical practitioners and clinical laboratories to notify all clinically- or laboratory-confirmed dengue cases to MOH within 24 hours from the time of diagnosis through fax or via a dedicated website [21] . The information required in the notification form includes unique personal identification number , name , date of birth , ethnic group , gender , residential and school or workplace addresses , and dates of diagnosis and onset of illness . If a dengue case is notified to MOH from multiple sources ( e . g . , from clinician and laboratory ) , duplicate records are removed after verification checks based on personal particulars captured in the notifications . Laboratory confirmation of dengue cases is based on non-structural protein 1 ( NS1 ) antigen detection , viral RNA detection by polymerase chain reaction ( PCR ) , or immunoglobulin M detection [22 , 23] . Source reduction remains the key strategy to suppress the vector population in Singapore’s integrated Aedes mosquito control programme , and it entails house-to-house checks , vector surveillance , community education , law enforcement and operational research . An enhanced approach for dengue control has been adopted after a series of reviews of the programme in the past ten years , with focus on inter-epidemic surveillance and control , risk-based prevention and intervention , and coordinated intersectoral cooperation [24] . The National Environment Agency ( NEA ) is responsible for vector surveillance and control . Dengue serotype is determined at the Environmental Health Institute ( EHI ) of NEA and the National Public Health Laboratory based on residual blood samples tested positive for DENV from the national virus surveillance programme [8] . We conducted a retrospective study of notified cases and laboratory confirmed dengue patients admitted to all public and private acute hospitals between 2003 and 2017 , so as to investigate the impact from the review and refinement of admission criteria . Inpatient information from all hospitals in Singapore is captured in the MediClaims database hosted by MOH , which contains electronic medical records that include up to three discharge diagnoses per patient based on the 9th and 10th revisions of the International Classification of Diseases ( ICD ) . MOH conducts annual check on the MediClaims database to ensure its completeness . We obtained the annual number of hospital admissions for all discharge diagnoses of dengue fever ( DF ) and dengue haemorrhagic fever ( DHF ) based on ICD-9 061 and 065 . 4 during 2003–2011 and ICD-10 A90 and A91 during 2012–2017 . Hospitalizations with discharge diagnosis of dengue are mostly laboratory-confirmed , in accordance with the recommendations by MOH for initial evaluation of a patient suspected to have dengue [21] . Analyses of hospitalization data included all diagnosis types ( principal or secondary ) for DF and DHF . It must be noted that the principal cause of death for patients who died in hospital may not be due to dengue . For computation of CFR , the number of deaths due to dengue was obtained based on data from the Singapore Registry of Births and Deaths . Foreigners who came to Singapore to seek medical treatment were excluded from data analyses of dengue cases , hospitalizations and deaths . Annual incidence rates and hospitalization rates of dengue cases were calculated based on the estimated mid-year total population obtained from the Singapore Department of Statistics and expressed as per 100 , 000 population in a given year . The Chi-square test for trend was used to evaluate the difference in proportions over time . We used two-sample independent z-tests to compare proportions between two groups for categorical variables . The Mann–Whitney U test was used to assess differences between any two groups for continuous variables . All statistical tests were two-sided , and statistical significance was taken as p < 0 . 05 . Statistical analyses were performed using R version 3 . 5 . 1 ( R Foundation for Statistical Computing , Vienna , Austria ) . As the data used had been collected for the purpose of mandated national public health surveillance , ethics approval was not sought for the study . All data analyzed were anonymized .
During the 15-year study period , the dengue incidence rate per 100 , 000 population ranged from 47 . 9 in 2017 to 413 . 6 in 2013 ( Table 1 ) . The CFR ranged from 0% in 2017 to 0 . 27% in 2006 . The proportion of dengue cases hospitalized plummeted from 97 . 3% in 2003 to an all-time low of 25 . 6% in 2014 ( p < 0 . 0005 ) . The proportion of deaths among hospitalized patients ranged from 0 . 14% in 2003 to 0 . 77% in 2009 . The median age of fatal dengue cases ranged from 31 to 74 years , while the median age of dengue patients who died in hospital ranged from 52 to 86 years . There were two large epidemics , each stretching over two years , in 2004–2005 and 2013–2014 , both of which were associated with a switch in the predominant serotype from DENV-2 to DENV-1 ( Table 1 ) . Another epidemic in 2007 was associated with a switch from DENV-1 to DENV-2 . The proportion of dengue cases hospitalized during these three epidemic periods was 93 . 2% in 2004–2005 , 58 . 1% in 2007 and 28 . 9% in 2013–2014 . While the 2004–2005 epidemic saw the largest proportion of cases hospitalized , a significant decline in the proportion hospitalized was observed in ensuing years ( p < 0 . 0005 ) ( Fig 1 and Table 1 ) . The gender-specific proportion of hospitalizations among dengue cases was consistently higher for women than for men during the study period , with the exceptions being in 2003 and 2006 ( S1 Table ) . More dengue cases of older age were hospitalized ( S1 Table ) ; the highest proportions were in older adults aged 45–64 years from 2003 to 2012 ( range: 50 . 8% to 100 . 0% ) and elderly persons aged ≥65 years from 2013 to 2017 ( range: 47 . 2% to 55 . 4% ) . In the majority of years after 2007 , the hospitalization rate per 100 , 000 population was highest in the elderly ≥65 years of age . About 0 . 6% to 1 . 5% of dengue patients in hospital were admitted to the ICU during the study period ( Table 1 ) . In the majority of years , this proportion was highest in elderly patients ≥65 years of age ( range 0 . 6% to 3 . 6% ) followed by older adults aged 45–64 years ( range 0 . 8% to 2 . 4% ) ( Fig 2 ) . Men constituted 58 . 2% to 65 . 4% of dengue cases ( S2 Table ) and 53 . 9% to 61 . 2% of hospitalizations ( S3 Table ) . Adults aged 25–44 years comprised about 40% to 50% of dengue cases and hospitalizations . While the proportion of hospitalizations in those aged 25–44 years declined from 46 . 3% in 2003 to 36 . 4% in 2017 , older adults ( aged ≥45 years ) accounted for an increasingly higher proportion over the 15-year period: from 21 . 0% to 29 . 9% in the age group 45–64 years , and from 4 . 9% to 16 . 8% in elderly patients ≥65 years of age ( all p < 0 . 0005 ) ( S3 Table ) . The median age of dengue cases remained about the same from 2007 onwards ( Fig 3 ) . On the other hand , the median age among hospitalized cases increased from 38 years in 2007 to 43 years in 2017 ( p < 0 . 0005 ) ( Fig 3 ) . The total number of hospital bed-years due to dengue halved from 136 in 2005 to 69 in 2013 ( Fig 4 ) . During the period from 2003 to 2017 , the overall mean LOS in hospital was 3 . 8 days ( range 1 to 121 days ) and the overall median was 3 days ( interquartile range [IQR] 2 to 5 days ) . The LOS in hospital remained stable over the years; both the annual mean and median LOS were 3 to 4 days . During the 15-year period , the overall mean ICU stay was 4 . 0 days ( range 0 to 65 days ) and the overall median was 2 days ( IQR 1 to 4 days ) . The annual mean ICU stay was 2 to 7 days and the annual median was 2 to 3 days . Both the annual mean and annual median LOS in hospital for the age group 25–44 years was 3 to 4 days . The annual mean LOS for older adults aged 45–64 years was 3 to 5 days and the annual median was 3 to 4 days . Hospital stays among elderly patients ≥65 years of age were significantly longer: the annual mean LOS was 5 to 8 days and the annual median was 4 to 6 days ( p < 0 . 0005 ) .
The proportion of dengue cases hospitalized saw a sharp reduction in recent years despite the larger scale of dengue epidemics . The proportion hospitalized in the last five years of the study period ranged from 25 . 6% ( in 2014 ) to 35 . 9% ( in 2017 ) , which was significantly lower than the 41 . 9% ( in 2012 ) to 50 . 6% ( in 2009 ) in the five-year period after the 2007 epidemic ( Table 1 ) . While there was a drastic decline in the proportion of dengue cases hospitalized due to refinement of hospital admission criteria in the aftermath of the 2004–2005 epidemic , there has been no concomitant increase in adverse outcomes , indicating that patients with more severe dengue requiring inpatient care and monitoring are being accurately identified and appropriately referred and admitted . Over the 15-year period , the proportion of patients admitted to the ICU remained below 2% , and the proportion who died in hospital was less than 1% . Early diagnosis of dengue was facilitated by NEA’s EHI , which offered NS1 antigen testing to primary care clinics at no cost from 2006 [25] . A series of annual educational seminars targeted at primary care practitioners has been held since 2011 in conjunction with the ASEAN Dengue Day . These initiatives facilitate early diagnosis and close monitoring of dengue in the primary healthcare setting [23 , 26] . Surveys of primary care practices over two time periods in 2011 and 2014 demonstrated increased confidence and better management of dengue among primary care practitioners , with fewer referrals to hospital [26 , 27] . Despite the implementation of more selective admission criteria by the hospitals , we did not observe a concomitant increase in adverse outcomes nationally , as shown by the overall CFR ( Table 1 ) . This may be attributed to a combination of factors , including early diagnosis , improved dengue clinical management in primary care settings , and more appropriate referral and hospital admission [11 , 25–27] . Early diagnosis and right-siting of dengue case management can help to optimize usage of limited healthcare resources while averting negative outcomes . The new admission criteria implemented in a tertiary care public hospital in 2007 resulted in a median cost saving of US$1 . 4 million ( 90th percentile US$2 . 7 million ) to patients in 2008 [13] . The impact of dengue epidemics on bed utilization rates was considerable even with the significant reduction in the proportion of dengue cases hospitalized . In the non-epidemic year 2012 , there were 1 , 931 hospitalizations accounting for 20 bed-years ( Fig 4 ) . In comparison , there were 7 , 054 and 4 , 572 hospitalizations accounting for 69 and 42 bed-years in the most recent epidemic years of 2013 and 2014 respectively ( Fig 4 ) . Our study revealed that a higher proportion of older dengue cases ( aged ≥45 years ) were admitted , and that older patients experienced more severe disease . The gap in median age between all dengue cases and hospitalized cases has widened since 2009 ( Fig 3 ) . Elderly patients ≥65 years of age comprised 17% of dengue hospitalizations in 2017 , compared to 5% in 2003 ( S3 Table ) . This was likely contributed to by both the MOH guideline to refer for hospital evaluation those aged ≥65 years as well as a shift in demographic profile of dengue cases towards older adults [17–19] . The proportion of dengue cases hospitalized among the elderly aged ≥65 years has been highest since 2013 ( ranging from 47% to 55% ) ( S1 Table ) . The proportion of dengue patients who died in hospital was also highest in the elderly , while the proportion admitted to the ICU in elderly patients was one of the highest of any age group . The ageing population has led to additional challenges in the clinical management of dengue cases [28] . A retrospective study of all adult dengue patients managed at a tertiary care public hospital between 2005 and 2008 found that elderly patients had atypical clinical presentations and were at higher risk of DHF , severe disease and hospital acquired infection ( HAI ) [29] . The factors contributing to prolonged LOS were dengue severity , age , comorbidity and HAI [29] . Our study revealed that the implementation of more selective admission criteria by the hospitals had not led to longer LOS; the overall mean was 3 . 8 days and the overall median was 3 days ( IQR 3 days ) from 2003 to 2017 . The overall proportion of hospitalized dengue cases who died over the 15-year period was 0 . 29% . In Malaysia , notifications for dengue infection represent only a small fraction of dengue incident cases ( 0 . 7% to 2 . 3% ) , and the proportion of dengue hospitalizations among estimated incident cases was about 3 . 0% to 5 . 6% based on data from 2001 to 2013 [30] . The mean LOS of dengue cases admitted to a tertiary care teaching hospital in Kelantan state of Malaysia during a six-year period from 2008 to 2013 was 4 . 88±2 . 74 days ( median 3 , IQR 3 , range 1–35 days ) , and 1 . 1% of the hospitalized cases died [31] . A retrospective cohort study using the National Inpatient Sample , the largest all-payer database of hospital discharges in the United States from 2000 to 2007 , found that the median LOS of hospitalized patients diagnosed with dengue was 3 days , with a range from 0 to 35 days [32] . A local study found that there has been a shift in the health-seeking behaviour of patients with dengue towards primary care: the proportion of dengue cases who sought medical attention at primary care clinics increased significantly from 14 . 8% in 2006 to 35 . 2% in 2015 , while those who sought hospital care declined from 71 . 7% to 48 . 8% [25] . We believe this contributed to the significantly lower proportion of dengue cases admitted to hospital after the 2007 epidemic as more dengue cases were appropriately treated and monitored by primary care physicians . Another study reported an increase in the proportion of DHF from 6% in 2004 to 21% in 2007 among adult dengue patients admitted to a tertiary care public hospital , which suggested an improved triage system that accurately identified those patients requiring inpatient care and monitoring [33] . The main strength of this study lies in the use of data from a national surveillance system for notification of dengue cases and a hospitalization database hosted by MOH for the purpose of capturing all patient discharge information submitted by accredited institutions ( both public and private ) . There are a few limitations to our study . Clinical data of hospitalized dengue cases were not available for analysis in our study . Using WHO case classification criteria for diagnosis of DHF published in 1997 [34] , a considerable proportion of DHF was either notified or diagnosed as DF , thereby its value as an indicator of disease severity was reduced . A systematic review involving studies in different countries and expert consensus meetings has suggested that the 1997 classification of DF , DHF and dengue shock syndrome ( DSS ) does not fully represent levels of disease severity [35] . We therefore opted to measure disease severity only in terms of admission to the ICU and deaths . While we could not determine if there had been delay in admission of dengue cases due to the more selective admission criteria resulting in adverse outcomes , those with more severe condition would most likely end up in hospital since there is good access to different levels of healthcare services in Singapore . Moreover , all dengue deaths reported to MOH are investigated and cross-checked with death data from the Singapore Registry of Births and Deaths , regardless of whether these fatal cases have been hospitalized or not . Comparison of trends may be limited by changes in laboratory tests for dengue and diagnostic practices over time and the advent of rapid laboratory diagnostic tools in later years . While the commonly used NS1 antigen assay was found to have high specificity , its sensitivity was significantly lower particularly in secondary infection [36] . During the dengue epidemic in 2007 , cross-sectional seroepidemiologic surveys conducted in seven outbreak areas in Singapore revealed an overall inapparent dengue rate of 78% [37] . This study using case surveillance notifications does not capture all DENV infections . The proportion of persons with dengue who were hospitalized would be lower if inapparent dengue infections were included instead of only symptomatic dengue cases notified to MOH . There may be hospital discharges for dengue which were not notified to MOH , but this proportion is expected to be small as dengue is a notifiable disease under the Infectious Diseases Act . In conclusion , the drastic decline in proportion of dengue cases hospitalized due to more selective admission criteria has not led to a corresponding increase in adverse consequences . Our study suggests that following the nadir in the proportion of dengue cases hospitalized after 2007 of about 26% , the subsequent increase after the 2013–2014 dengue epidemic was likely contributed to by both an increased risk of disease among the older population as well as the inclusion of age ≥65 years as a specific indication for hospitalization . In the light of Singapore’s ageing population and high hospital bed occupancy rates , coupled with periodic dengue epidemics , it is important to manage unexpected surges in dengue cases and ameliorate the strain on hospital systems with appropriate referral for treatment and right-siting of care . Further studies are needed to improve dengue management in older adults and further identify risk factors of severe disease in this age group .
|
Following the review of hospital admission criteria for dengue cases in the aftermath of the 2004–2005 dengue epidemic in Singapore , the proportion of dengue cases hospitalized plummeted from 72 . 6% in 2006 to the nadir of 25 . 6% in 2014 . There was no concomitant increase in adverse outcomes as a result of the more selective admission criteria . Median length of stay remained stable over the years; overall stay was 3 to 4 days and intensive care unit ( ICU ) stay was 2 to 3 days . Among hospitalized dengue patients , less than 2% were admitted to the ICU . Overall case fatality rate was low and remained below 0 . 5% . Elderly patients aged 65 years and older constituted the highest proportions of dengue cases hospitalized and patients admitted to the ICU .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"hospitalizations",
"medicine",
"and",
"health",
"sciences",
"elderly",
"geographical",
"locations",
"tropical",
"diseases",
"health",
"care",
"singapore",
"age",
"groups",
"neglected",
"tropical",
"diseases",
"infectious",
"diseases",
"intensive",
"care",
"units",
"primary",
"care",
"dengue",
"fever",
"hospitals",
"people",
"and",
"places",
"health",
"care",
"facilities",
"asia",
"population",
"groupings",
"viral",
"diseases",
"geriatrics"
] |
2019
|
A 15-year review of dengue hospitalizations in Singapore: Reducing admissions without adverse consequences, 2003 to 2017
|
A fundamental problem in developmental biology concerns how multipotent precursors choose specific fates . Neural crest cells ( NCCs ) are multipotent , yet the mechanisms driving specific fate choices remain incompletely understood . Sox10 is required for specification of neural cells and melanocytes from NCCs . Like sox10 mutants , zebrafish shady mutants lack iridophores; we have proposed that sox10 and shady are required for iridophore specification from NCCs . We show using diverse approaches that shady encodes zebrafish leukocyte tyrosine kinase ( Ltk ) . Cell transplantation studies show that Ltk acts cell-autonomously within the iridophore lineage . Consistent with this , ltk is expressed in a subset of NCCs , before becoming restricted to the iridophore lineage . Marker analysis reveals a primary defect in iridophore specification in ltk mutants . We saw no evidence for a fate-shift of neural crest cells into other pigment cell fates and some NCCs were subsequently lost by apoptosis . These features are also characteristic of the neural crest cell phenotype in sox10 mutants , leading us to examine iridophores in sox10 mutants . As expected , sox10 mutants largely lacked iridophore markers at late stages . In addition , sox10 mutants unexpectedly showed more ltk-expressing cells than wild-type siblings . These cells remained in a premigratory position and expressed sox10 but not the earliest neural crest markers and may represent multipotent , but partially-restricted , progenitors . In summary , we have discovered a novel signalling pathway in NCC development and demonstrate fate specification of iridophores as the first identified role for Ltk .
Understanding mechanisms determining the selection of specific fate choices by multipotent precursors is of fundamental importance in developmental and stem cell biology . Neural crest cells ( NCCs ) are a favoured model for investigation of fate specification mechanisms , being multipotent precursors of diverse cell-types , including craniofacial cartilage , peripheral neuronal and glial cell-types and pigment cells [1] . The mechanism driving specification of multipotent progenitors in the neural crest ( NC ) to fate-restricted cell types is controversial . Multipotent NC stem cells with broad potential have been isolated from embryos , even from post-migratory locations , leading to the hypothesis of direct fate restriction , whereby local signals instruct multipotent stem cells to adopt specific fates ( reviewed in [2] . In contrast , numerous studies indicating that NCCs include partially-restricted cells has suggested progressive fate restriction as an alternative model ( reviewed in [3] , [4] ) . Thus , multipotent precursors gradually lose the potential to generate certain derivative cell-types , forming partially-restricted precursors before eventually becoming specified to an individual fate . The number and character of these intermediate precursors in vivo remains largely undefined . The molecular mechanisms underlying fate restriction also remain poorly understood . Genetic analysis in mouse and zebrafish identifies key transcription factors required for specification of several or individual fates . Perhaps the best characterised example is that of Microphthalmia-related transcription Factor ( Mitf ) , which is pivotal for melanocyte specification [5] , [6] , [7] . Sox10 is required in multipotent NCCs to drive transcription of Mitf and other transcription factors ( e . g . [8] , [9] , [10] , [11] , [12] ( reviewed in [13] , while Pax3 acts synergistically with Sox10 to regulate the mouse Mitf promoter [11] , [14] . Extracellular signals are also important in NCC fate specification . For example , Wnt signals are required for melanocyte specification via transcriptional activation of Mitf [15] , [16] , [17] . Signalling by these pathways acts together with intrinsic factors such as Sox10 and Pax3 to induce specific cell-fates . Further signals remain to be identified and for some NC-derived fates , including other pigment cell-types , no fate specification factors have yet been identified . Surprisingly , even where key factors have been identified , in most cases the receptors mediating these signals are unknown . Leukocyte tyrosine kinase ( LTK ) was first identified as an insulin receptor-like receptor tyrosine kinase ( RTK ) expressed in mouse haematopoietic cells [18] . Within the insulin receptor superfamily , LTK is most closely-related to anaplastic lymphoma kinase ( ALK ) [19] . In mammals LTK's function remains unknown , although it is expressed in pre-B and B lymphocytes and in the adult brain [18] , [20] . It is widely expressed in human leukaemias [21] and is a candidate locus contributing to the multigenic autoimmune disease , systemic lupus erythematosus ( OMIM#152700 ) [22] . Mutations in the zebrafish shady ( shd ) locus were identified in a large-scale mutagenesis screen [23] . Iridophores , an iridescent pigment cell widespread in anamniotes , are reduced in number in shd mutants . Here we show that other NC derivatives are not affected in shd mutants . We demonstrate that shd encodes the zebrafish orthologue of LTK and functions cell-autonomously within the NC . We show that strong shd mutants lack iridoblast lineage markers , including ltk which is highly reduced from the very earliest stage in premigratory NCCs . Later some NCCs die by apoptosis , after failing to become specified to iridophore or other fates . Zebrafish sox10 ( also known as colourless ) mutants share the strong iridophore phenotype of shd mutants and show consistent defects in fate specification of non-ectomesenchymal NC derivatives [9] , [24] , [25] , [26] , [27] . We have previously proposed that fate specification defects may underlie the sox10 mutant iridophore defect [13] , [26] . As expected , late iridophore markers are absent . At earlier stages ltk expression is detected in an increased number of cells compared with wild-type ( WT ) siblings . Our data indicates that these ltk-expressing cells are most likely partially-restricted precursors . Together , these observations identify the first loss of function mutants for this poorly characterised RTK and suggest that Ltk mediates iridophore fate specification from multipotent NCCs .
shd mutant embryonic phenotypes formed a clear allelic series; homozygotes for strong alleles ( e . g . shdty82 ) have very few ( <3 ) iridophores and die as larvae , those for weaker alleles ( e . g . shdty9 ) show reduced numbers of differentiated iridophores in the embryo and are adult viable , but phenotypically normal ( Figure 1A , 1B , 1D , 1E , 1G , 1H ) . In all cases , and in contrast to all iridophore mutant phenotypes identified before [23] , remaining iridophores in shd mutants were invariably normally pigmented and hence appeared normally differentiated . Independent screens for adult pigment pattern mutants identified further mutants with reduced iridophores in the body and eyes ( j9s1 , j9e1 and j9e2 ) . The most severe of these , j9s1 , also lacked late stripe melanophores , and appeared similar to the roseb140 phenotype [28]; SLJ unpublished ) ( Figure 1C , 1F , 1I ) . Mapping and complementation testing showed these to be allelic to shd ( data not shown ) . These adult viable alleles had no detectable abnormal embryonic phenotype as homozygotes , although in transheterozygous combination with shdty82 they showed an embryonic iridophore phenotype of intermediate strength ( data not shown ) ; thus the adult viable alleles are also hypomorphic alleles . Iridophores are derived from the NC [29] . However , both of the other pigment cell-types , melanophores and xanthophores , as well as peripheral neuronal , glial and skeletogenic derivatives examined with specific markers showed comparable numbers and patterns in shd mutants and WT siblings ( Figure S1 ) . We conclude that the shd mutant phenotype is restricted within the NC to the iridophore lineage . We then used genetic chimaeras formed by transplanting WT cells , labelled with rhodamine and biotin-conjugated dextran beads , into shdty82 mutants to ask whether the shd iridophore phenotype resulted from cell-autonomous function within the NC [27] . Of 291 host embryos receiving WT cells , 188 survived the procedure to 3 dpf and could be scored for iridophore phenotype . Of these , 40 embryos ( 21 . 2% ) were identified as shd mutant hosts by the general absence of iridophores and 22 were chimaeric , containing some rhodamine fluorescent cells . Of these , seven embryos had iridophore counts above normal shdty82 mutant levels ( Table S1 ) . In these individuals , most or all iridophores exhibited biotin tracer and were derived from WT donor cells , consistent with cell-autonomous shd gene function ( Figure S1I , S1L , and S1O ) . As expected , some , but importantly not all , of these embryos also had up to two unlabelled iridophores in the expected sites ( dorsal stripe or lateral patch ) for the occasional ‘escaper’ iridophores seen in shdty82 mutants . Thus , these studies demonstrated that shd acts cell-autonomously in iridophore lineage development . In the absence of candidate genes , we utilised a positional cloning approach to identify shd . Linkage analysis mapped shd within 0 . 1 cM of marker z10985 on linkage group 17 ( Figure 2A ) . Assuming average recombination frequencies , we reasoned that this marker likely lay within c . 70 kb of the shd gene and that both might be contained within the insert of a single genomic PAC . We isolated three PACs containing z10985 from the PAC706 library [30] . Pulse field gel electrophoresis analysis of the PAC inserts defined an overlapping contig spanning 207 . 5 kb of genomic sequence ( Figure 2B ) . Microinjection of PAC DNA into 1- or 2-cell stage zebrafish embryos demonstrated that PAC3 , but not the others , partially rescued the shdty82 mutant iridophore phenotype ( Figure 2C ) , indicating that PAC3 contained a functional shd gene . Analysis of the PAC contig sequence ( Sanger Centre zebrafish genome project ) using the NIX gene prediction package identified one gene fully contained within PAC3 . This gene encodes an RTK , of the insulin receptor class and most similar to human ALK and LTK ( Figure 2E ) . Sequencing cDNAs for this ALK/LTK-like gene from AB WT embryos identified distinct isoforms generated by alternative splicing ( Figure S2 ) . BLAST searches identified closely-related RTKs from the zebrafish ( XM_686872 , XM_001342889 and XM_687805 ) and other vertebrates ( Table S2 ) in the NCBI databases . We used multiple sequence alignments and phylogenetic estimation protocols to determine the likely relationships between our cDNAs and these genes ( Figure 2H , Figure S3 ) . Our phylogenetic analyses identified; i ) the invertebrate genes as a clear outgroup to the vertebrate homologues; ii ) zebrafish sequence XM_686872 , on chromosome 17 but physically distant from shd , as the zebrafish ALK orthologue , which we name alk; iii ) our cDNAs and XM_001342889 and XM_687805 as zebrafish Ltk sequences . These latter genes resolve in build 7 of the zebrafish genome ( Zv7 ) to one locus , the zebrafish orthologue of LTK , which we name ltk . We note that while all mammalian Ltk proteins are predicted to lack MAM ( meprin/A5/µ ) domains , thought to mediate protein-protein interactions [31] , the predicted chicken and zebrafish Ltk proteins , and all Alk proteins , possess them ( Figure 2E ) . These data strongly suggested that the shd mutations identified the ltk gene . Injection of WT embryos with a translation-blocking ltk morpholino oligonucleotide generated strong morphant phenocopies of shd mutants in a dose-dependent manner ( Figure 2D; data not shown ) . Embryos injected with double doses ( 18 . 4 or 36 . 8 ng ) of a 5 bp mismatch morpholino showed no phenotype ( 0/406 injected ) , whereas siblings injected with single doses ( 9 . 2 or 18 . 4 ng ) of ltk morpholino showed a substantial proportion of embryos with a severe loss of iridophores ( 27/264 injected ) , as well as many with weaker phenocopies . The knockdown phenotype precisely phenocopied the shd mutant allelic series , supporting the conclusion that shd is ltk . Confirmation that shd is ltk came from identification of ltk point mutations in three shd mutant alleles by sequencing the Ltk coding region from RNA extracted from shd homozygotes ( Figure 2E , Figure S2 ) . The shdty82 mutation 2356A>T ( taking the A in the translation initiation codon as +1 ) results in a premature STOP codon , generating a truncated predicted protein lacking the tyrosine kinase domain , consistent with the strong mutant phenotype . This mutation fortuitously generates an RFLP allowing us to confirm that segregation of this mutation correlated perfectly with the embryonic phenotype ( Figure 2F ) . shdj9e2 mutates a splice donor site , resulting in a transcript with an in-frame deletion of exon 26 encoding a variant of Ltk in which the tyrosine kinase domain activation loop [32] is deleted ( Figure 2G ) . It is likely that this variant protein will be less readily activated , consistent with the weak mutant phenotype . shdj9s1 mutation , 2275C>T , results in a P759S substitution in the extracellular region . This nucleotide change is the only difference between the shdj9s1 allele and the cDNA sequence for the WT C32 allele that the mutation was isolated on ( data not shown ) . Moreover , this nucleotide change is not found on the 3 sequenced BACs or PACs , each derived from a different haplotype , that span this region , further supporting that the P759S substitution generates a mutant protein . Interestingly , this proline residue is conserved in chicken and mammalian LTKs , as well as in the LTK orthologue Drosophila Alk , but not in the corresponding tetrapod ALKs . Mutations affecting this residue in Drosophila Alk have not been reported to date [33] . Together , our data unambiguously identify shd as zebrafish ltk , showing that the shd mutant iridophore phenotype results from loss of Ltk signalling within the developing NC . This is the first time that a vertebrate Ltk loss of function phenotype has been defined . To clarify the role of ltk in iridophore development , we determined the spatiotemporal pattern of ltk gene expression by whole-mount mRNA in situ hybridisation ( ISH ) . Our cell-autonomy studies predicted NC expression of ltk , but expression might be restricted to differentiated iridophores or be found at earlier stages in NC development . Here we focus on expression in NCCs and their derivatives , but we also saw ltk expression in notochord from 18–24 hpf ( Figure 3A–3C ) and prominently in brain and swim bladder from 3 dpf ( data not shown ) . From 48 hpf onwards , ltk-expressing cells formed a series of spots along the dorsal and ventral stripes , as well as on the eye , a pattern strikingly reminiscent of differentiated iridophores ( Figure 1A and 1B , Figure 3J , 3O; 3XY and RNK , unpub . data ) . Consistent with this , the pattern was identical to that of ednrb1 , the only characterised iridophore marker ( Figure 3O , 3U , 3AA , and 3AE ) [34] . To test definitively if iridophores express ltk , we photographed the dorsal stripe iridophore pattern of individual WT embryos at 72 hpf , processed the embryos for ltk expression and then photographed the ltk pattern; ltk-expressing cells ( Figure 3J ) and differentiated iridophores ( Figure 3E ) showed an excellent correlation . Thus , at least in these late stages , ltk expression in NC derivatives is restricted to iridophores . Initial ltk expression in a subset of NCCs was seen near the eye at 18–24 hpf ( Figure 3A–3C ) . Between 26 and 30 hpf , these cells spread over the pigmented retinal epithelium from the dorsal surface of the eye ( Figure 3F , 3K , and 3Q ) . This widespread scattered distribution was then maintained , but the density in a ring around the lens increased ( data not shown ) , consistent with the WT pattern of corneal iridophores ( Figure 1A ) . Plastic sections showed cells on the eye were superficial to the pigmented retinal epithelium ( Figure 3W ) , consistent with a NC origin . Expression of ltk in trunk and tail NC at early stages was very dynamic , with transient expression in a subset of premigratory NC spreading progressively more posteriorly between 18 and 28 hpf ( Figure 3A–3C , 3H , and 3M ) . These premigratory NCCs were bilaterally arranged dorsolateral to the neural tube ( Figure 3D ) . From 26 hpf onwards , a few ltk-expressing cells were migrating on the medial , but never the lateral , migration pathway , as expected for iridoblasts ( Figure 3H , 3M , and 3S ) [29] . As early as 30 hpf , the pattern of strongly ltk-expressing cells in whole mount embryos and confirmed in plastic sections closely mimiced that later seen for iridophores ( Figure 3S and 3X ) . For example , in the dorsal stripe , cells were medially positioned and somewhat regularly spaced along the posterior trunk and tail . Hence , we interpreted these cells as iridoblasts and suggest that ltk-expression marks the iridophore lineage throughout their development . However , NC expression of ltk is seen very early in a subset of premigratory NCCs , which may include a subset of multipotent NCCs . shdty82 mutants usually showed no iridophores and hence a primary role for ltk in iridoblast proliferation was unlikely , since this would predict only a reduction in iridophores . Using phosphohistone H3 as a marker for proliferating cells , we were unable to detect a significant effect on NCC proliferation ( Figure S4 ) . To address a role in iridophore differentiation , we examined both known iridoblast markers , ednrb1 and ltk , reasoning that if ltk function was required only for iridophore differentiation these early markers would still be expressed normally in shd mutants . However , no ednrb1- or ltk-expressing cells were seen in shdty82 mutants at 50–72 hpf ( Figure 3P , 3V , 3AB , and 3AF ) and only reduced numbers in the weaker shdty9 mutants ( data not shown ) . We were unable to use ednrb1 as an iridophore marker at earlier stages since it is expressed in cells of multiple pigment cell lineages prior to 48 hpf [34] . Instead , we examined ltk expression in earlier embryos . At these stages shdty82 mutants could not be directly distinguished , but , from c . 20 hpf onwards , the expected 25% of embryos showed a consistent phenotype of severely reduced numbers of ltk-expressing cells . That these were shdty82 homozygotes was confirmed by RFLP genotyping of 20–24 hpf embryos prior to whole-mount in situ analysis; all homozygous WTs ( n = 68 ) showed normal ltk expression ( Figure 3B , 3C , 3H , 3M , 3S , 3Y , and 3AC; data not shown ) , wheareas all shdty82 homozygotes ( n = 67 ) showed the reduced pattern ( Figure 3I , 3N , 3T , 3Z , and 3AD; data not shown ) . Mutants consistently showed three main features: i ) ltk-expressing cells failed to spread across the eye and remained low in number ( Figure 3G , 3L , and 3R ) ; ii ) strong ltk expression was absent , with expression restricted to at most a few faintly expressing cells ( Figure 3I , 3N , and 3T ) , except iii ) from 35 hpf , a variable but always greatly reduced number of strongly expressing ‘escaper’ cells in the anterior trunk ventral stripe ( i . e . residual lateral patch cell clusters ) ( data not shown ) . In summary , throughout the stages when NCCs in WTs are specified to individual fates and begin to differentiate , shdty82 mutants showed a consistent phenotype of highly reduced numbers of ltk-positive cells , with ‘escaper’ cells with normal ltk expression restricted to a few cells on the dorsal eye and in the residual lateral patches . Both the presence of these escaper cells and the similar strong reduction in numbers of ltk-expressing cells in ltk morphants ( JM and RNK , data not shown ) argue against the possibility that absence of ltk-expressing neural crest cells reflects nonsense-mediated decay of ltk transcripts in this mutant . Thus , specification of almost all iridoblasts fails in shdty82 mutant embryos . We then investigated the fate of NCCs that failed to become specified as iridoblasts . In mitfa/nacre mutants , melanophore fate specification fails and increased iridophore numbers are seen , perhaps due to multipotent melanophore precursors adopting an iridophore fate in elevated numbers ( Lister et al . 1999 ) . Hence we considered whether some shd mutant iridoblast precursors might adopt another pigment cell fate . The late melanophore pattern in shdty82 mutants is overtly normal ( see Figure S1C , S1D , S1E , and S1H ) , and counts of melanophores in the dorsal stripe of 3 dpf shdty82 mutants ( mean±s . d . = 87 . 4±2 . 06 , n = 33 ) and their WT siblings ( 82 . 4±2 . 19 , n = 31 ) showed no significant difference ( Student's t-test , p = 0 . 100 ) ( Figure 4B ) . We then considered the possibility that an overproduction of melanoblasts or xanthoblasts at early stages might later be compensated by regulative processes . Hence , we asked whether shdty82 homozygotes showed elevated melanoblast or xanthoblast numbers at 30 hpf compared with homozygous WT siblings ( Figure 4C and 4D ) . Embryos were genotyped prior to mRNA whole mount ISH for , respectively , dopachrome tautomerase ( dct; [35] or guanine cyclohydrolase ( gch; [36] . Whilst dct is expressed exclusively in the melanophore lineage , gch expression , like other characterised xanthophore markers , is seen transiently in the melanophore lineage , as well as being strongly upregulated in xanthophore lineage cells [36] . In order to ensure that our counts focused as much as possible on the xanthophore lineage , we confined our attention to gch-expressing cells on the lateral pathway , since xanthoblasts do not utilise the medial pathway [29] . Counts of dct-positive melanoblasts in shdty82 homozygotes ( 128 . 0±3 . 77 , n = 35 ) and homozygous WT siblings ( 125 . 3±3 . 42 , n = 39 ) were statistically indistinguishable ( Student's t-test , p = 0 . 598 ) . Similarly , counts of gch-positive xanthoblasts in shdty82 homozygotes ( 97 . 8±5 . 65 , n = 23 ) and homozygous WT siblings ( 90 . 5±4 . 50 , n = 23 ) were indistinguishable ( Student t-test , p = 0 . 315 ) . Thus we found no evidence for a shift of iridoblast precursors to either a melanophore or xanthophore fate . In sox10 mutants , neural and pigment cell precursors that fail to become fate-specified are later ( 35–45 hpf ) lost by apoptosis [26] . We explored whether any NCCs in shdty82 mutants were lost by apoptosis . We generated shdty82 fish carrying the 7 . 2sox10∶gfp transgenic line ( J . R . Dutton and R . N . K . , in prep . ) , in which a 7 . 2 kb fragment of the zebrafish sox10 promoter [9] drives expression of GFP , robustly labelling all NCCs . We combined the TUNEL technique with immunofluorescent detection of GFP on 35–50 hpf embryos from crosses of shdty82 heterozygotes carrying this transgene , scoring TUNEL+/GFP+ NCCs in each embryo at a single time-point . Most embryos showed none , but approximately 25% ( 20/89 ) showed one double-labelled cell , consistent with the idea that NCC death was a feature of shd homozygotes . To test directly the hypothesis that NCC death was characteristic of shd mutants , we counted dying NCCs in the trunk and tail of live embryos from such crosses . Approximately 25% of embryos showed up to 2 cells with an apoptotic morphology in either premigratory or medial pathway GFP+ NCCs at 30–50 hpf . We sorted such fish and genotyped them by iridophore phenotype at 3 dpf ( Figure 4A ) . This data confirmed that apoptosis of NCCs was significantly elevated ( Student's t test , p<0 . 0001 ) in shdty82 homozygotes ( mean±s . d . = 0 . 41±0 . 125 , n = 22 ) compared with their WT siblings ( 0 . 02±0 . 016 , n = 61 ) . Although the number of dying cells recorded in any individual embryo was low , we note that embryos were only examined once . Given that apoptotic morphology is a very transient characteristic ( RNK , unpub . obs . ) , our observations indicate that a significant number of NCCs that failed to become iridoblasts were most likely lost by apoptosis in shd mutants . In sox10t3 mutants iridophores are almost invariably absent , whereas in sox10m618 occasional , normally differentiated escaper iridophores are seen in the dorsal and ventral stripes [27]; the iridophore phenotype is thus directly comparable with that of shdty82 mutants ( Figure 5A and 5B ) . Furthermore , our previous studies of the sox10 mutant neural crest phenotype showed several features shared with the shd phenotype , specifically the absence of fate-switching and late death of neural crest cells[26] . Given the general failure of fate-specification of non-ectomesenchymal derivatives in sox10 mutant fish and mice [8] , [9] , [10] , [11] , [12] , [13] , [24] , [25] , [37] , [38] , [39] , [40] , we expected iridoblast specification markers to be absent from the earliest stages in sox10t3 mutants . Hence , we examined ltk expression in sox10t3 mutants and their WT siblings ( Figure 5C–5H ) . Unlike WTs , but just like in shdty82 mutants , sox10t3 mutants almost entirely lacked ltk expression on the eye and in the dorsal , ventral and yolk sac stripes at 48 hpf , although a few escaper cells were seen in the anterior trunk ventral stripe ( Figure 5G and 5H ) , as expected since unspecified pigment cell precursors undergo apoptosis [26] . In contrast , a striking , but unexpected , ltk phenotype was seen at earlier stages in these mutants . In WT siblings at 24 and 30 hpf , ltk expressing cells are seen in the trunk and tail in premigratory NCCs , on migration , in the ventral stripe or clustered behind the otic vesicle ( Figure 5C and 5E ) . In contrast , in sox10t3 mutants , ltk-expressing cells were increased in number compared with WT siblings , and were restricted to premigratory NC ( Figure 5D and 5F ) . Our previous single NCC labelling studies showed that a large proportion of NCCs in sox10 mutants fail to migrate , a defect also detected by whole-mount ISH for sox10 [26] . We asked whether these sox10 mutant NCCs might be trapped in an early NCC state , but two very early NC markers , snail2 and foxd3 [41] , [42] , showed identical expression in sox10t3 mutants and their WT siblings ( Figure 6A–6H ) . In addition , cells expressing these early NC markers were located more posteriorly than the ltk-expressing cells in both sox10 mutants and WT siblings ( Figure 6E–6J ) . We conclude that ltk is not expressed in the developmentally youngest NCCs , but only in those at a slightly more mature stage . Early sox10 expression is widespread in NCCs [26] . Furthermore , sox10 expression is associated with multipotency of NC stem cells [39] . We asked whether ltk-expressing cells in sox10 mutants showed sox10 expression ( Figure 7 ) . For these experiments we used sox10m618 embryos , which also show NCCs trapped in a premigratory position , since sox10 transcripts are apparently destabilised in sox10t3 mutant embryos [26] . Interestingly , many of these sox10-expressing cells showed ltk-expression at 30 hpf ( Figure 7C and 7D ) . In contrast , WT siblings had largely mutually exclusive sox10 and ltk expression , suggesting that ltk-expressing cells were specified iridoblasts by this stage in WT embryos ( Figure 7A and 7B ) . Together these data show that ltk expression comes on in NCCs after the early premigratory NC markers snail2 and foxd3 , and that it labels cells that have developed beyond the initial premigratory NCC state , but which may retain multipotency .
We present a combination of genetic mapping , morpholino-mediated knockdown , and molecular lesion data that unambiguously identifies the shd locus as encoding zebrafish ltk . In mammals , ALK and LTK are distinguished by the presence or absence respectively of MAM domains . However , our phylogenetic analysis shows that presence of MAM domains is the ancestral condition in the ALK/LTK subfamily . Their loss is unique to mammalian Ltks , and not generally diagnostic of the LTK family . Our phylogenetic analysis also suggests that vertebrate Alk and Ltk arose by a gene duplication event early in the vertebrate lineage and are thus co-orthologues of the Drosophila Alk and C . elegans T10H9 . 2 genes . The functions of this subfamily of RTKs remain poorly understood . Human LTK is expressed in pre-B lymphocytes and various other tissues , but its endogenous function remains entirely unknown . In Drosophila , Alk signalling specifies visceral muscle pioneers [43] , [44] and regulates axonal guidance [45] and in C . elegans it functions in synapse differentiation [46] . Mouse Ltk knockouts have not been described . Thus we identify the first vertebrate model for studying ltk gene function . We demonstrate a key role for ltk in NC-derived pigment cell development . We see initial low level ltk expression in a subset of premigratory NCCs followed by persistent robust expression in the iridophore lineage . Furthermore , iridophore lineage markers are absent from shd/ltk mutants , suggesting a very early role in iridophore development , most likely fate specification of iridoblasts from multipotent neural crest cells . Previously , there has been no data concerning the timing when iridophore fate specification occurs , although an informative comparison can be made with melanophore development for which fate specification can be defined precisely as the time when mitfa is first expressed . The timing of this varies along the antero-posterior axis , but at approximately 21 hpf has just begun in the posterior trunk [47] . Two other very early markers of melanoblasts , dct and kit , also begin to be expressed in this region from approximately 21 hpf [35] , [47] , [48] . Thus , in the case of the melanophore , specification begins in the posterior trunk approximately 21 hpf and this is also reflected by expression of two other very early melanoblast marker genes . We show that ltk is already expressed more broadly , throughout the posterior trunk , at this stage . In an mitfa mutant , defects in melanoblasts are seen from 23 hpf at least [47]; in contrast , defects in melanoblast markers in strong mutants for the kit gene , an RTK important for melanoblast survival , are absent at 24 hpf , but detectable at 36 hpf [35] . In addition , the severity of the shdty82 mutant phenotype , with iridophores totally absent , fits well with the mitfa mutant phenotype , but contrasts with that of the survival mutant , kit . Thus , using the best analogy available , the nature and timing of the defects in shd/ltk mutants , being already visible at 20 hpf , best reflects a defect in iridoblast specification . Previous identification of numerous mutants affecting iridophore development shows many in which iridophore differentiation is clearly abnormal , with cells looking duller or whiter than in WT siblings [23] . Hence , the normal differentiation of any ‘escaper’ iridophores , even in shdty82 homozygotes , strongly argues against a later role for Ltk in iridophore differentiation . Maintenance of ltk expression in premigratory NCCs is dependent upon Ltk function since in shd , but not sox10 mutants , ltk expression was absent from iridophore precursors prior to their death . We saw no evidence for fate-switching between different pigment cell fates and although shd mutants show elevated NCC death , this is clearly a secondary effect since it occurs well after ( c . 35–50 hpf ) the initial fate specification phenotype . These two features are strongly reminiscent of our previous characterisation of the sox10 mutant phenotype and suggest important parallels between the cell-biological defects in sox10 and shd/ltk mutants . Our ltk expression data is most immediately interpreted as suggesting that ltk is expressed from a very early point in iridophore specification , but is restricted to the iridophore lineage . However , the very early low level expression pattern in rows of premigratory cells , could also be interpreted as suggesting a transient early phase of expression in multipotent neural crest cells , before being upregulated and maintained in those cells that become specified to an iridophore fate . Our sox10 mutant data , whilst initially unexpected , supports the hypothesis that ltk is expressed initially in a subset of multipotent neural crest cells . Zebrafish sox10 mutants have reductions in non-ectomesenchymal NC derivatives , including melanophores and iridophores [27] . Detailed investigation of melanocyte defects in sox10 mutants showed definitively that Sox10 regulation of mitfa transcription , and thus melanocyte fate specification , was the primary defect [9] . A direct test of whether a similar mechanism applies to iridophores will require the identification of ‘master regulator’ transcription factor ( s ) for this cell-type . However , our analysis of sensory and enteric neuron defects in sox10 mutant zebrafish , together with studies of glia , sympathetic neurons and melanocytes in Sox10 mutant mice identify a common theme of failure of fate specification from multipotent precursors resulting from impaired transcription of ‘master regulator’ transcription factors [8] , [9] , [10] , [11] , [12] , [13] , [24] , [25] , [37] , [38] , [39] , [40] . In this context , under the assumption that ltk expression was the earliest known marker for the iridoblast lineage and that ltk expression simply marked iridophore specification , we predicted that sox10 mutants would not show ltk expression at all . Consequently , we were initially surprised to find a prominent accumulation of ltk-expressing neural crest cells . The overabundance of these ltk-expressing cells in sox10 mutants , tightly clustered in a premigratory position , suggested that these cells were early NCCs . However , they did not express the early NCC markers , snail2 and foxd3 , and were found only more anteriorly ( i . e . in developmentally older cells ) , suggesting that they were a distinct population of partially-restricted progenitors , consistent with the progressive fate restriction model . These cells do express sox10 , a gene required for maintenance of multipotency in at least neural precursors [39] , [49] . We propose that in WT embryos ltk is expressed transiently in multipotent NCCs , but that in sox10 mutants , where fate specification is prevented , these cells remain trapped in this partially restricted progenitor state ( Figure 8 ) . The alternative interpretation of our data , that ltk expression is restricted to , and indeed defines , specified iridoblasts would lead to the conclusion that iridoblast fate specification occurs in sox10 mutants , but that further development fails . Whilst plausible , we do not favour this model because our data is fully consistent with the specification model shown to have general applicability to all other fates examined to date . Furthermore , we have previously shown that crestin , a general and early marker of differentiated lineages [50] , is not expressed in these cells in sox10 mutants [25] . Since in sox10 mutants all pigment cell precursors fail to migrate , whereas neural precursors migrate normally [26] , we speculate that the ltk-expressing cells may be multipotent pigment cell progenitors [1] , [51] . A definitive test of our proposals will require development of zebrafish neural crest cell culture or of tools to definitively fate map ltk-expressing cells . Our data identifies a novel RTK pathway mediating NC development . RTKs have diverse roles in development , including in fate specification . Indeed , Drosophila Alk functions in specification of visceral muscle pioneers [43] , [44] . However , in pigment cell development RTK function has been shown to be important for proliferation , survival and migration , but not fate specification [36] , [48] , [52]–[56] . For example , zebrafish kit mutants show a partial reduction in melanoblast numbers from approximately 36 hpf [35] , [48] and fms mutants show a failure of xanthoblast migration approximately 28 hpf [36] . In contrast , shdty82 mutants show a phenotype that is both earlier and much more severe than in these other RTK mutants . Indeed , our data reveal that iridophore fate specification occurs very early , with timing equivalent to that of melanocyte fate specification [5] , despite the later differentiation of iridophores ( c . 42 hpf for iridophores , c . 25 hpf for melanophores ) . We show that in zebrafish Ltk is crucial for specification of a particular pigment cell type , the iridophore , from NCCs . Our data contributes to understanding how pigment cell fate specification from these multipotent cells occurs . The challenge for the future will be to identify the genetic interactions between ltk , sox10 and other genes determining pigment cell fate choice . The ubiquitous nature of iridophores in fish , amphibians and reptiles suggests that Ltk function in NC is likely to be widespread . At least some birds , including doves , show iridophore-like cells in their iris [57] , but their embryological origin is unclear , so examination of Ltk expression in appropriate avian embryos will be revealing . Iridophores have been lost in mammals , yet Ltk has been evolutionarily conserved . Strong shd mutant alleles are homozygous lethal [23] , but this lethality cannot be attributed to the iridophore phenotype , and perhaps results from conserved functions in brain [20] . Further characterisation of defects in shd mutants will allow identification of any conserved roles . Finally , our data suggest simple , visual in vivo screens for LTK inhibitors which may be of utility considering the growing links of these RTKs to autoimmune disease [22] , [58] .
shd ty82 , shd ty9 and shdty70 have been described [23] . shdj9s1 was identified as a spontaneous mutation in AB stocks , and shdj9e1 in an early pressure screen for adult phenotypes [59] . shdj9e2 was identified in a non-complementation screen with shdj9s1 . WIK11 WT was used to generate the reference mapping crosses . WT cDNA was amplified from the AB line . All studies conformed to local and UK national ethical guidelines . Embryos were imaged on an Eclipse E800 ( Nikon ) using a U-III or DS-U1 camera ( Nikon ) or an LSM Meta confocal ( Zeiss ) microscope . Embryos were processed as previously [60] . Embryos for antibody staining were processed as previously described [61] . Antibodies used: mouse anti-Hu C/D ( Molecular Probes ) ; rabbit anti-phospho-Histone H3 ( Upstate Biotechnology , Cat#06-570 ) ; mouse anti-GFP , goat anti-mouse Alexa488 and anti-rabbit Alexa546 ( Molecular Probes ) . TUNEL assays were carried out using an ApopTag® Peroxidase In Situ Apoptosis Detection Kit ( Chemicon , Cat#:S7100 ) according to manufacturer's instructions . Assessment of cell-autonomy was performed as described before [27] . Labelled cells were detected by rhodamine fluorescence in the live embryo and by peroxidase detection of biotinylated tracer in embryos fixed after photographing the iridophore pattern . Heterozygous F1 fish from the mapping cross were incrossed and separate pools of F2 homozygous shd mutants and their WT siblings were used for simple sequence length polymorphism analysis [62] . Linkages from the pools were confirmed and refined by genotyping 1000 individual mutant embryos . The PAC 706 genomic library ( RZPD ) was screened with the marker z10985 by PCR; three positive PAC clones , BUSMP706P14181Q2 ( PAC1 ) , BUSMP706N10265Q2 ( PAC2 ) and BUSMP706O16107Q2 ( PAC3 ) , were provided by RZPD . 450 pg of PAC DNA was injected in 2-cell stage embryos from shdty82/+ carrier cross . Morpholino antisense oligonucleotides ( Gene Tools ) were injected into fertilized WT eggs at the one-cell stage , at concentrations up to 35 ng per embryo and incubated at 28 . 5°C until 72 hpf . ltk morpholino sequence , 5′-agtttgtcgagtaatataatccat-3′; mismatch control , 5′-actttctccagtaatattatgcatg-3′ . Three PACs containing z10985 were sequenced by the Danio rerio Sequencing Project at the Wellcome Trust Sanger Institute ( Sequences of PACs 1–3 have accession numbers BUSM1-181P14 , BUSM1-265N10 and BUSM1-107O16 respectively and can be accessed from http://www . sanger . ac . uk/Projects/D_rerio/ ) . Predicted genes were identified using Nucleotide Identify X ( NIX ) software ( UK Human Genome Mapping Project Resource Centre ) . WT cDNAs were isolated using SMART RACE PCR ( Clontech ) and long PCR using the Herculase enhanced DNA polymerase ( Stratagene ) , cloned in Zero Blunt TOPO ( Invitrogen ) and sequenced commercially ( Oswel ) . ltk cDNAs were amplified from shdty82 and shdj9e2 embryos and sequenced directly . RFLP analysis of shdty82 and siblings was performed on PCR fragments amplified from genomic DNA prepared from single embryos . Sequences were aligned using ClustalW multiple sequence alignment software . WT and mutant cDNA sequences have been deposited in Genbank under accession numbers 1051419 and 1057253 respectively . Individual embryos , or heads of individual embryos if embryos were to be subsequently processed for in situ hybridisation , were placed in 96-well plates , washed three times with PBS , digested at 55°C in 2 mg/ml Proteinase K ( Roche ) for 4 hours , then heated at 95°C for 10 min to inactivate enzyme . Diagnostic PCR was performed using forward ( 5′-CTAACTCAAAGCAGTTTCGT-3′ ) and reverse ( 5′-GTAACGTCATGAGCAGATAA-3′ ) primers and the following PCR programme: 3 mins at 94°C; 35 cycles of 30 sec at 94°C , 30 sec touchdown from 55–47°C , 30 s at 72°C; 10 mins at 72°C . PCR products were then cut with NheI and run on agarose gel to reveal diagnostic bands: 420 bp ( WT ) , 360 and 130 bp ( shdty82 ) . The cloned sequences were compared with homologous sequences in a phylogenetic analysis using a Bayesian method [63] implemented in MrBayes ( v3 . 1 . 2 ) and a maximum likelihood analysis implemented in TREEFINDER [64] applied to each of four alignments . See Table S2 and Figure S3 . Diagnosis of protein subdomains utilised interpro ( http://www . ebi . ac . uk/interpro/ ) .
|
Stem and other multipotent cells generate diverse cell-types , but our understanding of how they make these decisions , which is important for their therapeutic use , is incomplete . Neural crest cells are an important class of multipotent cells and generate multiple stem cell types . We have looked at how pigment cells are made from the neural crest in the zebrafish . The silver shine familiar in so many fish is due to specialised mirror-like pigment cells , called iridophores . We show that these cells are missing in zebrafish shady mutants . We identify the shady gene as encoding a cell signalling receptor , leukocyte tyrosine kinase ( Ltk ) , that has recently been associated with human auto-immune disease . We show that in zebrafish this gene is most likely required to make iridophores from neural crest cells . Thus , we identify a novel pathway required for diversification of these multipotent cells . Our work defines the first role for Ltk in a vertebrate . It provides a mutant resource that will allow us to discover the full breadth of roles for this important gene . Furthermore , the loss of iridophores forms a simple visual screen for inhibition of LTK function and might well have implications in drug discovery .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology/cell",
"differentiation",
"developmental",
"biology/developmental",
"molecular",
"mechanisms",
"developmental",
"biology/stem",
"cells"
] |
2008
|
Leukocyte Tyrosine Kinase Functions in Pigment Cell Development
|
Sensitization to Anisakis spp . can produce allergic reactions after eating raw or undercooked parasitized fish . Specific IgE is detected long after the onset of symptoms , but the changes in specific IgE levels over a long follow-up period are unknown; furthermore , the influence of Anisakis spp . allergen exposure through consumption of fishery products is also unknown . To analyse the changes in IgE sensitization to Anisakis spp . allergens over several years of follow-up and the influence of the consumption of fishery products in IgE sensitization . Total IgE , Anisakis spp . -specific IgE , anti-Ani s 1 and anti-Ani s 4 IgE were repeatedly measured over a median follow-up duration of 49 months in 17 sensitized patients . Anisakis spp . -specific IgE was detected in 16/17 patients throughout the follow-up period . The comparison between baseline and last visit measurements showed significant decreases in both total IgE and specific IgE . The specific IgE values had an exponential or polynomial decay trend in 13/17 patients . In 4/17 patients , an increase in specific IgE level with the introduction of fish to the diet was observed . Three patients reported symptoms after eating aquaculture or previously frozen fish , and in two of those patients , symptom presentation was coincident with an increase in specific IgE level . IgE sensitization to Anisakis spp . allergens lasts for many years since specific IgE was detectable in some patients after more than 8 years from the allergic episode . Specific IgE monitoring showed that specific IgE titres increase in some allergic patients and that allergen contamination of fishery products can account for the observed increase in Anisakis spp . -specific IgE level . Following sensitization to Anisakis spp . allergens , the absence of additional exposure to those allergens does not result in the loss of IgE sensitization . Exposure to Anisakis spp . allergens in fishery products can increase the specific IgE level in some sensitized patients .
The nematode Anisakis spp . is a parasite of marine mammals that can parasitize humans when a raw or undercooked fish containing live Anisakis spp . L3 is consumed . Ingestion of L3 causes an acute and self-limiting infection that can manifest with abdominal pain , nausea , vomiting or diarrhoea . Infection causes a strong polyclonal humoral immune response , and IgM , IgA , and IgG antibodies are detected after one month of infection [1] . In some patients , an IgE-mediated immune response is also triggered , and in those patients , allergic symptoms , such as urticaria , angioedema and anaphylaxis can develop after sensitization and re-exposure to the allergens of this parasite . The rise in specific IgE is usually accompanied with an increase in total IgE in the first month after the presentation of allergic symptoms , and serial serological analysis of both specific and total IgE values have been proven useful in the diagnosis of gastro-allergic anisakiasis [2] . To avoid the appearance of symptoms , sensitized patients are advised to consume frozen or heat-treated fishery products because these treatments kill larvae to prevent new parasitism [3–5] . Several groups have investigated the kinetics of specific antibody production in experimental animal models [review in 6] , but the results of those studies may not be applicable to the human immune responses to this parasite . Studies of the changes over time of the level of specific IgE to Anisakis spp . in sensitized patients have shown the persistence of IgE sensitization up to 38 months after the onset of symptoms [1 , 2 , 4 , 7] . However , those studies did not report variations in the specific IgE levels at different follow-up time points . The aim of this study was to analyse the changes in Anisakis spp . -specific IgE levels through repeated measures during a longer follow-up period than previously reported and to compare IgE sensitization between patients whose diets did not include fishery products and subjects who regularly consumed fishery products .
To analyse the kinetics of the IgE response , Anisakis spp . -allergic patients with at least 30 months of follow-up after symptom presentation were selected for this study . A total of 17 patients ( six males ) with a median age of 53 years ( IQR = 45–57 years ) were diagnosed as being allergic to Anisakis spp . because they reported allergic ( urticaria , angioedema or anaphylaxis ) and/or gastrointestinal ( vomiting , diarrhoea , or abdominal pain ) symptoms within 24 h after eating raw or undercooked fish or seafood . One patient reported symptoms after eating cooked fish ( scorpion fish cake ) . Five patients had grass pollen allergy and one of them had dog dander allergy . The data collected at the first visit are shown in Table 1 . Allergy was confirmed by a positive prick test and/or detection of specific IgE to Anisakis spp . and undetectable levels of IgE to shrimp , Ascaris lumbricoides , fish and mites . To assess new sensitization to these allergens , specific IgE levels were quantified at the last visit , and they remained undetectable in all patients . Measurements of the levels of total and specific IgE to Anisakis spp . and clinical evaluations were performed during successive visits . All patients were advised at the first visit to avoid consumption of raw and undercooked fish and to eat farmed fish and deep-frozen fishery products; however , patients with levels of specific IgE to Anisakis spp . higher than 100 kU/L were initially instructed to consume a fish-free diet for six months . This study was approved by the Ethics Committee of the Hospital Carlos III ( Madrid , Spain ) , and all included subjects were asked to sign an informed consent form . The serum total and specific IgE measurements were performed with a Phadia 250 instrument ( Thermo Fischer Scientific , Phadia , Madrid , Spain ) according to the manufacturer’s instructions . The detection range for total IgE was 2–5000 kU/L . Regarding positivity for specific IgE antibodies , values >0 . 7 kU/L [8] were considered positive for IgE to Anisakis spp . , and values >0 . 35 kU/L were considered positive for IgE to the other allergens . Live Anisakis spp . larvae in the third stage of the life cycle ( L3 ) were obtained from parasitized hake ( Merluccius merluccius ) at local markets in Madrid , Spain . L3 were extracted from fish tissue , washed in PBS and immediately frozen at -20°C until use . Then , L3 were ground in a Potter-ELV homogenizer and sonicated at 18 w for 5 s . Protein extracts were obtained after centrifugation at 16 , 000 g and 4°C for 10 min . Recombinant ( r ) Ani s 4 and rAni s 1 were obtained as previously reported [9 , 10] . In addition to the total and specific IgE measurements , IgE immunoblotting was performed with the parasite crude extract , recombinant ( r ) Ani s 1 and rAni s 4 . Proteins extracted from L3 ( 15 μg ) , rAni s 1 ( 3 μg ) and r Ani s 4 ( 3 μg ) were subjected to electrophoresis at 120 V on a 4%-20% Tris-glycine gel ( Bio-Rad , Hercules , CA , USA ) . Thereafter , proteins were transferred to nitrocellulose membranes by applying a constant current of 1 . 3 A for 7 min in a Trans-Blot Turbo Instrument ( Bio-Rad ) . Membranes were blocked with PBS , 0 . 05% Tween 20 and 1% BSA for 1 h at room temperature and then incubated with 10 mL of the sera of Anisakis spp . -allergic patients ( 1/20 ) overnight . After washing with PBS , the membranes were incubated for 2 h with 10 mL of a 1:1000 dilution of a monoclonal anti-IgE antiserum ( 1 mg/mL; Ingenasa , Madrid , Spain ) . After additional washes , the membranes were incubated with 10 mL of a 1:20 , 000 dilution of an alkaline phosphatase–labelled goat anti-mouse antiserum ( Sigma-Aldrich , St . Louis , MO , USA ) . Finally , the membranes were washed and incubated with the BCIP-NBT ( Sigma-Aldrich ) substrate for 30 min . Statistical analyses were performed using SPSS 20 . 0 software ( IBM Corporation , NY , USA ) . Quantitative variables are described as medians and interquartile ranges ( IQR ) . The Mann-Whitney U test was used to compare the values of total IgE and specific IgE quantified at the first and last visits . A paired-samples comparison between the baseline and last visit values was performed using the Wilcoxon signed-rank test . Regression was used to analyse the trends in the changes in specific IgE values . Changes in specific IgE values over time were estimated with linear and non-linear regression models , and the best fit was selected . A p–value of <0 . 05 was considered statistically significant . For the statistical analysis , all values of Anisakis spp . -specific IgE >100 kU/L were assigned a value of 101 kU/L [11] .
At first visit , three patients reported gastrointestinal , allergic ( n = 6 ) or allergic and gastrointestinal symptoms ( n = 8 ) after eating raw or undercooked fish , except P14 who reported urticaria after eating cooked fish ( scorpion fish cake ) . Raw anchovies in vinegar sauce were the most frequently consumed fish related to the onset of symptoms ( Table 1 ) . The median time elapsed from the allergic episode to the first visit was one month ( IQR = 0 . 85–7 months ) . The median follow-up duration was 49 months ( IQR = 38–83 months ) . P5 , P7 , P9 and P13 reported previous allergic episodes associated with raw fish consumption . At the first visit , using IgE immunoblotting , in addition to Anisakis spp . -specific IgE , specific IgE to rAni s 1 was detected in all patients , and specific IgE to rAni s 4 was detected in six patients ( Table 1 ) . Although patients were advised to eat aquaculture and previously frozen fish , five patients did not include fish or seafood in their diet during the follow-up period ( P2 , P4 , P7 , P11 and P16 ) , and they were considered to have not been re-exposed to Anisakis spp . antigens or allergens . Positive specific IgE to Anisakis spp . values ( >0 . 7 kU/L ) were detected in all patients throughout the follow-up period , except for P14 , who had a specific IgE to Anisakis spp . value of 0 . 6 kU/L at the end of a 112-month follow-up period ( Table 1 ) . Baseline total IgE ( 557 kU/L , IQR = 242–1648 kU/L ) and Anisakis spp . -specific IgE values ( 79 kU/L , IQR = 20–101 kU/L ) were significantly higher than the final total IgE value ( 184 kU/L , IQR = 57–410 kU/L; p <0 . 01 ) and the Anisakis spp . -specific IgE titre ( 13 kU/L , IQR = 3–21 kU/L , p< 0 . 01 ) ( Table 1 ) . The median value of the specific IgE decrease was 76% over the follow-up period ( IQR = 60%-88% ) . The paired-samples comparison between the measurements at baseline and the last visit showed significant decreases in both total IgE ( p < 0 . 01 ) and Anisakis spp . -specific IgE titres ( p < 0 . 01 ) . No significant differences were found between the patients who were and were not re-exposed to Anisakis spp . in the decrease in the Anisakis spp . -specific IgE level over the follow-up period ( 76% , IQR = 51%-84% and 88% , IQR = 66%-92% , respectively; p = 0 . 28 ) or in the follow-up duration ( 53 months , IQR = 37–96 and 45 months , IQR = 40–60 , respectively; p = 0 . 72 ) . Regression was used to analyse the trend in the changes in Anisakis spp . -specific IgE values . The Anisakis spp . -specific IgE values underwent an exponential or polynomial decay trend in 13/17 patients , including the patients who had not included fish in their diet during the follow-up period ( Fig 1 and S1 Fig ) . However , the changes in the Anisakis spp . -specific IgE levels from baseline to last visit in four patients ( P1 , P3 , P6 and P15 ) did not fit a regression model ( Fig 2 ) . The Anisakis spp . -specific IgE level in P1 decreased to 1 kU/L at 10 months , and then increased and remained at > 12 kU/L throughout the remainder of the follow-up period . The Anisakis spp . -specific IgE level in P3 initially showed a 70% decrease ( from 50 kU/L to 14 kU/L ) and , then , increased up to 37 kU/L at one year later . The changes in the Anisakis spp . -specific IgE level in P6 were similar to those in P1 , with a decrease of specific IgE titre until reaching 3 kU/L at 12 months and an obvious increase at the next visit ( 50 kU/L ) . P15 showed a moderate level of Anisakis spp . -specific IgE ( 10 kU/L ) at the first visit; it decreased at one month ( 6 kU/L ) and increased at the following visit . The Anisakis spp . -specific IgE increases in these four patients were coincident with the introduction of aquaculture and frozen fish into their diet ( Fig 2 ) . The paired samples comparisons in these patients revealed no significant differences between baseline total IgE ( 927 kU/L , IQR = 367–1604 kU/L ) and last visit total IgE ( 457 kU/L , IQR = 397–573 kU/L; p = 0 . 27 ) or between baseline Anisakis spp . -specific IgE ( 66 kU/L , IQR = 15->100 kU/L ) and last visit Anisakis spp . -specific IgE ( 23 kU/L , IQR = 10–45 kU/L; p = 0 . 07 ) . When the total and specific IgE titres and the duration of follow-up for the patients whose data fit a regression model were compared to those of these four patients , only total IgE at last visit was found to be higher in the latter group ( 184 kU/L , IQR = 57–410 kU/L vs . 457 kU/L , IQR = 397–573 kU/L , p < 0 . 01 ) . Three patients reported symptoms during the follow-up period . P3 experienced localized acute urticaria in the hands after eating cod that had been frozen for 72 h at home ( Fig 2 ) . P5 reported diarrhoea after eating aquaculture sea bass ( Fig 1 ) . In these two patients , the appearance of symptoms was coincident with an increase in Anisakis spp . -specific IgE level . P3 and P5 did not report new episodes in the subsequent visits . P13 showed generalized acute urticaria after eating commercial frozen hake ( Fig 1 ) . Unfortunately , this episode occurred at the end of the follow-up period , and we could not assess if the level of IgE against Anisakis spp . had varied . Other common causes of acute urticaria were discarded . To determine if the changes in specific IgE over time are related to changes in the recognition pattern of allergens , IgE immunoblotting was performed at different time points . As expected , baseline IgE immunoblotting showed different patterns for the parasite crude extracts [12] . rAni s 1 was detected by all patients , and rAni s 4 was detected by six patients ( Figs 1 and 2 and S1 Fig ) . The intensities of the bands corresponding to rAni s 1 paralleled the Anisakis spp . -specific IgE levels . During the follow-up period , the rAni s 4 bands disappeared prior to the rAni s 1 bands , suggesting that Ani s 4 is an early marker of Anisakis spp . infection . On the other hand , the increase in Anisakis spp . -specific IgE over time observed in some patients was associated with an increase in the intensity of some proteins in the parasite crude extract and rAni s 1 ( Fig 2 ) .
Our study analysed the changes in the values of total IgE and specific IgE to Anisakis spp . over time using a longer follow-up period ( 31–118 months ) than previous studies ( 6–38 months ) [1 , 2 , 4 , 7] . Repeated measures were performed throughout the study period and included detection of Ani s 1 and Ani s 4 . Ani s 1 is a major and heat stable allergen [12] and Ani s 4 is a pepsin and heat-resistant allergen , and its clinical relevance has been verified because it is associated with anaphylaxis [9] . Exposure to the live fish-borne parasite Anisakis spp . L3 can produce an acute and self-limiting infection in humans with allergic and gastrointestinal symptoms . To avoid the appearance of symptoms , patients are advised to consume frozen or heat-treated fishery products because these treatments kill larvae and thus prevent new parasitism [3–5] . An alternative is to consume aquaculture fish because the risk of exposure to Anisakis spp . larvae in farmed fish has been shown to be minimal , because no Anisakis spp . larvae have been found in their viscera or flesh [5 , 13] . Twelve patients regularly consumed aquaculture and frozen fish during the study period , and five patients decided to stop consuming fish throughout the follow-up period . The decrease in the Anisakis spp . -specific IgE levels in these patients , who were considered to have not been re-exposed to parasite material , was higher than those found previously in patients on a fish-free diet ( 57% ) with a follow-up period of 13 months , and this discrepancy was probably due to our longer follow-up period [4] . On the other hand , we did not observe significant differences in the total or Anisakis spp . -specific IgE values between the patients who did and did not consume fish during the follow-up period [4]; thus , the rate of Anisakis spp . -specific IgE decay does not seem to be influenced by the consumption of previously frozen fish . Our results show that IgE sensitization to Anisakis spp . allergens persists over several years since specific IgE was detectable in some patients after more than 8 years from the allergic episode . Similar results were observed in a nine-year follow–up study of adult subjects sensitized to food allergens [14] . In our study , specific IgE to Ani s 1 was detectable up to 118 months from the onset of symptoms , and this result agrees with that obtained in a study with 6–38 months of follow-up [7] and suggests that Ani s 1 can be detected in both recent and old Anisakis spp . infection cases . Therefore , the persistent IgE sensitization observed in Anisakis spp . allergic patients who were not exposed to parasite allergens for several years indicates that Anisakis spp . allergens induce long-lived IgE responses . The analysis of the changes over time in specific IgE level in the patients that were not re-exposed to Anisakis spp . allergens and in some patients who regularly consume aquaculture and frozen fish shows that the decrease in specific IgE titres fit a non-linear regression model . However , we observed that the Anisakis spp . -specific IgE level in four patients ( P1 , P3 , P6 and P15 ) increased at some time points during follow–up . This increase in Anisakis spp . -specific IgE could be due to sensitization to other allergen sources that have been reported to cross-react with Anisakis spp . allergens , i . e . , other parasite nematodes , mites and crustaceans [15–17] . However , according to the available patient data , this explanation does not seem to be valid because our patients were not sensitized to allergens that cross-react with Anisakis spp . allergens at baseline , and no new onset hypersensitivity to those allergens was found during the follow-up period . In addition , the increase in Anisakis spp . -specific IgE was parallel to the increase in Ani s 1 detected by immunoblotting , which supports the hypothesis that the observed changes in Anisakis spp . -specific IgE were actually due to exposure to Anisakis spp . allergens . Another explanation for this observation is re-infection with live L3 . However , we do not believe this is the case because the patients reported symptoms after eating frozen or aquaculture fish , and it is unlikely that an acute parasitism episode went unnoticed by these patients , as they had experienced one previously . On the other hand , previously unrecognized allergens have been detected in the course of gastro-allergic anisakiasis [1] , which was not observed in our patients according to the results of IgE immunoblotting . Accordingly , it has been shown that adult patients allergic to aeroallergens did not acquire sensitization to new allergens , but they exhibited a pre-established profile of allergens after antigen exposure . The levels of allergen-specific IgE in previously sensitized allergic patients decreased in the absence of allergen exposure and increased upon allergen exposure [18] . Because the raise in specific IgE levels is coincident with the introduction of fishery products to the diet , a more plausible explanation for the changes in IgE sensitisation is that exposure to allergens present in those fishery products is involved in the increase in the Anisakis spp . -specific IgE level . Some Anisakis spp . antigens have been shown to be stable and to maintain their capacity to bind IgE after different freezing and heat treatments [19 , 20] . Furthermore , Anisakis spp . antigens have been detected in farmed salmon and processed fish products [21] , which could contribute to the variations in the trends of IgE sensitization observed in this study . The clinical significance of these results is difficult to determine because very few patients ( n = 3 ) reported symptoms after the initial acute parasitism episode in our study . It has been proposed that the high rate of Anisakis spp . parasitism of fish in our region would result in frequent contact with parasite material , which would cause symptoms in sensitized subjects exposed to it when previously frozen or heat-treated fish is included in their diet [22] . However , our results suggest that the exposure to Anisakis spp . antigens present in fishery products may contribute to the persistence and even to the increase in Anisakis spp . -specific IgE level . The increase in the Anisakis spp . -specific IgE level associated with the appearance of symptoms indicates that exposure to Anisakis spp . material could be involved in the onset of symptoms in some Anisakis spp . allergic patients [12] . However , more studies are required to evaluate the clinical relevance of the re-exposure to Anisakis spp . in sensitized patients . In conclusion , we have shown that IgE sensitization to Anisakis spp . allergens can last more than 8 years . Specific IgE monitoring showed that specific IgE titres increase in some allergic patients and that allergen contamination of fishery products can account for the observed increase in Anisakis spp . -specific IgE level .
|
The nematode Anisakis spp . is a parasite of marine mammals that can parasitize humans when a raw or undercooked fish containing live Anisakis spp . larvae is consumed . As a result , gastrointestinal and/or allergic symptoms are reported . The allergic reaction can be diagnosed by quantifying the serum specific IgE against some parasite allergens . To avoid the appearance of symptoms , sensitized patients are advised to consume frozen or heat-treated fishery products because these treatments kill larvae to prevent new parasitism and allergic reactions . We have analysed the changes in Anisakis spp . -specific IgE levels through repeated measures during a long follow-up period of sensitized individuals whose diets did not include fishery products and subjects who regularly consumed fishery products . We have found that IgE sensitization to Anisakis spp . allergens persists over several years ( more than 8 years in some cases ) even in subjects who did not include fishery products into their diets . On the other hand , specific IgE titres increase in some allergic patients during the follow-up period and this increase was coincident with the introduction of previously frozen fish into their diet . These findings indicate that Anisakis spp . allergens induce long-lived IgE responses .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"molecular",
"probe",
"techniques",
"immunology",
"immunoblotting",
"vertebrates",
"diet",
"marine",
"biology",
"animals",
"marine",
"fish",
"physiological",
"processes",
"clinical",
"medicine",
"nutrition",
"fisheries",
"eating",
"aquaculture",
"allergens",
"molecular",
"biology",
"techniques",
"anisakis",
"research",
"and",
"analysis",
"methods",
"fishes",
"marine",
"and",
"aquatic",
"sciences",
"molecular",
"biology",
"agriculture",
"allergies",
"clinical",
"immunology",
"earth",
"sciences",
"physiology",
"nematoda",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2016
|
Changes over Time in IgE Sensitization to Allergens of the Fish Parasite Anisakis spp.
|
The secondary metabolome provides pathogenic fungi with a plethoric and versatile panel of molecules that can be deployed during host ingress . While powerful genetic and analytical chemistry methods have been developed to identify fungal secondary metabolites ( SMs ) , discovering the biological activity of SMs remains an elusive yet critical task . Here , we describe a process for identifying the immunosuppressive properties of Aspergillus SMs developed by coupling a cost-effective microfluidic neutrophil chemotaxis assay with an in vivo zebrafish assay . The microfluidic platform allows the identification of metabolites inhibiting neutrophil recruitment with as little as several nano-grams of compound in microliters of fluid . The zebrafish assay demonstrates a simple and accessible approach for performing in vivo studies without requiring any manipulation of the fish . Using this methodology we identify the immunosuppressive properties of a fungal SM , endocrocin . We find that endocrocin is localized in Aspergillus fumigatus spores and its biosynthesis is temperature-dependent . Finally , using the Drosophila toll deficient model , we find that deletion of encA , encoding the polyketide synthase required for endocrocin production , yields a less pathogenic strain of A . fumigatus when spores are harvested from endocrocin permissive but not when harvested from endocrocin restrictive conditions . The tools developed here will open new “function-omic” avenues downstream of the metabolomics , identification , and purification phases .
The secondary metabolome provides filamentous fungi with a biologically active panel of molecules , deployed in the presence of competing/host organisms or specific microenvironmental factors , and increasingly found to afford both physical and competitive fitness to the producing fungus [1] . Although the study of fungal secondary metabolism has reached the ‘omics era , with the development of tools for efficient genetic exploration [2] and the improvement of HPLC and LC-MS methods [3] , a significant challenge remains in identifying the biological activity of the purified compounds . The minute quantity of metabolites ( nano- to micrograms ) collected from the latter methods and the large number of isolated compounds ( hundreds to thousands ) are important limiting factors in this endeavor . Thus , as the methods for identifying SM gene clusters and the compounds they produce are becoming well established [4] , there is an increasing need for improved assays , compatible with the fungal metabolomics process , that can reveal the biological activity of metabolites produced and break a bottleneck in scientific advancement . Aspergillus spp . SMs are of particular interest in medical research as the genus is genetically accessible , produces a plethora of bioactive compounds [5] , and contains several opportunistic pathogenic species including A . fumigatus and A . nidulans whose SMs are assessed in this study [6] , [7] . Though it is likely that a number of factors together contribute in making these species effective pathogens , SMs play an important role in the virulence of Aspergillus-related diseases as direct toxins and modulators of the immune response [8] , [9] . As the innate immune response is the primary line of defense against fungal spores in the lung , inhibition of essential functions of these cells may confer to the fungi an ability to evade immune clearance , and increase its pathogenicity . These findings highlight the necessity of mapping the interactome between fungi and host organisms to establish the pathomechanism of fungal diseases as well as to bioprospect . Leading to this current study are the series of works showing LaeA , a global regulator of secondary metabolism to be a virulence factor not only in pathogenic Aspergilli [10]–[13] but in all filamentous pathogenic fungi assessed to date [12] , [14] , [15] . Examination of the laeA mutant in A . fumigatus implicated unidentified SMs in development of invasive aspergillosis [10] , [11] . As several studies have shown A . fumigatus culture filtrates to inhibit neutrophil chemotaxis [16]–[18] , we considered it possible that LaeA-regulated SMs could be chemotaxis inhibitors . Complicating this hypothesis , however , is the fact that LaeA regulates dozens of SM clusters , all of which can produce multiple derivatives from the same biosynthetic pathway , whose purification results in small available quantities [19] . Traditional in vitro neutrophil migration models , often performed in well-plates , do not allow a good level of control over the migration microenvironment , do not allow imaging of the cells during the migration process , and require large amounts of purified compound ( micrograms to grams in hundreds of microliters to milliliters ) . Advances in microscaled assays have demonstrated enabling characteristics , with the ability to use microliters or less of reagents , develop high-throughput applications , and design assays with more control over the micro-environment [20] . The development of open systems , that interface with existing fluid handling equipment , contributed in making microscaled assays more accessible and better suited for screening libraries of individual compounds [21] . Further , these approaches enable the development of functional cell-based assays , such as arrayed leukocyte recruitment assays [22] . However , these have not been applied for identifying SMs modulating the fungal-immune interaction , nor for screening of fungal SMs . Appropriate in vivo models are also required to support in vitro progress . Current models such as Galleria [23] or Drosophila [24] , are accessible and inexpensive but do not fully recapitulate the vertebrate immune system . More relevant models such as the murine model are logistically challenging , require excessive amounts of purified SMs , and are still fraught with deficiencies in visualization of innate cell response [25] . Recently , the development of the zebrafish embryo model has proven to be a vertebrate model well suited for leukocyte studies as it is readily accessible , small , and transparent and has been established as a model system for infectious diseases including fungi [26] . Here , we demonstrate a two-tiered screening approach for identifying the immunosuppressive and neutrophil recruitment inhibitory activity of LaeA-regulated SMs , capitalizing on advances of microscale in vitro systems and an original zebrafish model . An arrayed microfluidic in vitro neutrophil recruitment platform was developed , compatible with manual and automated pipettes , allowing for rapid assessment of the neutrophil recruitment inhibition properties of purified Aspergillus SMs . Passive open microfluidic methods were employed for creating arrayed gradient-generation devices operable in typical biological laboratory settings and minimizing reagent use . Bioactive metabolites identified from this platform were then assessed in an in vivo zebrafish recruitment assay . The zebrafish assay is enabling as it significantly reduces the quantity of compounds required and provides a quickly assessed window into innate immune response . Using this approach , we report the identification of the neutrophil recruitment inhibition activity of endocrocin , an A . fumigatus SM , and further characterized its localization in spores of the growing fungus .
In order to systematically test the large number of compounds provided by liquid chromatography methods , a microscale platform for assessing neutrophil chemotaxis properties was designed . The use of tubeless microfluidic-interfacing methods minimizes dead volumes and static gradient-generation methods further minimizes volume requirements [22] , [27] . Thus , neutrophil recruitment can be assessed with as little as 3 µL of purified compound , using only a simple micropipette . In this approach , the gradient is generated in a reproducible way by leveraging a flow bypassing method based on ensuring that no undesired flow passes through the gradient channel , rather a second flow path of significantly lesser fluidic resistance diverts the bulk of the flow ( Figure 1A , B ) . The reliability of the fluidic handling mechanism used , and the simplicity of the design , enables rapid loading protocols and large arrays of microdevices . In order to achieve the highest throughput possible , batch-processing algorithms were developed , enabling the quantification of neutrophil migration properties from an endpoint phase-contrast image of the migrating neutrophils . Together , these advances allow the processing by hand of up to 300 migration data-points per experimental run in an embodiment that can be readily interfaced using handheld pipettes and automated liquid handlers . The static gradient-generation device was verified using fluorescent dyes and food colorant ( Figure 1B ) , demonstrating that the gradient establishes in minutes as assessed by fluorescent imaging of AlexaFluor488 ( Figure 1C ) . The microfluidic gradient , once established , was able to maintain steady for several hours , allowing the reliable measure of neutrophil ( or other leukocyte ) migration over the course of a typical neutrophil experiment ( Figure 1C ) . The microfluidic channel can be entirely prepared in 3 simple pipetting steps , which require seconds , and can be interface-able with electronic and robotic liquid handlers ( Figure 1D ) . Using the flow generation method described , as well as the static gradient-generation methods , we demonstrate the ability to create large arrays of gradient devices which can be prepared in large batches allowing the operation of several hundred per day ( Figure 1E ) . As these contain very little dead volumes ( being devoid of tubes and actuation equipment ) a single blood draw of 10–20 mL is sufficient to operate thousands of these channels , provided the appropriate liquid handling and data acquisition equipment is used . In order to systematically test the large number of compounds provided by liquid chromatography methods , the microscale platform was used to screen for compounds inhibiting neutrophil recruitment . The purified compounds were introduced in conjuncture with a known chemoattractant ( fMLP ) in the gradient-source reservoir , causing the recruitment of neutrophils into the gradient microchannels ( Figure 2A , B ) . A decrease in the number of neutrophils recruited compared to the positive control suggests an inhibitory effect and is considered as a positive hit for the screen . Using this method , a set of 20 purified compounds ( each inserted in a volume of 3 µL at 10 µM , n = 9 ) from A . nidulans and A . fumigatus was screened ( Figure 2C ) . Results show that four compounds – 8-hydroxyl emodin , austinolide , F-9775B , and endocrocin – display a significant decrease in neutrophil migration . As in vitro assays do not account for the complexity of whole organisms , we developed a low-volume in vivo neutrophil recruitment assay based on advances in zebrafish models . Zebrafish are a very attractive model for developing neutrophil research assays as they are transparent , allow the observation of immune cells in vivo , low-cost , and importantly have a strong immunological resemblance to the human system [26] . However , most neutrophil recruitment assays in zebrafish to date were performed through a wounding assay in which the fish is wounded by a needle stab or a cut in the fin or the body [28] . This approach has been successful but requires the individual manipulation of each fish , which is not amenable to screening applications . We developed a neutrophil recruitment inhibition assay based on the ability of a known chemoattractant , leukotriene B4 ( LTB4 ) , to diffuse through the skin of the zebrafish and induce neutrophil recruitment out of the Caudal Hematopoietic Tissue ( CHT ) . In conjuncture with a neutrophil migration inhibitor added to the well in which the zebrafish are bathing , the level of recruitment can be reduced and easily assessed ( Figure 3A ) . We validated this assay using LY294002 , a known neutrophil chemotaxis inhibitor targeting PI3K , and found that neutrophil recruitment to the CHT was entirely suppressed ( Movie S1 and S2 ) . This assay is rapid and accessible as many zebrafish are treated simultaneously in a multi-well plate and it does not require the use of wounding or micro-injection methods . Furthermore , the fish can be readily fixed and neutrophils quantified using simple optical microscopy ( Figure 3B ) . Based on availability , three of the four compounds identified in the in vitro screen were tested for their inhibitory properties to neutrophil chemotaxis . Results show that only endocrocin displayed a significant reduction in the number of neutrophils recruited to the tail fin ( Figure 3B , C ) . We further assessed the general cytotoxicity of the compound , and did not observe a reduction in zebrafish survival after 4 days at endocrocin concentrations up to 10 µM ( data not shown ) . Similarly , endocrocin did not induce an increase in neutrophil death in the timescale and concentrations used in the microfluidic assays ( Figure S1 ) . We investigated properties of endocrocin production in order to gain insight on its mode of interaction with a host . Endocrocin is a LaeA regulated anthraquinone whose biosynthesis pathway was recently identified in A . fumigatus [29] . Given the potential role of endocrocin for modulating the immune response , we sought to characterize the mode of production and tissue specificity of this SM . We analyzed the production of endocrocin in both the WT strain and a ΔencA strain of A . fumigatus grown on GMM-agar at temperatures varying from 25°C to 42°C ( Figure 4A ) . At temperatures below 35°C , the WT fungus produces endocrocin , while at higher temperatures it does not . As expected , the ΔencA strain did not produce endocrocin . The location of endocrocin production was characterized by analyzing crude extracts from different fractions of the fungal culture grown on solid medium: the conidia , the mycelia ( top agar ) , and secreted metabolites ( bottom agar ) . Endocrocin was only observed in the conidial fractions and not in the fractions containing mycelia ( top agar ) or soluble secreted factors ( bottom agar , Figure 4B ) . We identified the functional range of endocrocin inhibition by performing a dose response assay of endocrocin using the microfluidic neutrophil migration platform . Results show that a significant reduction of neutrophil inhibition can be observed at concentrations as low as 100 nM , with an inhibition of up to 40% for concentrations of 10 µM ( Figure 5 ) . Finally we assessed the pathogenicity attribute ( s ) of products from the endocrocin gene cluster in vivo by using the established Toll-deficient Drosophila invasive aspergillosis model [30] , [31] . This model , while not possessing the immune system of mammalians , is utilized successfully for studying virulence of A . fumigatus and , importantly , meets the affordable and rapid methodology of this study . Drosophila were inoculated with A . fumigatus wild-type and ΔencA spores harvested from 25°C ( endocrocin stimulating ) or 37°C ( endocrocin restrictive ) environments . Attenuated virulence was observed when flies were inoculated with spores of ΔencA grown at 25 but not 37°C , consistent with the observed temperature-dependent production of endocrocin ( Figure 6A , B ) .
Modern metabolomic techniques provide the ability to identify and isolate hundreds of microbial compounds in academic research settings . Downstream of these techniques , there is an increasing need for low-volume platforms that can help map the multi-kingdom interactome in a low-cost laboratory setting . We present here a novel and economical approach for identifying immunosuppressive properties of microbial compounds . The two-tiered approach offers an accessible solution for performing a screen both in vitro , on human primary neutrophils , and in vivo , on whole organism zebrafish models . Both methods described are compatible with natural product extraction methods with small yields and are scalable as they allow batch processing . The combination of assessing the response of human primary cells - a more precise model for medical studies - and using zebrafish for a more holistic in vivo assessment of specific innate immune responses , covers a highly relevant scope of models for predicting the effect of metabolites in humans . While a multitude of microfluidic gradient and neutrophil migration platforms have been demonstrated to date , this is the first systematic assessment of the neutrophil migration properties of Aspergilli SMs in a microfluidic embodiment . It is worthy to note , however , that due to material interaction and a higher surface-to-volume ratio in microscale devices , it is possible for the compound tested to be sequestered by the material and not be able to diffuse to the neutrophils , leading to false negatives . Using this microscaled approach , we find that four of the compounds tested displayed migration inhibitory properties . The in vivo zebrafish assay complements the in vitro assay as it uses an equally scalable and potentially automatable approach , as specific manual treatment of the fish used for each compound is not necessary . One potential caveat is the necessity for the compound to traverse the skin barrier in order to have an effect on the neutrophils . In this aspect , the assay can be made more sensitive by pre-incubating the fish in the purified compounds prior to adding the chemoattractant . Together , these assays represent a new avenue for the discovery of immunosuppressive properties of microbial SMs . As an example , we found that one of the compounds selected from the in vitro screen , endocrocin , displayed significant leukocyte recruitment inhibitory properties . Endocrocin was only recently identified as a LaeA regulated SM in A . fumigatus [28] . Coupling the observations that laeA loss yields reduced virulent strains and that several studies have shown uncharacterized components of A . fumigatus filtrates inhibit neutrophil chemotaxis , we suggest that endocrocin is one of these components . The fact that endocrocin did not result in zebrafish death nor increase neutrophil death under the assay conditions used ( Figure S1 ) further supports a more specific function for this metabolite than mere cytotoxicity . The study of endocrocin showed that it is located in the fungal spores , and may be released upon contact with a humid environment such as the host lung tissues or upon early germination of spores that managed to evade macrophage clearance . Previous studies have found that certain fungal metabolites , associated with the cell wall of the inhaled spores , have the ability to interact detrimentally with the lung epithelia [32] , [33] . This study provides insight into the role of spore-borne metabolites as “protective” constituents during early lung colonization , in which the spores are pre-armed with metabolites that could provide the germinating spore with an advantage over the immune system . Spores containing these “protective” attributes that could modulate or negate early host innate confrontations – the first line of defense towards Aspergillus infections – may provide the fungus with leverage during this initial host-pathogen arms race . Thus , spore-borne metabolites , highly dependent on the micro-environmental conditions in which the spores originate ( temperature , nutrient sources , microbial interactions , etc ) may play an important and understudied role in the pathomechanism of fungal opportunistic diseases .
Purified fungal metabolites were obtained by culturing Aspergillus species ( A . fumigatus and A . nidulans wild-type and mutant stains ) in liquid shake or solid agar conditions . For the liquid shake culture , the supernatant was collected and small molecules extracted by freeze drying and methanol extraction . For solid agar cultures , the agar was homogenized and soaked in 800 mL of 1∶1 CH2Cl2/Methanol for 24 h . After filtration , the combined extract was evaporated in vacuo to yield a residue , which was suspended in water ( 500 mL ) and then partitioned with ethyl acetate ( 500 mL ) three times . The combined ethyl acetate layer was evaporated in vacuo to afford a crude extract ( the weight for each deletant is list below ) . The crude extract was applied to a Si gel column ( Merck , 230 to 400 mesh , ASTM , 20×80 mm ) and eluted with 250 mL CH2Cl2/Methanol mixtures of increasing polarity ( fraction A , 1∶0; fraction B , 19∶1; fraction C , 9∶1; fraction D , 7∶3 ) . Each fraction was examined by high performance liquid chromatography–photodiode array detection–mass spectrometry ( HPLC–DAD–MS ) and the fraction contained target natural products was applied to a gradient HPLC on a C18 reverse phase column ( Phenomenex Luna 5 µm C18 [2] , 250×10 mm ) with a flow rate of 5 . 0 mL/min and measured by a UV detector at 254 nm . The gradient system was acetonitrile ( solvent B ) and 5% Acetonitrile/H2O ( solvent A ) both containing 0 . 05% Tetra-Fluoroboric Acid . The neutrophils were purified from whole blood of consenting self-reported healthy donors . A volume of 7 mL of whole blood was placed in a 15 mL conical tube and 7 mL of Polymorphoprep liquid ( Axis Shield , UK ) was layered above . The conical tube was centrifuged for 20 min at 1200 rpm followed by 10 min at 1700 rpm . The neutrophil layer was removed , placed in a 50 mL conical tube , and 1× PBS was added to fill the tube up to 50 mL . The conical tube was subsequently centrifuged for 10 min at 1500 rpm . The pellet was re-suspended in 9 mL of de-ionized H2O for 30 s after which 1 mL of 10× PBS and 40 mL of 1× PBS were added . The conical tube was centrifuged again for 10 min at 1500 rpm , and the pellet was re-suspended in 1 mL of PBS . The neutrophils were counted and the cell suspension was diluted to a final concentration of 4 million per mL . A silicon-based mold for the microfluidic device-arrays was fabricated by creating the design on Illustrator ( Adobe , USA ) , and printing on a high-resolution film ( Imagesetter , Inc , Madison , USA ) . The main channels of the microfluidic device were designed to be 1 mm wide and 400 µm tall , while the gradient channel is 35 µm tall , 150 µm wide , and 1 mm long . A photo sensitive epoxy , SU-10 ( Microchem , USA ) , was spun on a 150 mm diameter wafer ( WRS , USA ) to a thickness of 35 µm and backed for 10 min at 95°C . The first film defining the migration channels was placed on the wafer and exposed to 200 mJ of UV light from an Omnicure light source ( EXFO , Canada ) , after which the wafer was post-backed at 95°C for 5 min . A second layer of epoxy , SU-100 , was spun on the wafer to a thickness of 400 µm , backed for 90 min at 95°C , exposed with 1200 mJ of UV light with the second film defining the microfluidic channels , and backed for 30 min . A third layer of SU-100 was spun and exposed with the same parameters as the second layer , albeit with the third film defining the ports . The wafer was developed in SU-developer ( Microchem , USA ) for 3 hours on a shaker , washed with acetone , rinsed with iso-propanol , and dried using compressed air . Microfluidic arrays were designed and fabricated using silicone polymer-based soft-lithography methods by replicating a silicon-SU8 master mold [34] . In brief , a silicone polymer , PDMS ( Sylgaard 184 , Dow Corning , USA ) , was mixed with a ratio of 1∶10 of curing agent to monomer base and placed in a vacuum for 30 min for degassing . The molding process was performed by placing , in order , on a hot plate a transparency sheet ( Cheap Joe's , USA ) , the silicon-based mold , degassed PDMS , a second transparency sheet , a 1 mm thick layer of silicon foam ( Mc Master , USA ) , a rectangle of glass , and a 5 kg weight . The hot plate was heated to 80°C for 3 hours and cooled down completely prior to PDMS removal . The PDMS layer was peeled off the silicon-based mold , bonded non-covalently in a polystyrene Omnitray dish ( NUNC , USA ) . The Omnitray was filled with 20 mL of PBS and placed in a vacuum chamber for 15 min in order to fill the microdevices . The superfluous PBS was removed , 4 µL of mHBSS ( 1× PBS containing 0 . 1% HSA and 0 . 2 mM of HEPES buffer ) were placed on the output well and 4 µL were placed on each input well in order to replace the fluid contained in each device . Purified dried compounds were resuspended in DMSO to 1 mM and subsequently diluted 1∶100 in mHBSS containing 100 nM of fMLP – a known chemoattractant . 4 µL of neutrophil suspension at 4 . 106 cells per mL was inserted into the sink channel of the device , followed by 3 µL of compound in the source channel , and the Omnitrays were placed in a CO2 incubator . After 45 min , the Omnitrays were removed , placed on a Nikon Eclipse microscope , and phase-contrast images of the migration channels were taken at 10× magnification . Image analysis was performed using the software package Je'Xperiment developed in-house ( source code available on sourceforge . net or upon request ) allowing the identification and quantification of the neutrophils invading the migration channel on a batch processing scale . A Wilcoxon rank-sum test was performed on sets of 5 data points to evaluate the statistical significance of the effect of each compound . Those with a p-value of less than 0 . 05 were considered statistically significant . Zebrafish were maintained according to the protocols approved by the University of Wisconsin-Madison Research Animal Resources Center . Zebrafish larvae at 3 days post fertilization were placed in the wells of a 24 well-plate ( 20 per well ) . The larvae were pre-incubated in 500 µL of E3 ( egg water ) containing 10 µM of the purified fungal compound in DMSO . After 1 hour , LTB4 was added to each well at a final concentration of 30 nM . The larvae were fixed after 30 min and stained using Sudan Black [26] . Numbers of neutrophils recruited to the ventral fin of each fish were counted manually and representative images were taken with a phase contrast upright microscope . Statistical significance was estimated using a Kruskal-Wallis test followed by Dunn's Multiple Comparison test in the software GraphPad Prism ( USA ) . The animal handling protocols were performed according to the Guide of the Care and Use of Animals of the National Institutes of Health . Aspergillus fumigatus strains used in this study are listed in Table S1 . Strains were maintained as glycerol stocks and activated on glucose minimal media ( GMM ) . Conidia were harvested in 0 . 01% Tween 80 and enumerated using a hemocytometer . Strain TFYL7 . 1 was constructed by targeted gene deletion of encA using a deletion cassette made via double-joint fusion PCR described in [29] ) into strain AF293 . 1 . Internal primers to encA was used to confirm the absence of its open reading frame and single integration of the deletion cassette was confirmed via Southern analysis using two different restriction digest profiles as described in [29] ) ( data not shown ) . For temperature-dependent characterization of endocrocin production , A . fumigatus WT and ΔencA strains were point-inoculated at 1×104 conidia/inoculum onto solid GMM and incubated at temperatures ranging from 25–42°C without light selection . A 1 . 2 cm diameter core was removed from the middle of the fungal culture and homogenized in 2 mL of 0 . 01% Tween 80 . The homogenized mixture was extracted with equal volumes of ethyl acetate and routine vortexing at room temperature over the course of 30 min . The mixture was centrifuged for 5 min at 3 , 500 rpm and 1 mL of the ethyl acetate layer was removed and allowed to evaporate at room temperature to yield a dried crude extract . For TLC analysis , the crude extract was reconstituted in 50–100 µL of ethyl acetate and 5–10 µL was spotted onto a 250 µM analytical silica plate ( Whatman , Cat 4410-222 ) and subjected to a toluene: ethyl acetate: formic acid ( 5∶4∶0 . 8 ) resolving phase . Plates were visualized at 366 nm using a FOTO/Analyst® Investigator gel imaging system ( Fotodyne Inc ) . For tissue specific extraction of endocrocin , a suspension of 1×106 conidia in molten GMM top agar ( 0 . 75% w/v agar ) was overlaid over solid GMM bottom agar ( 1 . 5% w/v agar ) . SM from different developmental parts of the fungus were obtained as follows: conidia was obtained by gently tapping the fungal culture with the petri dish lid side down as described previously [35] , The remaining conidia and conidiophores were removed by gently scraping the surface of the fungal culture with 0 . 01% Tween 80 followed by multiple washes to remove residual conidia/conidiophores . The mycelia ( free of aerial conidiophores/conidia ) were obtained by peeling off the top agar from the culture plate . The bottom agar that is now free of mycelia ( as inspected under the microscope ) was extracted . A diagram that depicts the various sampling parts of the culture plate can be found in Figure 3B . Extraction process and TLC analyses were performed as described above . We developed an image-processing algorithm that is able to identify un-labeled neutrophils in a phase-contrast image and quantify their number of migration distance ( Figure1F ) . Used in conjuncture with a software platform developed in-house , called Je'Xperiment ( Source code available on sourceforge . net or on request ) , which allows batch processing and data mining , we show a first step towards creating a high-throughput solution for quantifying the effect of fungal SMs on the inhibition of leukocyte migration . Flies were generated by crossing flies carrying a thermosensitive allele of Toll ( Tl r632 ) with flies carrying a null allele of Toll ( Tl I-RXA ) [36] . Two- to four day old adult female Toll-deficient flies were used in all of the experiments . Twenty flies were infected with each A . fumigatus strain used in this study . A . fumigatus isolates ( ΔencA and wild-type ) were grown on yeast extract agar glucose ( YAG ) either at 37°C or 25°C . Conidia were collected in sterile 0 . 9% saline from 2 days old cultures . The conidial concentration suspension was determined by using a hemacytometer and adjusted to 1×108 per mL . The dorsal side of the thorax of 20 CO2 anesthetized flies was punctured with a thin ( 10 µm ) sterile needle that had been dipped in a concentrated solution of A . fumigatus conidia ( 107 spores/mL ) . As a negative control group , Toll-deficient flies were punctured with a 10 µm sterile needle and monitored daily for survival . Flies that died within 3 h of the injection were considered to have died as a result of the procedure and were not included in the survival rate analysis . The flies were housed in a 29°C incubator to maximize expression of the Tl r632 phenotype [31] . The Toll-deficient flies were transferred into fresh vials every 3 days . Fly survival was assessed daily over 10 days . Each experiment was repeated 3 times on different days and at the same time of the day to eliminate variability due to circadian rhythm . Neutrophils were obtained from whole blood of self-reportedly healthy donors , from which we obtained informed and written consent at the time of the blood draw with approval of the University of Wisconsin-Madison Center for Health Sciences Human Subjects committee .
|
Several fungal pathogens produce bioactive small molecules , commonly known as secondary metabolites ( SMs ) that contribute towards disease development in susceptible hosts . Genome assessment of human pathogenic Aspergillus species indicates these fungi have the capabilities of producing hundreds of SMs , most of which are currently not characterized for their effect on human health and the immune system . This lack of knowledge is directly correlated to the difficulties of obtaining assayable quantities of pure metabolites . To overcome this roadblock in assessing the potential impact of SMs on the immune system , our laboratories have developed a two-tiered cost-effective , high-throughput program utilizing microfluidic platforms and a novel zebrafish model to identify SMs inhibiting neutrophil chemotaxis . Using minimal and physiologically relevant amounts of SMs , this systematic approach has identified the A . fumigatus spore SM , endocrocin , as a potent chemotaxis inhibitor . Interestingly , the production of endocrocin is temperature dependent and virulence studies with the endocrocin null mutant implicates the temperature at which the fungus forms spores as a factor in disease development .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"bioengineering",
"biomedical",
"engineering",
"biotechnology",
"mycology",
"fungi",
"immunology",
"biology",
"microbiology",
"immune",
"response",
"engineering"
] |
2013
|
Low-Volume Toolbox for the Discovery of Immunosuppressive Fungal Secondary Metabolites
|
Myosin VI has been studied in both a monomeric and a dimeric form in vitro . Because the functional characteristics of the motor are dramatically different for these two forms , it is important to understand whether myosin VI heavy chains are brought together on endocytic vesicles . We have used fluorescence anisotropy measurements to detect fluorescence resonance energy transfer between identical fluorophores ( homoFRET ) resulting from myosin VI heavy chains being brought into close proximity . We observed that , when associated with clathrin-mediated endocytic vesicles , myosin VI heavy chains are precisely positioned to bring their tail domains in close proximity . Our data show that on endocytic vesicles , myosin VI heavy chains are brought together in an orientation that previous in vitro studies have shown causes dimerization of the motor . Our results are therefore consistent with vesicle-associated myosin VI existing as a processive dimer , capable of its known trafficking function .
Class VI myosins are found in a variety of organisms from Caenorhabditis elegans to human , and in a variety of cell types ( reviewed in [1 , 2] ) . Unlike other characterized myosins , they move toward the pointed end of an actin filament [3] , and so are capable of functions unique from other myosins . For example , during clathrin-mediated endocytosis , myosin VI is implicated in trafficking vesicles that have recently shed their clathrin coat , denoted uncoated vesicles ( UCV ) . The motor transports UCV from the periphery of a cell to its interior , presumably along actin filaments in the cell periphery that are oriented with pointed ends directed toward the cell interior [4–6] . The motor's heavy chain contains an N-terminal catalytic head followed by a unique myosin VI insert and an IQ motif , each of which can bind a single calmodulin [7 , 8] . The calmodulin binding domains are followed by a tail domain ( TD ) that is predicted to be highly α-helical . The C-terminal domain is the motor's cargo-binding domain ( CBD ) , a region implicated in association of the motor with its protein cargo [9–11] ( Figure 1A ) . Myosin VI heavy chains have been hypothesized to dimerize [7 , 12 , 13] . This model is supported by single-motor optical trap assays that utilized a motor construct containing a GCN4 leucine-zipper domain in the C-terminal region of the TD , ensuring dimerization of the motor even under the dilute ( pM ) conditions of single-molecule assays [14] . This dimer walks processively along actin , meaning it takes numerous steps along a filament before dissociating [15] . Its stepping is highly coordinated , with mechanical strain regulating the biochemical behavior of the molecule , resulting in head-to-head communication and proper in vivo function [16 , 17] . Surprisingly , however , Lister et al . [18] demonstrated that the myosin VI heavy chain , when purified from a baculovirus expression system or observed in extracts from rat kidney fibroblastic tissue culture , exists as a monomer . Though an ensemble of monomeric motors may be capable of myosin VI's predicted trafficking function , such an ensemble is not ideal for trafficking because the motor has a high duty ratio [16] , and so monomers attached to actin would work against newly attached and stroking motors . On the other hand , as a coordinated processive dimer , the motor would be well suited to traffic cargo efficiently with relatively few motors , as demonstrated by in vitro studies of a myosin VI dimer [15 , 17] . We thus speculated that , in regions of the cell where myosin VI performs trafficking function , dimerization of the motor occurs in a regulated manner . Park et al . [19] demonstrated that monomeric myosin VI motors lacking the CBD can dimerize in vitro if they are brought into close proximity , suggesting that myosin VI may be capable of in vivo dimerization in regions of high local motor concentration . However , the CBD appears to somewhat inhibit this dimerization , indicating that dimerization may require a proper positioning of the monomers . Given the above considerations , we hypothesize that two myosin VI CBDs are precisely positioned close together when loaded onto a vesicle and that this positioning orients the motor appropriately for dimerization ( Figure 1B ) . This would then allow the motor to perform its predicted trafficking function as a processive dimer [5] . Similar regulation of a motor protein between a monomer and dimer has been proposed for the C . elegans kinesin Unc104 [20 , 21] . However , mechanisms of dimerization for Unc104 and myosin VI are both inferred from in vitro data , and there is a lack of evidence indicating that this dimerization occurs in vivo . Here , we provide evidence for formation of a myosin VI dimer in vivo .
We conducted our studies in ARPE-19 cells , a human retinal pigment epithelial cell line that was one of the lines used by Dance et al . [6] in their in vivo studies of myosin VI . Dance et al . demonstrated in various cell lines that , during clathrin-mediated transferrin uptake , endogenous myosin VI colocalizes with transferrin-containing UCV [6] . Figure 1C shows a green fluorescent protein ( GFP ) image of an ARPE-19 cell expressing full-length myosin VI with an N-terminal GFP ( GFP-FL ) . After transfection with the GFP-FL construct , cells exhibited two distributions of GFP fluorescence: a homogeneous GFP haze throughout the cytosol and small , bright GFP puncta that exist throughout the cell , though are often more dense in the cell periphery ( Figure 1C ) . Dance et al . [6] observed similar colocalization for endogenous myosin VI . To verify that the myosin VI puncta correspond to UCV , we observed endocytosis of transferrin conjugated with Alexa 647 dye ( Alexa647-Tfn ) . Transferrin is known to be internalized via the clathrin-mediated endocytic pathway . GFP puncta showed a high degree of colocalization with internalized fluorescent transferrin immediately after internalization ( Figure 1C ) . We next sought evidence for precise positioning of myosin VI motors on UCV . As discussed in the introduction , we speculated that myosin VI heavy chains , although likely to be monomeric in the cytoplasm [18] , are brought into close proximity on its cargo , allowing the motor to function as a dimer . Two primary myosin VI–truncated CBD constructs were used for these studies ( Figure 1A ) . The first is a myosin VI containing the CBD as well as 17 residues from the TD N-terminal to the CBD ( we refer to this simply as the CBD construct ) . The second construct is the CBD construct with a leucine zipper ( GCN4 ) attached at its N-terminus ( GCN4-CBD ) , which forces it to dimerize [14] . These constructs were made fluorescent by inserting a monomeric GFP isoform [22] at the N-termini of CBD ( GFP-CBD ) and of GCN4-CBD ( GFP-GCN4-CBD ) . After transfection of ARPE-19 cells with these constructs , cells exhibited the same two distributions of GFP fluorescence as cells transfected with GFP-FL: a homogeneous GFP haze throughout the cytosol and small , bright GFP puncta throughout the cell ( Figure 1D and 1E ) . We also observed endocytosis of transferrin conjugated with Alexa 647 dye in these cells . GFP puncta showed a high degree of colocalization with internalized fluorescent transferrin immediately after internalization ( Figure 1D and 1E ) . Thus , our constructs have maintained their ability to associate with UCV similarly to endogenous myosin VI [6] . In movies of cells expressing our GFP-tagged myosin VI–CBD constructs , UCV exhibited motion throughout the cell , with UCV toward the periphery of the cell typically exhibiting slower velocities relative to those further into the cell ( see Videos S1–S3 ) . For cells expressing GFP-FL , the slower motion likely corresponds to myosin VI–dependent movement of UCV through the thick actin-mesh at the cell periphery . This peripheral mesh is particularly thick in ARPE-19 cells , and UCV travel a net distance of approximately 2 μm through the actin in a process that takes on the order of 5 min [5] . For cells expressing myosin VI constructs lacking the catalytic head , the GFP construct competes with the endogenous motor for binding to the UCV , and acts as a dominant negative . In these cells , the slower motion corresponds to Brownian-like motion with a slow drift toward the interior of the cell [5] . The faster motion observed deeper in the cell may result from UCV moving on microtubules , consistent with predictions that a microtubule network is involved in intracellular trafficking of UCV from the early to late endosomes [23] . Fluorescence resonance energy transfer ( FRET ) is the nonradiative transfer of energy between fluorophores occurring when the emission spectrum of an excited fluorophore overlaps with the absorption spectra of a fluorophore in very close proximity ( within 10–100 Å ) [24] . FRET between identical fluorophores ( homoFRET ) serves as an ideal way for detecting homo-oligomeric protein configurations [25 , 26] . According to our proposed mechanism for myosin VI function when bound to its cargo ( Figure 1B ) , the CBDs of UCV-associated myosin VI heavy chains are positioned to bring together the heavy chains . By analogy , this mechanism also predicts that , for two UCV-associated GFP-CBD constructs , the CBDs are positioned to bring into close proximity their associated GFPs . Thus , homoFRET of GFP-CBD on the UCV serves as a readout of our proposed mechanism . GFP-GCN4-CBD serves as a positive control for detection of homoFRET . The leucine zipper forces the construct to form a constitutive dimer , resulting in close association of the GFPs adjacent to the GCN4 coiled coil and subsequent homoFRET . As a negative control , we used a construct that is identical to GFP-CBD , except the GFP is located at its C-terminus ( CBD-GFP ) . For this construct , we expect that GFPs are not likely to be positioned to undergo homoFRET , even if the CBDs are brought close together . To quantify the levels of GFP-FL , GFP-CBD , GFP-GCN4-CBD , and CBD-GFP in the cytosol and on UCV , we used multiphoton microscopy to gather confocal images of GFP fluorescence from ARPE-19 cells expressing these constructs . A sample image collected for GFP-CBD is shown in Figure 2 ( top ) . Fluorescence emission was collected with multichannel plate photomultiplier tubes capable of photon counting . For an imaged cell , the mean photon count was calculated in numerous regions corresponding to the UCV and the cytosol ( for example , see Figure 2 , top ) , and these values were averaged to arrive at the cell's mean fluorescence intensity at each localization . The mean fluorescence intensities for multiple cells were then averaged to arrive at the overall mean fluorescence intensities at UCV and in the cytosol for each construct ( see Materials and Methods ) . The total expression of all constructs was similar , as were their levels on UCV and in the cytosol ( Figure 2 , middle ) . We excited the fluorophores of our GFP-tagged CBD constructs with pulsed , polarized excitation , and observed subsequent changes in fluorescence emission polarization , as quantified by the fluorescence anisotropy , over time . The emission polarization is initially aligned with the excitation polarization , resulting in a high initial anisotropy , and becomes randomized over the lifetime of the fluorophore through two processes: ( 1 ) rotational diffusion of the GFPs and ( 2 ) energy transfer to GFPs in close proximity ( on the order of the Förster's radius ) [27] . Rotational diffusion and homoFRET each result in exponential decays in anisotropy which , for large proteins , occur on very different time scales [26 , 28]: homoFRET results in a rapid anisotropy decay , and rotational diffusion results in a slower decay ( Figure 3 , top ) . From the former , we can detect processes that bring GFPs into close proximity , and from the latter , we can infer the size of the rotating object . We measured fluorescence anisotropy following polarized multiphoton excitation with a pulsed laser ( ∼12-ns repetition rate ) using time-correlated single-photon counting ( TCSPC ) [26 , 29] ( see Materials and Methods ) . Pico-second time-resolved anisotropy decays were measured for our three GFP-tagged CBD constructs in the cell periphery , both in the cytosol and at the UCV . An example of the regions selected for these measurements is shown in Figure 2 ( top ) . We selected UCV in the cell periphery to be sure the construct is associated with UCV in the peripheral actin network . To further ensure that we selected for these vesicles , we collected an image encompassing the area of the measured UCV both before and after the measurement , an interval lasting approximately 1 min . UCV that remained in the observation volume before and after the measurement corresponded to slowly moving UCV that were associated with the peripheral actin network and so were selected for analysis . We fit two decay models to each empirical anisotropy decay: ( 1 ) a single exponential decay and ( 2 ) the sum of two exponential decays ( see Materials and Methods and Figure 3 ) . These fits revealed two classes of decay profiles . For the first class , the profiles were well fit by a single exponent; the addition of another exponent had little effect on the fit . These profiles describe decay in anisotropy through only a single process , presumably rotational diffusion . For a second class , the decay was not fit well by a single exponent , but the addition of a second exponent resulted in a good fit ( for example , see Figure 3 , bottom ) . These profiles describe decay in anisotropy through two exponential processes , both homoFRET and rotational diffusion . In this manuscript , we describe in detail the best fits for all empirical decays , the first class of decays to a single exponent and the second class to the sum of two exponents ( Figure 4 and Table 1 ) . As a control to test the instrumentation , we transfected cells with monomeric GFP . The GFP homogeneously filled the cytoplasm , and anisotropy decay profiles collected from cytosolic GFP were well fit by a single exponent ( Figure S1 ) . The time scale of this decay ( ∼25 ns ) is consistent with previous measurements of GFP tumbling in the cytosol [30] . For CBD-GFP , anisotropy decays measured both in the cytosol and at UCV were well fit by a single exponent ( Figure 4 , middle ) , consistent with our expectation that the construct does not undergo homoFRET and that anisotropy decreases only through fluorophore rotation . The time scale of this decay in the cytosol is consistent with tumbling , and the decay is considerably slower at the UCV , consistent with a slowed rotation due to association of the CBD-GFP with a UCV ( Table 1 ) . Anisotropy decays collected for GFP-GCN4-CBD both in the cytosol and at UCV could not be fit by a single exponential but were well fit by the sum of two exponents ( Figure 4 , right ) , consistent with our expectation that anisotropy decreases both through rotational diffusion and through homoFRET of the dimeric construct . The time scales for the fast and slow decays are consistent with homoFRET and tumbling , respectively . As with CBD-GFP , the decay corresponding to rotation is slower at the UCV relative to the cytosol due to association with the UCV ( Table 1 ) . Anisotropy decays collected from GFP-CBD in the cytosol were well fit by a single exponent . Anisotropy decays collected at UCV , however , were only well fit by the sum of two exponents ( Figure 3 , bottom , and Figure 4 , left ) . The rapid anisotropy decay at UCV is consistent with a homoFRET process . Both the slower decay at the UCV and the decay in the cytosol are consistent with rotational diffusion . As with the other GFP-constructs , the decay describing rotation is slower at the UCV compared to the cytosol ( Table 1 ) . From these data , we infer that GFP-CBDs are positioned on UCV to bring their N-termini together ( Figure 1B ) . The lack of homoFRET in the cytosol confirms that this precise positioning requires the construct to be loaded onto the vesicle . This result is consistent with our prediction that a precise positioning of CBDs on a vesicle orients heavy chains in close proximity . For anisotropy decays measured in the cytosol , the time scale describing rotational diffusion ( Table 1 ) provides information about the size of the GFP construct . Because these decay times are longer than the fluorescence lifetime of GFP ( ∼3 ns ) , unambiguous molecular weights cannot be determined . However , we can infer relative sizes of our constructs from these decays . Rotation of GFP-CBD is slower than for GFP alone ( Figure S1 ) , consistent with slowed rotational diffusion of the fluorophore when attached to the myosin VI construct . The 2-fold difference in the decay time suggests that the molecular weight of GFP-CBD is twice that of GFP ( molecular weight , 27 kDa ) , consistent with a monomeric form of GFP-CBD ( molecular weight , 58 kDa ) . The rotational decay time of GFP-CBD is also similar to that of CBD-GFP , and both constructs rotate faster than GFP-GCN4-CBD ( Figure S1 ) . Thus , both GFP-CBD and CBD-GFP appear to be smaller than a similarly sized GFP construct known to dimerize , supporting our prediction that both are monomeric in the cytosol . To further demonstrate that homoFRET occurs when GFP-CBD is associated with UCV , we determined both the steady-state fluorescence emission and steady-state fluorescence anisotropy throughout cells expressing GFP-CBD ( Figure 5A ) . Steady-state anisotropy represents the integral over time of an anisotropy decay profile , and so it is reduced by both rotational diffusion and homoFRET ( Note the areas under the curves in Figure 3 , top ) , though these processes cannot be distinguished by steady-state analysis [31] . In all cells analyzed , steady-state anisotropy for GFP-CBD was clearly lower at UCV relative to the surrounding cytosol ( Figure 5A ) . Considering only the effects of fluorophore tumbling , we would have expected steady-state anisotropy to be lower in the cytosol , where rotational diffusion is more rapid . The observed pattern of steady-state anisotropy is thus consistent with a further reduction in steady-state anisotropy at the UCV due to homoFRET . To quantify this , we manually selected from the periphery of each cell numerous UCV as well as 30 regions in the cytosol , similar in size to the UCV , and calculated the steady-state anisotropy at these regions ( see Materials and Methods ) . We observed that the mean steady-state anisotropy of these regions at either localization is not consistent from cell to cell , due to a variety of factors such as cell thickness . However , for all ten cells analyzed , we observed that the mean anisotropy at the UCV was consistently lower than the mean anisotropy in the cytosol . An example of analysis of a single cell is shown in Figure 5B . As a control , we also determined the steady-state fluorescence anisotropy and emission throughout cells expressing CBD-GFP . Because CBD-GFP does not exhibit homoFRET at UCV , its steady-state anisotropy should be dictated solely by its rotational diffusion . Thus , we expect a higher anisotropy at UCV relative to the surrounding cytosol , in contrast to the pattern observed for cells expressing GFP-CBD . We observed this expected pattern of steady-state anisotropy for all cells analyzed ( Figure 5C ) , and again we quantified this by calculating steady-state anisotropy at numerous UCV regions and at 30 regions corresponding to the cytosol . For all nine cells analyzed , we observed that the mean steady-state anisotropy at the UCV was higher than the steady-state anisotropy in the cytosol . An example of analysis of a single cell is shown in Figure 5D . Our time-resolved and steady-state anisotropy experiments demonstrated that GFP-CBD undergoes homoFRET when localized to UCV . Though we hypothesize that this is the result of precise positioning of the construct on a UCV , we must also consider the possibility that GFPs are brought into close proximity simply due to the crowding of a high density of GFP-CBD on the UCV surface . From our analysis of steady-state GFP-fluorescence emission images of cells expressing GFP-CBD and CBD-GFP , we observed that both constructs are expressed to similar levels in our cell line ( Figure 2 , middle ) . Furthermore , both constructs exhibit similar ratios of GFP intensities on UCV and in the cytosol , indicating that they are loaded onto vesicles at similar densities and so are similarly crowded ( Figure 2 , bottom ) . Thus , if homoFRET of GFP-CBD were the result of crowding , we would also expect CBD-GFP to exhibit homoFRET when loaded onto UCV . Since this is not the case ( Figure 4 , middle ) , crowding cannot be the cause of homoFRET . Instead , the CBD must be positioned on a UCV so that its N-termini are brought together , resulting in homoFRET from an N-terminal ( and not a C-terminal ) GFP . We confirmed this conclusion by examining UCV associated with varying densities of GFP-CBD . If homoFRET were the result of crowding , then reducing the construct concentration on the UCV would reduce close packing of fluorophores and subsequently reduce the occurrence of homoFRET . On the other hand , if homoFRET results from precise positioning of GFP-CBD on UCV , then , even at low densities , the construct will undergo homoFRET . To differentiate between these mechanisms , we calculated the steady-state fluorescence emission intensity at the UCV regions selected from the previous steady-state anisotropy analysis ( see Figure 5 and Materials and Methods ) . Using these measurements , we probed for effects of GFP-CBD density on homoFRET by looking for effects of varying steady-state fluorescence emission intensity on steady-state fluorescence anisotropy ( Figure S2 ) . To determine the degree to which these measured values of anisotropy and intensity are related , we calculated the Pearson product-moment correlation coefficient ( r ) . For ten cells expressing GFP-CBD , the UCV regions of nine cells showed no significant correlation between steady-state fluorescence emission intensity and anisotropy ( p > 0 . 05 for nine cells , p = 0 . 02 for one cell ) . The lack of correlation indicates that the extent of homoFRET does not depend on the density of UCV-associated GFP-CBD . This supports our conclusion that homoFRET results from precise positioning of GFP-CBD on the vesicle , and not crowding of the fluorophores . As expected , when we performed a similar analysis for GFP-CBD regions in the cytosol , where the construct does not undergo homoFRET , we observed no correlation between fluorescence emission intensity and anisotropy . The same is true for the UCV and cytosolic regions of cells expressing CBD-GFP , which does not undergo homoFRET at either localization , and for GFP-GCN4-CBD , which undergoes homoFRET at both localizations due to precise positioning of its fluorophores ( unpublished data ) . In summary , understanding the in vivo functional form of a molecular motor is essential to understanding its function . Our data suggest that , although myosin VI exists as a monomer in the cytosol , heavy chains are brought into close proximity on UCV , allowing the motor to function as a dimer . Consistent with our model , Spudich et al . [13] reported that a myosin VI tail construct , when bound to artificial lipid vesicles in vitro , can be linked as dimers upon addition of a zero-length cross-linker . Through this mechanism , myosin VI is able to processively traffic its vesicular cargo through the actin meshwork in the cell periphery [32] .
GFP constructs were derived from the GFP-HM6Tail+LI construct from Dance et al . [6] , which consists of a myosin VI–CBD construct in the pEGFP-C3 expression vector ( Clontech , http://www . clontech . com ) . GFP-CBD was made from GFP-HM6Tail+LI by changing residue 206 of the GFP from alanine to lysine ( A206K ) , which reduces the proclivity of GFP to dimerize [22] . This was achieved through site-directed mutagenesis using the primer 5′-CCTGAGCACCCAGTCCAAGCTGAGCAAAGACCCCA-3′ and the QuikChange Site-Directed Mutagenesis Kit ( Stratagene , http://www . stratagene . com ) . To make GFP-GCN4-CBD , the leucine zipper from the myosin VI/GFP plasmid described in [33] was amplified using the primers 5′-CCCGAATTCTGGAAGACATGAAACAGCTCGAGGACAAAGTAGAGGAGCTGCTGTCCAAG-3′ and 5′-GCCCGCGGCTCCCCGACCAGCTTCTTAAGTCTCGCAACCTCATTTTCTAGATGG-3′ . The resulting PCR product was cut with EcoRI and SacII , and inserted into the MCS of the GFP-CBD plasmid . To make CBD-GFP , CBD was amplified from the GFP-CBD plasmid using the primers 5′-CGCCGCGGATGAGGATTGCCCAGAGTGAAGCCGAGCTCATCAGTGATGAGGCCC-3′ and 5′-TTGGATCCGCCTTTAACAGACTCTGCAGCATGGCTGTTGCATAGGTGGGCCGAGCCTG-3′ . The resulting PCR product was cut with SacII and BamH1 , and inserted into the multiple cloning site ( MCS ) of the pEGFP-N1 expression vector ( Clontech ) containing the A206K GFP mutation . The A206K mutation was made in pEGFP-N1 using the site-directed mutagenesis described above . ARPE-19 cells were purchased from American Type Culture Collection ( ATCC , http://www . atcc . org ) . Cells were grown at incubating conditions ( 37 °C and 5% CO2 ) in medium + serum ( DMEM/F-12 [GIBCO-Invitrogen , http://www . invitrogen . com] , 1% fungizone [GIBCO] , 1% L-glutamate [GIBCO] , 10% FBS [GIBCO] , 1 . 5 M HEPES , 100 U/ml penicillin , and 100 mg/ml streptomycin ) . Before transfection , cells were grown in imaging dishes that were polylysine-coated , with a translucent bottom appropriate for fluorescence imaging . To transfect cells , a transfection mixture , consisting of 75 μl of serum-free media ( SFM; medium+serum lacking FBS ) , 6 μl of TransIT Transfection Reagent ( Mirus Bio Corporation , http://www . mirusbio . com ) , and 1 μg of plasmid DNA , was added to the cell culture , which is in 0 . 75 ml of medium + serum . Cells were imaged 10–20 h after transfection . Transferrin was labeled with Alexa 647 ( Alex647-Tfn ) using the Alexa Fluor 647 Protein Labeling Kit ( Molecular Probes , http://probes . invitrogen . com ) . To observe uptake of Alexa647-Tfn , cells grown in imaging dishes were starved in SFM for 2 h at incubating conditions . The media was removed , and 150 μl of 10 μg/ml Alexa647-Tfn in SFM was applied to the cells . The cells were left at incubating conditions for 30 min . Cells were fixed by washing in M1 buffer ( 150 mM NaCl , 5 mM KCl , 1 mM MgCl2 , 1 mM CaCl2 , 20 mM HEPES [pH 7 . 4] ) and then adding 150 μl of 4% paraformaldehyde in M1 buffer . Cells were incubated for 25 min at room temperature , washed with M1 buffer , and imaged . For colocalization experiments , the transferrin-internalization protocol was begun 10 . 5 h after transfection of cells with the GFP construct . Briefly , experiments were done on a Nikon TMD fluorescence microscope with a cooled back-illuminated , 16-bit charge-coupled device ( CCD ) camera ( Nikon , http://www . nikonusa . com ) . Different filter sets were used to image Alex 647 and GFP . Images were collected using Metamorph software ( Molecular Devices , http://www . moleculardevices . com ) . Fluorescence imaging was carried out exactly as described [34] . Live-cell measurements of fluorescence anisotropy were made using TCSPC and pulsed multiphoton excitation . Details of the method and analysis will be described elsewhere ( D . Goswami , K . Gowrishankar , M . Rao , and S . Mayor , unpublished data ) . Briefly , steady state and time-resolved anisotropy measurements of fluorophores excited by multiphoton excitation were made on a Zeiss LSM 510 Meta microscope ( Carl Zeiss , http://www . zeiss . com ) with 63× 1 . 4 numerical aperture ( NA ) objective coupled to the femtosecond-pulsed Tsunami Titanium:Sapphire tunable pulsed laser ( Newport , http://www . newport . com ) . Parallel and perpendicular emissions were collected simultaneously into two Hamamatsu R3809U multi-channel plate photomultiplier tubes ( PMTs; Hamamatsu Photonics , http://www . hamamatsu . com ) using a polarizing beam splitter ( Melles Griot , http://www . mellesgriot . com ) at the non-descanned emission side . TCSPC was accomplished using a Becker & Hickl 830 card ( Becker and Hickl , http://www . becker-hickl . de ) , operating in a stop–start configuration [35] . For multiphoton excitation of GFP or fluorescein in cells , we used 920-nm excitation wavelength . At this wavelength , the two-photon absorption cross section for GFP is higher , enabling lower laser excitation power , and autofluorescence signals are minimized . The repetition rate of the pulsed laser is 80 . 09 MHz ( 12 ns ) . Steady-state imaging was accomplished using a pixel residence time of 102 μs/pixel , setting the detection time resolution in the Becker and Hickl card to one . Thus , a full image ( 512 × 512 pixel ) was collected over 62 s . For time-resolved anisotropy measurements , the time resolution was 12 . 2 ps . The beam was “parked” at a single point using routines available in the Zeiss software . The parked beam was placed at the center of the field to maintain uniformity of G-Factor , and photons were collected for 30–50 s . Photons were collected at a maximum rate of 0 . 1 MHz to ensure that TCSPC conditions were strictly met . Because of the low laser power , less than 10% bleaching was observed during a measurement . The instrument response function ( IRF ) was measured using 10–16-nm gold particles dried on a coverslip as a second harmonic generator; full width at half maximum ( FWHM ) of IRF is approximately 60 ps . In our experimental setup , the steady-state anisotropy measured while the laser beam is parked at a single point ( Table S2 ) was always higher than the steady-state anisotropy measured using the scanning mode ( Figure 5B and 5D ) . This is attributed to a small but detectable depolarization of the excitation laser beam in the scanning mode when using high NA objectives , which is also seen for measurements of a monomeric GFP solution . This effect is negligible for objectives with NA less than 0 . 8 ( unpublished data ) . A high NA objective was required to discern GFP associated with UCV as puncta distinct from the GFP cytosolic haze . Fluorescence lifetime and anisotropy decay analyses were done essentially as described [26 , 36 , 37] , with minor modifications in the analysis procedure . Briefly , the experimentally measured fluorescence decay is a convolution of the IRF with the intensity decay function . The intensity decay data were fit to the appropriate equations by an iterative reconvolution procedure using a Levenberg-Marquardt minimization algorithm . When fitting the models to the decay profiles , ro was constrained to a small window to improve the ability of the fitting algorithm to find the optimal fit . A constrained range of values for ro ( 0 . 43 ± 0 . 3 ) was used for all fits described in the manuscript . This range of values was obtained from unconstrained fitting of cytoplasmically expressed GFP fluorescence emission anisotropy decays ( n = 8 ) . These fits provided reliable values for the initial anisotropy because the fluorophore does not undergo homoFRET and because the time scale of rotation is much slower than the time scale of the measurement . When fitting a model describing two exponential decays to the decay profiles , the two decay times were somewhat constrained to a wide range of values . These constraints involved large windows centered on the expected decay times for the physical processes involved ( homoFRET and rotational diffusion ) . Again , these constraints improved the ability of the fitting algorithm to find the optimal fit . It is important to note that the empirical anisotropy decay profile is the convolution of the real-time behavior of the fluorophore with the IRF . This distortion results in an apparent fast decay at the start of all measurements that is an artifact and does not represent anything physical . This artifact is apparent because our sampling rate of 12 . 2 ps is smaller than the width of the instrument's IRF ( ∼60 ps ) . This effect is also apparent in the empirical decay for a monomeric GFP in the cytoplasm ( Figure S1 ) . The G-Factor was estimated using a fluorescein solution and setting the anisotropy at late times to 0 . 005 . Fluorescence and anisotropy decays were considered well fit if three criteria were met: reduced χ2 was less than 1 . 4 , residuals were evenly distributed across the full extent of the data , and visual inspection ensured that the fit accurately described the decay profile . For analysis of steady-state images , N nearest-neighbor averaging of a 512 × 512 array of pixel values refers to the following calculation: the pixel value at location ( row = i , column = j ) was set to the mean value of pixels spanning rows i − N to i + N and columns j – N to j + N . This calculation was done using software developed in Matlab ( The MathWorks , http://www . mathworks . com ) . Steady-state anisotropy was calculated from steady-state parallel- and perpendicular-polarization images . To account for differences in the optical paths traversed by the perpendicular and parallel emissions , a G-Factor correction was applied to the data as follows . We collected steady-state emission images from a fluorescein sample . Because fluorescein tumbles rapidly relative to the time scale of our measurements , fluorescein provides a pixel-by-pixel readout of the detector output from a source emitting identically in both polarizations . We created a G-Factor image from the parallel and perpendicular fluorescein emission images , and perpendicular images from subsequent experiments were multiplied by this G-Factor image to apply the appropriate correction . To create this G-Factor image , dividing the parallel and perpendicular fluorescein images pixel by pixel is insufficient , and results in a correction that is artificially too large . This is because pixel values in the parallel and perpendicular images exhibit Poisson photon noise . To reduce artifacts arising from dividing signals containing noise , the images must first be averaged so as to increase the signal-to-noise ratio . Using simulated data , we determined the extent of averaging required to sufficiently reduce the G-Factor artifact while retaining information about G-Factor variation across the image ( unpublished data ) . To create our G-Factor image , we first applied three-nearest-neighbor averaging to the parallel and perpendicular fluorescein images and then divided the averaged images pixel by pixel . A new G-Factor image was created for each day of experiments . After applying the G-Factor correction to the data , anisotropy was calculated at UCV and in cytosolic regions using software developed in ImageJ [38] . Regions were manually selected in the parallel image and were transferred to the perpendicular image . The mean perpendicular- and parallel-polarization emission intensities were calculated for each region , and from these , the steady-state fluorescence anisotropy and fluorescence emission intensity were calculated using the relations: and where I‖ and I⊥ are the calculated intensities of fluorescence emission with polarization parallel and perpendicular to the excitation polarization , respectively , I is the total fluorescence emission intensity , and r is the fluorescence anisotropy .
|
Myosin VI is a molecular motor implicated in diverse cell processes , including trafficking endocytic vesicles into the cell , transporting proteins to the leading edge of a migrating cell , and anchoring stereocilia to the hair cells of inner ear sensory epithelia . The motor has been studied in both a monomeric and dimeric form in vitro and is reported to exist as a monomer in the cytoplasm of cells . Because the functional characteristics of the motor are dramatically different for these two forms , an understanding of the activity of myosin VI requires an understanding of its functional form in vivo . To probe the role of myosin VI in vesicle trafficking , we labeled myosin VI truncations with a fluorescent protein and studied the positioning of these constructs on endocytic vesicles . We observed nonradiative transfer of energy between the fluorescent proteins , a process that can only occur if they are brought extremely close together . Our results indicate that , when myosin VI heavy chains bind to endocytic vesicles , they are precisely positioned very close together . Work from other laboratories indicates that myosin VI heavy chains brought together in this manner are capable of dimerization . Our results are therefore consistent with vesicle-associated myosin VI existing as a processive dimer , capable of myosin VI's known trafficking function .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"cell",
"biology",
"in",
"vitro",
"biophysics"
] |
2007
|
Precise Positioning of Myosin VI on Endocytic Vesicles In Vivo
|
Chemical and nutrient signaling are fundamental for all cellular processes , including interactions between the mammalian host and the microbiota , which have a significant impact on health and disease . Ethanolamine is an essential component of cell membranes and has profound signaling activity within mammalian cells by modulating inflammatory responses and intestinal physiology . Here , we describe a virulence-regulating pathway in which the foodborne pathogen Salmonella enterica serovar Typhimurium ( S . Typhimurium ) exploits ethanolamine signaling to recognize and adapt to distinct niches within the host . The bacterial transcription factor EutR promotes ethanolamine metabolism in the intestine , which enables S . Typhimurium to establish infection . Subsequently , EutR directly activates expression of the Salmonella pathogenicity island 2 in the intramacrophage environment , and thus augments intramacrophage survival . Moreover , EutR is critical for robust dissemination during mammalian infection . Our findings reveal that S . Typhimurium co-opts ethanolamine as a signal to coordinate metabolism and then virulence . Because the ability to sense ethanolamine is a conserved trait among pathogenic and commensal bacteria , our work indicates that ethanolamine signaling may be a key step in the localized adaptation of bacteria within their mammalian hosts .
Chemical and nutrient signaling mediate diverse biological processes , and underlie interactions among the mammalian host , the resident microbiota , and invading pathogens [1] . Ethanolamine is abundant in cell membranes , as a component of phosphatidylethanolamine as well as in modified lipid molecules such as N-acylethanolamines [2] . These ethanolamine-containing compounds play important roles in mammalian cell signaling and influence diverse physiological effects , including cytokinesis , immunomodulation , food intake and energy balance [2–4] . Ethanolamine is abundant in the intestinal tract due to the turnover and exfoliation of enterocytes and bacterial cells [5 , 6] , and intracellular pools of ethanolamine are maintained by low and high affinity uptake systems as well as through internal recycling of phosphatidylethanolamine [7–10] . Bacterial pathogens compete for nutrients with the resident microbiota and rely on environmental cues to control virulence gene expression . Salmonella enterica serovar Typhimurium ( S . Typhimurium ) is a facultative intracellular pathogen and a leading cause of acute gastroenteritis , which can progress to systemic infection in susceptible individuals [11] . S . Typhimurium encodes two type three secretion systems ( T3SSs ) that are important for pathogenesis . S . Typhimurium uses the T3SS encoded within the Salmonella pathogenicity island ( SPI ) -1 to invade intestinal epithelial cells and penetrate to the lamina propria [12] . There , S . Typhimurium is taken up by macrophages , where it survives and replicates . Intracellular survival is mediated by the T3SS and effectors encoded in SPI-2 [13–15] . Ethanolamine can serve as a carbon and/or nitrogen source for bacteria in the intestine as well as within epithelial cells [16 , 17] . The aim of this work was to determine whether S . Typhimurium relies on ethanolamine as a signal to coordinate gene expression and augment virulence in vivo . Here , we show that the intramacrophage environment promotes expression of the ethanolamine utilization transcription factor EutR , which directly activates SPI-2 . Moreover , we demonstrate that EutR signaling during systemic infection is specific to the intracellular environment and is important for robust S . Typhimurium dissemination . Altogether , our findings suggest that ethanolamine , an intrinsic component of bacterial and mammalian cell membranes , functions as a signal to modulate metabolism and virulence and suggest a new layer of complexity in chemical signaling that underlies pathogenicity .
Genes encoding for ethanolamine metabolism are clustered in the eut operon [18] ( Fig 1A ) . In the Enterobacteriaceae , expression of this operon is regulated by the eut-encoded transcription factor EutR . EutR is constitutively expressed at low levels from its own promoter and binds to the promoter region immediately upstream of eutS . In the presence of ethanolamine and vitamin B12 , EutR activates transcription of this operon [19 , 20] . In enterohemorrhagic Escherichia coli ( EHEC ) , EutR senses ethanolamine to activate virulence gene expression in vitro , independently of ethanolamine metabolism [19 , 21 , 22] . To determine whether EutR influences S . Typhimurium disease progression during infection , we generated an eutR deletion strain ( ΔeutR ) that cannot sense ethanolamine as well as an eutB deletion strain ( ΔeutB ) that lacks the large subunit of the ethanolamine ammonia lyase , and thus is unable to catabolize ethanolamine . The eutR and eutB mutations did not result in a general loss of fitness , as the ΔeutR and ΔeutB strains exhibited no measurable growth defects in vitro ( Fig 1B ) . Importantly , the eutB mutation is nonpolar as this mutant can respond to ethanolamine ( Fig 1C ) . Subsequently , we performed competitive infections in which streptomycin-treated mice were orally infected with an equal mixture of wild type ( WT ) and ΔeutB ( ΔeutB::CmR ) strains or the WT and ΔeutR ( ΔeutR::CmR ) strains . S . Typhimurium infection presents as intestinal outgrowth , invasion of epithelial cells , and subsequent uptake by macrophages and dissemination to secondary lymphoid tissue . Therefore , to monitor the course of S . Typhimurium infection , we analyzed the number of recovered bacteria from the intestinal contents , the colon , and the spleen . At 2 and 4 days post infection ( dpi ) , the ΔeutR and ΔeutB strains were significantly outcompeted by the WT strain in intestinal contents ( Fig 1D and 1E ) . These data underscore the importance of ethanolamine metabolism in S . Typhimurium colonization of the intestinal tract , and these findings are consistent with previous work by Thiennimitr et al . , who showed that ethanolamine metabolism provides a growth advantage to S . Typhimurium during intestinal colonization [17] . Defects in ethanolamine metabolism have been reported to result in mild or no attenuation during S . Typhimurium systemic infection [23 , 24]; however the role of EutR , specifically , in contributing to dissemination has not been investigated . Therefore , to assess this , we harvested the colons and the spleens of infected mice at 4 dpi . The ΔeutR and ΔeutB strains were both recovered at significantly lower numbers than WT from the colon and spleen; however , the competition indices measured from the spleen from the ΔeutR/WT infections were significantly greater than between WT and the ΔeutB strain ( P = 0 . 002 ) . These findings led us to hypothesize that EutR plays a more extensive role in S . Typhimurium pathogenesis that is distinct from its function to promote ethanolamine metabolism . To test this , we performed competition infections between the ΔeutB and ΔeutR ( ΔeutR::CmR ) strains . At 2 days post infection ( dpi ) , the ΔeutR and ΔeutB strains were recovered at similar numbers from intestinal contents ( Fig 1F ) , indicating that at this initial stage of colonization , EutR functions to drive ethanolamine metabolism . However , at 4 dpi , which is a time point consistent with the progression to systemic infection [25] , the ΔeutR strain was significantly outcompeted by the ΔeutB strain ( Fig 1F ) . Significantly , although equal numbers of the ΔeutR and ΔeutB strains were recovered from the colon , the ΔeutR strain was significantly outcompeted by the ΔeutB strain in the spleen ( Fig 1F ) . These data suggest that EutR , independent of its role in ethanolamine metabolism , is important to S . Typhimurium dissemination during infection . To further explore how ethanolamine signaling contributes to S . Typhimurium dissemination , we examined S . Typhimurium virulence gene expression in vitro . To examine ethanolamine-mediated expression of SPI-1 , we measured expression of sipC , a SPI-1 encoded translocase that plays a role in invasion of epithelial cells [26] . For this , we grew S . Typhimurium in LB , which induces SPI-1 expression [27] as well as in DMEM used for cell culture assays . In both cases , expression of sipC was slightly decreased when ethanolamine was included in the culture medium ( Fig 2A and 2B and S1 Fig ) . To determine whether the ethanolamine-dependent decrease in sipC expression impacted S . Typhimurium invasion of epithelial cells , we infected HeLa cells with either WT S . Typhimurium or the ΔeutR strain . WT S . Typhimurium and ΔeutR invaded HeLa cells at nearly equivalent levels in the presence or absence of ethanolamine ( Fig 2C and 2D ) . Because we did not observe EutR-dependent effects on epithelial invasion under the specified conditions , the influence of ethanolamine at this stage in S . Typhimurium dissemination was not pursued further . Next , we investigated whether ethanolamine impacted SPI-2 expression . SsrB is a SPI-2 encoded transcriptional regulator that is required for expression of all the SPI-2-encoded genes , as well as for expression of effectors and virulence genes encoded outside of SPI-2 [28–31] . To test the influence of ethanolamine , we measured expression of ssrB in low magnesium , minimal medium , a condition that induces SPI-2 expression [32] ( S2 Fig ) without supplementation or with supplementation of 250 μM or 5 mM ethanolamine . These concentrations were used because 250 μM ethanolamine was the lowest concentration with which we could readily detect EutR expression ( S3 Fig ) , whereas 5 mM is similar to ethanolamine concentrations in the gastrointestinal tract [33] . When 250 μM ethanolamine was added to the SPI-2 inducing medium , expression of ssrB was significantly increased compared to medium without supplementation , but unchanged when 5 mM ethanolamine was added ( Fig 3A ) . These data suggest that ethanolamine may enhance the response of S . Typhimurium in adapting to the intramacrophage environment . Expression of SPI-2 is tightly regulated and is induced specifically in the intracellular environment [34] , or in conditions that mimic the intracellular environment . In accordance , ssrB expression was not induced in DMEM or LB when ethanolamine was supplemented to the medium ( Fig 3B and S4 Fig ) , indicating that ethanolamine in and of itself does not override additional regulatory factors that direct ssrB expression . However , ssrB expression was decreased in DMEM with the addition of 5 mM ethanolamine ( Fig 3B ) . Altogether these data raised the possibility that ethanolamine signaling enhances niche recognition . Macrophages play a significant role in the pathogenesis of S . Typhimurium infection by providing protected sites for intracellular replication and a means of dissemination [35] . Robust expression of EutR requires ethanolamine as well as the cofactor vitamin B12 [20] . S . Typhimurium synthesizes vitamin B12 under anaerobic conditions [36]; however , S . Typhimurium must acquire ethanolamine from the environment [18] . Therefore , we investigated whether the intracellular environment induces eutR expression . For this , we infected macrophages with S . Typhimurium in the absence of any exogenous ethanolamine or vitamin B12 . Subsequently , RNA was extracted from internalized S . Typhimurium at 3 , 5 , and 7 h post phagocytosis , and eutR transcript levels were analyzed and compared to eutR transcript levels from S . Typhimurium grown in the absence of macrophages . Expression of eutR was significantly increased in phagocytized S . Typhimurium throughout infection compared to cells grown in the absence of macrophages ( Fig 3C and S5 Fig ) . Moreover , neither vitamin B12 or ethanolamine alone activated eutR expression in tissue culture medium , indicating that the intramacrophage environment is conducive to EutR-dependent signaling . The addition of ethanolamine and vitamin B12 to SPI-2 inducing medium or DMEM resulted in an increase in expression of the eut operon ( as indicated by eutS expression ( Fig 1A ) ) that corresponded with an increase in eutR expression ( S6 Fig ) . Notably , the eut operon was not induced within macrophages ( Fig 3D ) . These findings indicate that the ethanolamine metabolic genes are not activated within the intramacrophage environment . Interestingly , the expression pattern of eutR in internalized S . Typhimurium was similar to ssrB expression ( Fig 3E ) ; therefore , we hypothesized that EutR regulates SPI-2 expression . SPI-2 contains four major operons that encode a T3SS , chaperone and effector proteins , as well as the transcriptional regulator SsrB ( Fig 4A ) . To test our hypothesis , we examined transcription of ssrB and one gene from each of the other major operons encoded in SPI-2 using RNA harvested from phagocytized WT or ΔeutR S . Typhimurium strains . Transcription of ssrB was significantly decreased in the ΔeutR strain compared to WT ( Fig 4B and S7 Fig ) , and we measured a concomitant decrease in expression of all the SPI-2 operons , as well as the SPI-2-associated effector sifA ( Fig 4B and S8 Fig ) . Expression of SPI-2 encoded and associated factors enhances the intrinsic ability of S . Typhimurium to withstand and disrupt host defense mechanisms [37 , 38] , and these data revealed that EutR influences this critical aspect of S . Typhimurium virulence . SsrB is a response regulator that comprises a two component system with the sensor kinase SsrA ( also referred to as SpiR ) [13 , 15] . SsrA autophosphorylates in response to the acidic environment of the Salmonella-containing vacuole ( SCV ) within host cells [39 , 40] , which initiates a signaling cascade that promotes SsrB activity as well as ssrAB expression [41] . Importantly , the ssrB gene contains its own promoter [41] . The genetic data indicated that EutR influenced expression of ssrB and downstream targets , but that EutR did not impact ssrA expression ( Fig 4B and S8 Fig ) , indicating that EutR may regulate SPI-2 expression by binding the ssrB promoter . To examine this , we purified an EutR::MBP fusion protein . Electrophoretic mobility shift assays ( EMSAs ) indicated that EutR directly binds the ssrB promoter to activate expression of SPI-2 ( Fig 4C ) . To confirm specificity of binding , EMSAs with purified MBP alone as well as competitions assays with unlabeled probes were performed . MBP alone did not bind the ssrB promoter ( Fig 4D ) . Furthermore , EutR binding was outcompeted by the addition of unlabeled ssrB probe; however , the addition of unlabeled kan probe , as a negative control reaction , showed no competition ( Fig 4D ) . Consistent with these results , there was a significant enrichment of the ssrB promoter when EutR-DNA interactions were analyzed using in vivo chromatin immunoprecipitation followed by qPCR ( Fig 4E ) . As a positive control , we also measured enrichment of the eutS promoter , an established binding target of EutR [19] , and observed similar enrichment of both targets ( Fig 4E ) ; strB DNA was used as a negative control . Control of ssrB expression is complex and also includes activation by additional TCS PhoP/PhoQ and EnvZ/OmpR , which respond to signals within the Salmonella containing vesicle ( SCV ) [41 , 42] . Our findings suggest that EutR-dependent activation of ssrB enables S . Typhimurium to integrate intrinsic information regarding the host cell through ethanolamine signaling with SCV-specific signals to coordinate efficient spatiotemporal expression of SPI-2 and SPI-2 associated effectors . Next , we tested the consequences of EutR-dependent activation of SPI-2 on S . Typhimurium fitness during macrophage infection . Following infection of RAW or peritoneal exudate macrophages ( PEMs ) , the ΔeutR strain was recovered at significantly lower numbers compared to the WT strain ( Fig 5A–5C and S9 Fig ) . Additionally , complementation of the ΔeutR strain with eutR expressed from the native promoter ( eutR+ ) restored intracellular survival to WT levels during primary macrophage infection ( Fig 5B ) . To verify that the defect in the ΔeutR strain was not the result of a defect in ethanolamine metabolism , we assessed survival of the ΔeutB strain within PEMs . The ΔeutB strain was recovered at similar numbers to WT and at significantly higher numbers than the ΔeutR strain ( Fig 5C ) . Importantly , the ΔeutR mutant grows similarly to WT and ΔeutB strains in SPI-2 inducing medium and tissue culture medium with or without the addition of ethanolamine ( Fig 5D and 5E ) , confirming that the decrease in intracellular survival is not a result of a EutR-dependent growth defect . These findings indicate that ethanolamine-associated signaling , but not catabolism , impacts S . Typhimurium survival within macrophages . Moreover , these findings , in conjunction with lack of eut operon induction within macrophages ( Fig 3D ) , reveals that S . Typhimurium relies on EutR to direct gene expression in a manner that is particular to a specific niche . Next , we confirmed that EutR mediates dissemination specifically through intracellular survival in vivo . To further discriminate between ethanolamine-associated signaling and ethanolamine metabolism , we infected mice with equal numbers of the ΔeutR::CmR and ΔeutB strains by intraperitoneal injection . At 6 h pi , the ΔeutR strain was recovered at significantly lower numbers compared to the ΔeutB strain from the spleen ( Fig 6A ) . Furthermore , we assessed bacterial burden in the peritoneal cavity . At this site , there were no significant differences between the ΔeutR and ΔeutB strains in the total bacteria recovered ( Fig 6B ) , the majority of which were extracellular ( S10 Fig ) . However , the ΔeutR strain was recovered at significantly lower numbers compared to the ΔeutB strain in the phagocytized population of S . Typhimurium within the peritoneal cavity ( Fig 6C ) . These findings reveal that EutR augments S . Typhimurium fitness during systemic infection . Our findings differ from a previous study that used a genetic screen to identify genes important for systemic virulence [43] . Discrepancies may reflect differences in study design such as the age and genetic background of mice , route of infection , and/or duration of infection . Importantly , using in vitro and in vivo approaches , our data establish a genetic and functional role for EutR in S . Typhimurium systemic disease , and altogether , these results indicate that EutR contributes to the ability of S . Typhimurium to gauge and adapt to the intracellular environment in vivo . The in vitro studies identified targets of EutR-dependent gene regulation . To test our findings within the complexities of the in vivo environment , we assessed EutR-dependent regulation of ssrB using single strain infections and purified S . Typhimurium RNA from harvested spleens . Expression of ssrB was significantly decreased in the ΔeutR strain compared to WT ( Fig 7A and S11 Fig ) , which is consistent with the data presented in Fig 4B . Additionally , we measured expression of eutR and eutS in WT S . Typhimurium recovered from the spleen relative to S . Typhimurium grown in vitro . Notably , eutR expression was significantly increased in the spleen , whereas expression of eutS was not detectable ( Fig 7B ) . These data further highlight the dynamic role of EutR in S . Typhimurium pathogenesis from driving ethanolamine metabolism in the intestine to promoting virulence gene expression in later stages of disease . These findings reveal a novel signaling pathway critical for S . Typhimurium to enhance disease progression during infection . We propose a model in which S . Typhimurium relies on ethanolamine signaling through EutR to gauge distinct environments in the host and then modulate expression of genes encoding metabolism and virulence ( Fig 8 ) . The resident microbiota do not readily metabolize ethanolamine [33] . Thus , to establish infection , S . Typhimurium sidesteps nutritional competition by respiring ethanolamine in conjunction with tetrathionate , an electron acceptor generated specifically during intestinal inflammation [17 , 44] . Fermentation of ethanolamine provides very little growth [17]; hence , outside of the intestine , and in the absence of bacterial competition , S . Typhimurium preferentially utilizes alternative nutrients [23] . This enables EutR to direct expression of traits necessary for dissemination and systemic infection . Additional experiments are necessary to determine what factors influence the transition from driving metabolism to influencing virulence in the intestine . Genes encoding ethanolamine utilization are widespread in pathogenic bacteria as well as in members of the resident microbiota [45] , and the extracellular pathogen EHEC responds to ethanolamine to regulate virulence gene expression [22] . Therefore , ethanolamine signaling may be a conserved strategy used by diverse pathogens to coordinate metabolism and virulence in response to distinct host environments . Our findings highlight a sophisticated mechanism in which S . Typhimurium exploits an abundant and essential molecule within the host to gain specific information about the localized environment and modulate gene expression to overcome bacterial and host resistance mechanisms .
All strains and plasmids used in this study are listed in S1 Table . Luria-Bertani ( LB ) , Dulbecco’s Modified Eagle Medium ( DMEM ) ( Invitrogen ) , or minimal medium ( described below ) were used as indicated . Ethanolamine ( Sigma ) and/or vitamin B12 ( Sigma ) were supplemented to the media as indicated in the main text . Unless indicated otherwise , 150 nM vitamin B12 was added whenever ethanolamine was added to the growth medium . Antibiotics were used in the following concentrations: ampicillin ( 100μg/ml ) , streptomycin ( 100μg/ml ) , chloramphenicol ( 20μg/ml ) , and kanamycin ( 50μg/ml ) . Recombinant DNA and molecular biology techniques were performed as described previously [46] . S . Typhimurium SL1344 [47] and its derivatives were used in all experiments . The invG mutant ( strain AJK61 ) was a gift from James Casanova and was constructed as previously described [48] . Nonpolar eutR and eutB deletion strains were generated in WT and ΔinvG backgrounds using λ-red mutagenesis [49] . Briefly , PCR products ( obtained with primers listed in S2 Table ) were amplified from plasmid pKD3 or pKD4 with flanking regions matching eutR or eutB and then transformed into S . Typhimurium expressing the Red genes from plasmid pKD46 . The resistance cassette was resolved with flippase from temperature-sensitive plasmid pCP20 , which was then cured through growth at 42°C . Unresolved strains were used in the murine competition assays as indicated in the text and figure legends . All deletions were confirmed by sequencing . The eutR mutant was complemented with pCJA002 . Plasmid pCJA002 was constructed by amplifying S . Typhimurium genomic DNA using primers specific to the eutR gene , including 206 nucleotides upstream of the ATG start site ( listed in S2 Table ) . Amplified DNA was digested with HindIII and BamHI and inserted into pGEN-MCS [50] ( Addgene MTA ) . As controls , WT and the ΔeutR strains were transformed with empty pGEN-MCS vectors for use in the complementation experiments . The EutR::Flag strain was generated as described for the deletion strains , using pSUB11 as described [51] . All cultures were grown overnight in LB and then diluted 1:100 in the indicated medium and grown at 37°C . For RNA expression studies , cultures were grown until mid-logarithmic phase ( OD600 = 0 . 45–0 . 55 ) . Cultures grown in DMEM were incubated statically under a 5% CO2 atmosphere ( to mimic tissue culture conditions ) . SPI-2 inducing medium was prepared as previously described ( 100mM Bis/Tris-HCl pH 7 . 0 , 5mM KCl , 7 . 5mM ( NH4 ) 2SO4 , 0 . 5mM K2SO4 , 1mM KH2PO4 , 38mM glycerol , 0 . 1% casamino acids , and 8μM MgCl2 ) [32] , and cultures were grown aerobically with agitation [32] . For the in vitro studies , RNA was extracted from S . Typhimurium cells grown in culture medium as described or from phagocytized S . Typhimurium . Cells were resuspended in Trizol ( Life Technologies ) and RNA was purified using the RiboPure kit ( Ambion ) . For the in vivo studies , spleens were harvested at 2 dpi and homogenized in 1 mL Trizol per 100 mg tissue [52] . RNA was isolated using standard molecular biological procedures . Primer validation and qRT-PCR was performed as described previously [53] using primers listed in S2 Table . Briefly , RNA was extracted from three biological replicates , and qRT-PCR was performed in a one-step reaction using an ABI 7500 sequence detection system ( Applied Biosystems ) . Data were collected using the ABI software Detection 1 . 2 software ( Applied Biosystems ) . All data were normalized to the endogenous control strB ( main text ) or to 16S rRNA ( RNA was diluted 1:1000 ) as previously performed [54 , 55] . Controls were used as indicated in figure legends and analyzed using the comparative critical threshold ( CT ) method . The Student’s unpaired t test was used to determine statistical significance . RAW , J774 , and HeLa cells were routinely cultured in DMEM supplemented with 10% FBS and 1x penicillin-streptomycin–glutamine; peritoneal exudate macrophages ( PEMs ) were cultured in RPMI 1640 supplemented with 10% FBS , 20% L-929 conditioned medium , and 1x penicillin-streptomycin–glutamine . PEMs were isolated as described [56] . Antibiotics were omitted during bacterial infections . For epithelial cell infection bacterial cultures were grown under invasion-inducing conditions [27] . Briefly , overnight cultures were diluted back 1:100 and grown without agitation in LB until late logarithmic phase ( OD600 of approximately 1 . 0 ) at 37°C . Bacterial cells were washed and resuspended in 1x phosphate buffered saline before infection . HeLa cells were placed in DMEM or DMEM supplemented with 5 mM ethanolamine and 150 nM vitamin B12 . HeLa cells were infected at a multiplicity of infection ( MOI ) of 100 for 1 h and either lysed directly or treated with 100 μg/ml gentamicin for 30 min to kill any extracellular bacteria . Percent invasion was calculated as the number of intracellular bacteria as a percent of the directly lysed sample and normalized such that wild type was equal to 100% . For macrophage assays , we used an invG mutant ( deficient in cell invasion ) as the WT strain , and we generated corresponding ΔeutRΔinvG and ΔeutBΔinvG strains ( described above ) . These strains were used because invasive S . Typhimurium rapidly kills macrophages [57] . Additionally , expression of invasion-associated genes are down-regulated after entry into host cells; therefore , this strain more closely mimics S . Typhimurium as it is encountered by professional phagocytes after penetration of the epithelial barrier [58] . Gentamicin protection assays were performed according to published methods [14 , 57 , 59 , 60] . S . Typhimurium was grown overnight in LB , washed and re-suspended in PBS before incubation with macrophages ( without the addition of ethanolamine or vitamin B12 ) at an MOI of 50 . After 30 min of incubation , extracellular bacteria were killed with 100 μg/ml gentamicin treatment for 30 min , before replacement with media containing 10 μg/ml gentamicin for the remainder of the assay . Cells were lysed at indicated time points in 1% Triton-X and colony forming units ( cfu ) determined by serial dilutions and plating onto LB agar . After internalization , cells were treated with gentamicin and lysed to enumerate viable intracellular bacteria at time 0 h . Survival was calculated as previously described [14] . Briefly , the viable cfu at the indicated time points were determined as the percentage of this intracellular time 0 h population and normalized such that wild type was equal to 100% . For all assays , the Student t test was used to determine statistical significance . All experiments were approved by the Institutional Animal Care and Use Committee at the University of Virginia School of Medicine . For the colitis infections , female C57BL/6 ( 10–12 week old ) mice were given a single dose of 20 mg streptomycin 24 h prior to infection [61] . Mice were infected with an equal mixture of 5 x 108 cfu of the indicated strains . Fresh fecal pellets were collected daily , and mice were euthanized at 4 dpi to assess bacterial burden in the colon and spleen . Tissue samples were weighed , homogenized in 1 ml PBS , and bacterial numbers were quantified by plating serial dilutions of homogenates on MacConkey agar supplemented with streptomycin or with chloramphenicol . The competitive index was calculated as the ratio of ΔeutR to wild type ( WT ) or ΔeutB strains or the ratio of the ΔeutB to WT recovered normalized to the ratio in the inoculum . Statistical significance was determined by one-sample t test with an expected value of 1 . Comparisons between competitive indexes were performed using the Mann-Whitney U test . For the systemic competition experiments , mice were infected intraperitoneally ( i . p . ) with 1x105 cfu of the ΔeutR and ΔeutB strains . Spleens were harvested at 6 h and bacterial burden was assessed as described above . Bacterial burden in the peritoneal cavity was assessed as described [62] . Briefly , following euthanasia , 5mL of PBS was injected into the peritoneal cavity , aspirated , and immediately placed on ice . Samples were split into two aliquots , one receiving 100 μg/ml gentamicin treatment . After 30 minutes on ice , samples were washed , lysed and plated as described above . For RNA analyses , mice were infected by i . p . with 1x104 cfu of WT or the ΔeutR strain , and spleens were harvested at 2 dpi . Plasmid pDC24 or the empty vector pMAL-c5X was used for the EMSA assays . This plasmid was constructed by amplifying the eutR gene from the S . Typhimurium strain SL1344 with indicated primers ( S2 Table ) . The resulting PCR product was cloned into the Nco1/Sbf1 cloning site of vector pMAL-c5X . EutR was purified under native conditions as described [19] . Briefly , the MBP-tagged EutR protein was purified by growing the E . coli strain NEBexpress cells ( NEB ) containing pDC24 at 37°C in LB with glucose ( 0 . 2% final concentration ) and ampicillin ( 100 μg/ml ) to an OD600 of 0 . 5 , at which point IPTG was added to a final concentration of 0 . 3 mM and allowed to induce overnight at 18°C . Cells were harvested by centrifugation at 4000 x g for 20 min and then resuspended in 25 mL column buffer ( 20 mM Tris-HCl; 200 mM NaCl; 1 mM EDTA ) and lysed by homogenization . The lysed cells were centrifuged , and the lysate was loaded onto a gravity column ( Qiagen ) with amylose resin . The column was washed with column buffer and then eluted with column buffer containing 10 mM maltose . Fractions containing purified proteins were confirmed by SDS-PAGE and Western analysis , and the protein concentration was determined using a NanoDrop Spectrophotometer . PCR-amplified DNA probes ( listed in text and described in S2 Table ) were generated as previously described [19 , 63] . DNA probes were end-labeled with [γ-32P]-ATP ( Perkin-Elmer ) using T4 polynucleotide kinase ( NEB ) following standard procedures [64] . End-labeled fragments were run on a 6% polyacrylamide gel , excised , and purified using the Qiagen PCR purification kit . EMSAs were performed by adding purified EutR-MBP or MBP to labeled DNA in binding buffer ( 500 μg ml-1 BSA ( NEB ) , 50 ng poly-dIdC , 60 mM HEPES pH 7 . 5 , 5 mM EDTA , 3 mM dithiothreitol ( DTT ) , 300 mM KCl , and 25 mM MgCl2 ) . Ethanolamine ( 1 mM ) and vitamin nM B12 ( 150 nM ) were added to the reactions . Reactions were incubated for 25 minutes at 25°C . Then , a 1% Ficoll solution was added to the reactions immediately before loading the samples on the gel . The reactions were electrophoresed for approximately 6 h at 150 V on a 6% polyacrylamide gel , dried , and imaged with a phosphorimager ( Molecular Dynamics ) . ChIP was performed using an WT S . Typhimurium ( untagged EutR ) or with the S . Typhimurium eutR mutant transformed with the EutR::MBP plasmid . Strains were grown in DMEM supplemented with 5 mM EA , 150 nM B12 , and 0 . 5 μM IPTG until cells reached an OD600 of approximately 0 . 8 . Cross-linking and ChIP were performed based on established methods [65] . Formaldehyde was added ( 1% final concentration ) for cross-linking , and cells were incubated at room temperature for 20 min . Reactions were quenched with 0 . 5 M glycine , then samples were pelleted , resuspended in TBS , and washed . Cells were lysed with 2 mg/ml lysozyme and incubated at 37°C for 30 min . Subsequently , samples were placed on ice and sonicated . Insoluble cell debris was removed by centrifugation , and supernatants were saved . Immunoprecipitation was carried out by incubating samples with amylose beads ( NEB ) in buffer for 2 h at 4°C with gentle mixing . Beads were pelleted and washed . Then the samples were incubated for 10 min at 65°C in elution buffer with occasional gentle mixing . Samples were centrifuged and supernatants were collected . To reverse the cross-link , samples were boiled for 10 min and DNA was purified using the Qiagen PCR purification kit . For ChIP-quantitative PCR ( qPCR ) experiments , untreated chromatin was de-cross-linked by boiling for 10 min and purified , for use as the “input” control . Primers amplifying the strB gene were used as the negative control . The fold enrichment of each promoter represents the value of the immunoprecipitated DNA divided by the input unprecipitated DNA [66 , 67] . These values were normalized to the values obtained for each promoter precipitated using untagged EutR in order to account for non-specific enrichment . eutR , NP_461389; eutS , NP_461405; eutB , NP_461393; sipC , NP_461805; ssrA , NP_460357; ssrB , NP_460356; ssaB , NP_460358; sseA , NP_460362; ssaG , NP_460371; sifA , NP_460194 .
|
Chemical signaling underlies all cellular processes . Bacteria rely on chemical signaling to gain information about the local environment and precisely regulate gene expression . Ethanolamine is an abundant molecule within mammalian hosts that plays an important role in mammalian physiology and also serves as a carbon and nitrogen source for bacteria . Here we show that the foodborne pathogen Salmonella enterica exploits ethanolamine as a signal of distinct host environments to coordinate metabolism and virulence , which enhances disease progression during infection . The ability to sense ethanolamine is conserved in diverse bacteria; thus , these studies reveal that ethanolamine signaling may be important for bacterial adaptation to the mammalian host .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Ethanolamine Signaling Promotes Salmonella Niche Recognition and Adaptation during Infection
|
International seaports are hotspots for disease invasion and pathogens can persist in seaports even after ports are abandoned . Transmitted by fleas infected by Rickettsia typhi , murine typhus , a largely neglected and easily misdiagnosed disease , is known to occur primarily in large seaports . However , the significance of seaports in the occurrence of murine typhus has never been validated quantitatively . We studied the spatial distribution of murine typhus , a notifiable disease , in Taiwan . We investigated whether risk of infection was correlated with distance to international seaports and a collection of environmental and socioeconomic factors , using a Bayesian negative binomial conditionally autoregressive model , followed with geographically weighted regression . Seaports that are currently in use and those that operated in the 19th century for trade with China , but were later abandoned due to siltation were analyzed . A total of 476 human cases of murine typhus were reported during 2000–2014 in the main island of Taiwan , with spatial clustering in districts in southwest and central-west Taiwan . A higher incidence rate ( case/population ) was associated with a smaller distance to currently in-use international seaports and lower rainfall and temperature , but was uncorrelated with distance to abandoned ports . Geographically weighted regression revealed a geographic heterogeneity in the importance of distance to in-use seaports near the four international seaports of Taiwan . Our study suggests that murine typhus is associated with international seaports , especially for those with large trading volume . Thus , one of the costs of global trade in Taiwan might be elevated risks of murine typhus . Globalization has accelerated the spread of infectious diseases , but the burden of disease varies geographically , with regions surrounding major international seaports warranting particular surveillance .
Trade is commonly accompanied by the spread of infectious diseases and international seaports have long been hotspots for disease invasion [1] . The great expansion in trade and international networks in recent history has seen seaports increasingly receive imported pathogens and vectors [2 , 3] . For example , yellow fever has devastated seaports in the Americas due to the importation of the virus-infected mosquito Aedes aegypti ( a competent vector for yellow fever ) by ships [4] . Another new disease vector originating in Asia , Aedes albopictus , has also spread to seaports in both the Old and New Worlds [5 , 6] . Successful introduction of exotic diseases involve arrival , establishment of local transmission , and subsequent spatial dispersal [7] . In suitable environments , exotic pathogens or parasites can persist in invaded regions even though these pathogens or parasites have ceased to arrive at the seaports . For instance , plague introduced to the USA through San Francisco in 1899–1900 still circulates among prairie dogs in the deserts of the Southwestern United Sates [8 , 9] despite the absence of current importations . Likewise , helminths introduced by exotic rats have spread to indigenous mice on the California Channel Islands , with transmission persisting even after eradication of the rat hosts [10] . The probability of ongoing transmission following introduction to a new area is dependent on habitat suitability: for example , the availability of host species and/or vectors which may , in turn , be influenced by environmental conditions [11] . Thus , one legacy of past shipping events might be continuing circulation of exotic pathogens near receptive seaports; that is , although seaports have ceased to function , imported pathogens may persist in proximity to the abandoned seaports , if the conditions are suitable . Murine typhus is a rickettsial disease with a worldwide distribution , but its significance as a common causative agent of illness in tropical regions remains largely neglected [12] . Transmitted by fleas infected with Rickettsia typhi , people typically acquire murine typhus via contaminated flea faeces near the bite sites instead of directly from the flea bites [13] . The life cycle of R . typhi commonly involves the oriental rat flea Xenopsylla cheopis and commensal rats , particularly Rattus rattus and Rattus norvegicus [14] . However , in suburban Southern California and Southern Texas , R . typhi is instead maintained by the cat flea Ctenocephalides felis , the opossum Didelphis marsupialis and domestic cats [15–18] , and in Spain , dogs were found to host R . typhi [19] . It is well recognized that murine typhus is prevalent primarily in large seaports , probably due to the repeated introduction of infective fleas and rats [20] . Nevertheless , the significance of seaports in the occurrence of murine typhus has never been validated quantitatively . Likewise , while incidence of murine typhus is associated with the abundance of fleas , which is affected by climatic factors such as temperature , precipitation and humidity [20 , 21] , spatial analysis of the relationship between murine typhus and environmental variables remains very rare . The spatial distribution of murine typhus has been investigated in Lao PDR to confirm whether murine typhus is more common in urban areas , but only socio-economic risk factors have been included in the study [22] . Spatial clustering of murine typhus was also studied in Texas , but focusing on a comparison of clustering detection methods [23] instead of environmental correlates . In Taiwan , murine typhus is an endemic disease , with 13 to 44 human cases annually during 2005–2014 ( Taiwan Centers for Disease Control ( CDC ) ; http://nidss . cdc . gov . tw/ ) . The spatial pattern of murine typhus occurrence and the reasons for geographic heterogeneity have never been explored in Taiwan; instead , past studies have focused on clinical manifestations of the disease [24–28] . We conducted a retrospective investigation of the spatial distribution of murine typhus in Taiwan and explored its association with environmental and socioeconomic factors . Notably , we sought to determine whether murine typhus incidence was higher in areas closer to international seaports . Seaports that are currently in use and abandoned seaports were analyzed to identify the public health consequences of historical international trade . Occurrence of murine typhus could also be related to the presence of cats , dogs and cat fleas , as recently found in Spain and the United States of America [16 , 19] . However , the lack of information on the number of cats and dogs ( particularly stray ones ) and the spatial distribution of cat fleas in Taiwan hindered incorporation of this non-classic infection route in this research . The current study therefore focused on the classic rat-flea transmission cycle , which remains the primary route of infection all over the world [15] .
The case records were retrieved from the Taiwan National Infectious Disease Statistics System administrated by Taiwan Centers for Disease Control ( Taiwan CDC ) and no personally identifiable information were used as part of this study . This study focused on the main island of Taiwan . Small associated islets were excluded ( Kin-men , Ma-tou , Peng-hu , Little Liu-chiu , Ci-jin , Green , and Orchard islands ) because they frequently differ with regard to potentially important ecological characteristics ( e . g . , animal communities [29] ) . The basic geographical units used in this analysis were administrative districts ( within urban cities ) and townships ( within rural counties ) ; these are the smallest administrative areas to which murine typhus cases can be assigned . In this study , we use “district” to refer to both the urban districts and the rural townships . Human incidence of murine typhus from 2000 to 2014 was retrospectively analyzed in this study . Murine typhus is a notifiable disease in Taiwan . Blood samples from patients with suspected murine typhus are collected and sent to the Taiwan CDC for laboratory diagnosis . Samples were considered positive for murine typhus based on a positive real-time polymerase chain reaction ( PCR ) test or the detection of R . typhi-specific antibodies based on the indirect immunofluorescent assay ( IFA ) . The real-time PCR test targeted the 17-kDa antigen in Rickettsia spp . and the PCR products were sequenced and then assessed with the Basic Local Alignment Search Tool ( www . ncbi . nlm . nih . gov ) for resemblance to known Rickettsia spp . For IFA , each serum sample was applied to slides coated with R . typhi antigens ( Focus Technologies , Inc . , Cypress , CA , U . S . A . ) . Two IFA criteria were applied: ( 1 ) four-fold increase in R . typhi-specific immunoglobulin M ( IgM ) or IgG antibody in paired sera ( each for the acute and convalescent phase , with interval >14 days ) ; ( 2 ) positive for patient with IgM 1:80 dilution and IgG 1:320 dilution . Because infection may occur away from a patient’s residence , starting in 2003 , the presumptive location of infection was recorded as well as the patient’s residence . These data , along with gender , age , and date of symptom onset , are available from the Taiwan CDC . To more accurately assess the relationship between infection and environmental factors , we allocated cases of murine typhus ( 2003–2014 ) to the presumed district in which the infection occurred rather than the district in which the patient resided . For incidences during 2000–2002 , patient’s residence was used instead . The presumed district of infection and district of residence were the same for 97 . 1% of cases from 2003 to 2014 , so the use of patient’s residence from 2000 to 2002 is not considered problematic . Because yearly variation ( 2000–2014 ) in district population size was low ( 3 . 7% , average value of ( standard deviation divided by mean ) for all districts ) , population size for each district was represented by the mean value from 2000–2014 . Population size was obtained from the Department of Statistics of the Taiwan Ministry of the Interior ( http://sowf . moi . gov . tw/stat/month/list . htm ) , and the murine typhus incidence rate ( IR , number of cases per 100 , 000 people per year ) was calculated for inter-district comparisons . The presence of spatial autocorrelation of the murine typhus IR ( incidence rate ) was assessed using Moran’s I [30] . The locations of spatial clusters of murine typhus incidence were identified using local indicators of spatial association ( LISAs ) . LISAs can be treated as a local version of Moran’s I [30] , and can be used to detect local clusters of observations with similar or dissimilar values [31] . A map of LISAs clusters , thus , allowed the assignment of each district to one of five categories: high-high , which indicates a district with high IR surrounded by districts with high IR ( also called a hot spot ) ; low-low , a district with low IR surrounded by low-IR districts ( a cold spot ) ; low-high or high-low , a district with low IR surrounded by high-IR neighbors and vice versa; and not significant , which indicates a district with no significant local autocorrelation [32] . Inference for significance of Moran’s I and LISAs was based on 99 , 999 permutations using the GeoDa 0 . 9 . 5 software [33] , and empirical Bayes ( EB ) was applied to correct for large variation in population size among districts [32] , with population size as the base variable . The threshold of significance was set at P = 0 . 05 , and maps were displayed using QGIS 2 . 12 ( QGIS Development Team ) . We selected variables for analysis based on the availability of data and our knowledge of the study system . Twelve explanatory variables ( seven environmental variables , two socioeconomic variables , and three port distance variables ) were included in the study . Environmental variables included elevation ( elevation , meters ) , total annual rainfall ( rainfall , mm ) , mean annual temperature ( temperature , °C; calculated as the mean of 12 monthly mean temperatures ) , number of days with temperature higher than 30°C within a year ( daysT30 ) , relative humidity ( % ) and a selected list of land cover categories . Elevation was derived from a 40-m digital elevation model ( Aerial Survey Office of Taiwan Forestry Bureau ) . The four meteorological variables were obtained from Central Weather Bureau of Taiwan ( n = 390 meteorological stations , the Data Bank for Atmospheric Research is available at https://dbahr . narlabs . org . tw/ ) and were calculated over the period 1991 to 2013 . The spatial layers of climatic variables were generated at a spatial resolution of 1 km by interpolation ( 390 stations ) using Kriging in ArcGIS with a spherical variogram model [34] . We overlaid administrative district boundaries and calculated the mean values for elevation , rainfall , temperature , days over 30°C and relative humidity for each district . Land cover data were obtained from the Globcover database [35] using a spatial resolution of 30 arc seconds ( ca . 1 km ) and the initial land cover classes were merged to create a smaller number of land cover types likely to be important for R . typhi transmission . These include artificial structure and forests ( artificial surface and forest ) because human infection of R . typhi occurs mainly inside buildings [20] and we were interested in the potentially protective effects of forests . The proportion of each district that consisted of each of these land cover classes was calculated to provide a quantitative characterisation of the land cover . To assess the role of socioeconomic factors , average income ( income ) of each district for the year 2005 was obtained from the Fiscal Information Agency of the Taiwan Ministry of Finance ( http://www . fia . gov . tw/ ) . Population density for each district was obtained by dividing population size by the respective administrative area . In this study , distance to three different types of international seaports were analyzed for comparison with R . typhi infection: ( 1 ) currently in use ( n = 4 ) ; ( 2 ) operated mainly during the 19th century for the trade of commodities with mainland China , where murine typhus has long been prevalent along the coast [36] , but were largely abandoned later because of siltation ( n = 26 ) ; and ( 3 ) including both in-use and abandoned international seaports ( n = 28 ) . Two seaports which were operational during the 19th century remain in operation today , and so are included in all three of these categories . International seaports that are currently operated include Keelung , Taichung , Kaohsiung , and Hualien seaports ( Fig 1 ) . Keelung and Kaohsiung seaports have been in use since the 19th century while Taichung and Hualien seaports have been operated since the 1970s . 19th century Taiwanese seaports have been classified into ten categories based on the volume of seaborne goods handled [37] . Some of the ports with the largest amount of cargo handled were deemed international ports in this study , because each had direct marine traffic with mainland China [37] . These 26 international ports are mostly located along the coast although some are situated along rivers ( Fig 1 ) and only two of them ( Keelung and Kaohsiung ) continue to engage in international trade . Distance to international seaports was represented by the Euclidean distance from the geographical centroid of each district to the nearest ports . Overlay of the district boundaries on grids of environmental variables and the calculation of the nearest distance to ports were implemented in ArcGIS 10 . 2 . Correlation analysis was applied to assess the strength and direction of the association amongst the explanatory variables . Where variables were highly correlated with one another , only one of the variables was retained for subsequent non-spatial multivariate regression analysis to avoid multi-collinearity . Lastly , significant variables in the final non-spatial multivariate model were analyzed separately with a Bayesian spatial regression model and geographically weighted regression .
A total of 476 human cases of murine typhus were recorded during 2000–2014 , with an incidence rate of 0 . 14 cases per 100 , 000 residents per year; this was higher in males than in females ( 0 . 20 vs . 0 . 08; Chi-square test with Yates’ correction , χ2 = 85 . 0 , P < 0 . 001 ) . The incidence rate also varied with age ( χ2 = 167 . 1 , P < 0 . 001 ) and was higher in the 50–79 age range ( Fig 2A ) . There was also a significant seasonal variation in incidence rate ( χ2 = 114 . 6 , P < 0 . 001 ) , with rates higher in later spring and summer than in other seasons ( Fig 2B ) . Among the 349 districts , the number of cases of murine typhus during 2000–2014 ranged from zero to 16 cases , with more cases occurring in southwest and central-west Taiwan ( Fig 3A ) . The IR varied from zero to 3 . 1 cases per 100 , 000 residents per year and was higher in southwest and central-west Taiwan , along with central Taiwan ( Fig 3B ) . Incidence of murine typhus was not randomly distributed in Taiwan ( Moran’s I = 0 . 35 , P<0 . 0001 ) . Instead , the LISA map revealed that hot spots were present in southwest and central-west Taiwan while cold spots occurred in eastern Taiwan ( Fig 4 ) .
This research has examined the spatial distribution of murine typhus in Taiwan , and possible explanatory factors for this distribution , for the first time . We found spatial clustering of human cases of murine typhus in southwest and central-west Taiwan . The risk of infection was higher in areas closer to international seaports that are currently in use , particularly near Kaohsiung and Taichung seaports . However , the probability of infection was not significantly associated with distance to abandoned international seaports . Risk of murine typhus was also negatively associated with rainfall and temperature , after controlling for distance to in-use international seaports . It has been stated that ports are the primary foci of murine typhus transmission [20] . Nevertheless , to the best of our knowledge , this is the first study to quantitatively validate a negative association between risks of R . typhi infection and distance to seaports , based on an advanced spatial modeling approach . Higher risks of infection near ( active ) seaports suggest that these may be the source of infection , as a consequence of repeated introduction of infective rats and/or fleas from abroad in combination with the mild weather typically enjoyed by coastal cities that is also hospitable for rats and fleas [20] . In spite of the negative association between disease incidence and distance to operating international seaports , the IR of murine typhus and the importance of distance varies considerably among the four ports . Distinctly , negative association between IR of murine typhus and distance to operating seaports occurs primarily near the Kaohsiung and Taichung seaports ( Fig 6C ) . There have been no cases surrounding the Hualien seaport in eastern Taiwan and there are very few cases along the eastern coast of Taiwan ( Fig 3 ) , in stark contrast with the high prevalence along the western coast , particularly near the Kaohsiung and Taichung seaports . This is consistent with the finding of a higher seropositivity rate of R . typhi infection in shrews and rodents trapped in Kaohsiung seaport ( 26 . 1% ) and Taichung seaport ( 18 . 1% ) than the other eight seaports or airports ( including Keelung seaport of 0 . 7% and Hualien seaport of 1 . 7% , [44] ) . Such geographical variation could be due to the remarkable difference in trading volume among the four ports , with Kaohsiung dealing with the lion’s share of international cargo ( an annual mean of 112 million tons during 2011–2013 ) , followed by Taichung ( 60 million tons ) , Keelung ( 18 million tons ) , and Hualien ( 4 million tons ) ( Taiwan International Ports Corporation; http://www . twport . com . tw/ ) . The lack of cases in proximity to Hualien might be related to the smaller cargo volumes providing fewer opportunities for pathogen introduction although this could also be related to higher temperature and rainfall near this port ( Supporting information S1 Fig ) so that pathogen transmission cannot be easily sustained after being imported . It was also found that the spatial distribution of murine typhus differed from that of scrub typhus , another rickettsial disease transmitted by mites . In Lao PDR , murine typhus was more common in urban areas while scrub typhus was more common in rural regions [22] . This contrasting spatial distribution also occurs in Taiwan , where scrub typhus is much more prevalent in less developed eastern areas than industrialized western areas of Taiwan [45 , 46] . Although R . typhi was not detected in fleas in eastern Taiwan [47] , rickettsial strains similar to R . typhi have been detected in ticks and rodents in the same region [48 , 49] , indicating that murine typhus might also circulate in this region but may be overlooked by physicians . This could be due to low prevalence as revealed by the low seropositivity rate of R . typhi infection in shrews and rodents in Hualien seaport ( 1 . 7% , [44] ) . Our study suggests that murine typhus should be considered as a possible diagnosis when patients close to the Hualien seaport present with suspected rickettsial infections . Indeed , clinical manifestations of many rickettsial diseases ( e . g . high fever , headache , rash ) are so similar that identification of the etiologic agent is very challenging , especially in the tropics [50] . Under-reporting is thus likely to be common , particularly in Hua-lien , where the other rickettsial disease ( scrub typhus ) is very prevalent [45 , 46] and murine typhus might be readily excluded . Although it is expected that poorer hygiene in the 19th century vessels might render rats infested with fleas more likely to board ships and invade ports , we did not find quantitative evidence supporting higher risks of infections near ports that operated in the 19th century , but which have subsequently been abandoned . This suggests that local conditions might not be suitable for long-term sustained transmission , and as these ports have largely been abandoned since the late 19th century , there was little opportunity for recent introduction at these locations . Because more people work in operational than abandoned seaports , more food might be available to sustain a higher population of rats and fleas in operational than abandoned ports . However , whether the lower risks were the result of lower survival of rats , fleas or R . typhi in abandoned ports remains to be investigated due to a lack of systematic studies on these ports . Another possibility is that the transmission cycle is sustained to the current date in so few abandoned ports that the significance of abandoned ports cannot be established statistically . In other words , contemporary infection might continue in a few abandoned ports since the 19th century , but because infection has ceased in most abandoned ports , we were unable to recognize its significance when all obsolete ports are considered in spatial analysis . For example , while the two hotspots in southwest and central-west Taiwan are also close to abandoned seaports ( Fig 4 ) , the majority of obsolete ports have very low incidence; although a similar spatial pattern could also arise where there are so many abandoned ports that hotspots coincidently occur close to a few of them . It is very difficult to unpick the significance of in-use versus abandoned seaports although our results suggest that some in-use seaports are of more importance for contemporary murine typhus incidence than abandoned ports . One potential solution to this issue is to investigate the population genetic structure of R . typhi in Taiwan . Cargos moving through abandoned and operating seaports came from different locations: abandoned seaports are likely to have dealt with cargo mainly from coastal China , while in-use seaports are likely to deal with cargo mainly from other countries . Therefore , the genetic composition of R . typhi should differ based on origin prior to introduction to Taiwan . This information would allow a more comprehensive assessment of the importance of abandoned seaports in the contemporary spatial distribution of infection . Studying the population genetic structure would also help discern whether R . typhi is mostly imported ( i . e . genetic composition differs among international seaports ) or is spread from within Taiwan ( i . e . no spatial structure in genetic composition is observed ) . The current status of rats and fleas ( species and abundance ) in operational and abandoned ports could also be better understood when trapping rodents to investigate the genetic structure of R . typhi in fleas; this could help reveal how the non-sustained transmission of R . typhi in abandoned ports could be related to the survival of rats or fleas . Another limitation of the current study is that due to the lack of data on trading volume at abandoned seaports [37] , the probability of importation of R . typhi at each seaport is considered identical , even though the volume of trade varies considerably among ports . Anping , Lukang and Wanhua were regarded as the largest ports in the 19th century in Taiwan , but there was no evidence of spatial clustering around these obsolete ports ( Fig 4 ) , suggesting that historical trading volume might not be the primary determinant of contemporary murine typhus infection risks . Lastly , whereas serological assay is the primary method for diagnosis of murine typhus [51] , cross-reactivity can occur in human sera against R . typhi and R . felis antigens [52–55] , although it is unclear why similar cross-reactivity does not always occur [e . g . 56 , 57] . Potential serologic cross-reactivity suggests that confirmed cases of murine typhus based solely on IFA diagnosis may include some misdiagnosed R . felis infections ( also a flea-borne rickettsial disease ) . In Taiwan , molecular methods have detected R . felis or R . felis-like organisms in one patient [58] as well as in small mammals [49] and fleas [47 , 59 , 60] . Therefore , we cannot exclude the possibility that murine typhus cases confirmed by the Taiwan CDC may include some cases of R . felis and the case data might more accurately reflect infections of flea-borne rickettsial diseases ( caused by R . typhi or R . felis ) instead of infectious associated with R . typhi only . In Taiwan , however , sera of confirmed cases of murine typhus were not found to cross-react with R . felis antigen [61] and serum from the single patient detected with R . felis nucleotides did not cross-react with R . typhi antigen [58] . Based on this , the extent of R . felis infections in patients diagnosed with murine typhus is presumed to be minimal in Taiwan , but this warrants further investigation . Also awaiting validation is the importance of R . felis as the causative agent of human illness , which is recently questioned for its widespread distribution in cat fleas but few and spatially restricted human cases of flea-borne rickettsioses in California [62 , 63] and its ubiquity in a diverse array of arthropods and also in healthy people in Africa [51] . Whether R . felis is simply a symbiont of arthropods similar to Wolbachia [64] would therefore determine if infection of R . felis is required to be considered in epidemiological studies of murine typhus . Across Taiwan , rainfall and temperature were also significantly associated with murine typhus incidence , after controlling for the influence of distance to operating international seaports . Given the same distance to seaports , murine typhus incidence decreased with increasing rainfall and temperature . This suggests that after R . typhi was introduced at the ports , the probability of further inland invasion of rats , fleas or R . typhi may have been determined by local climates . It was found that in eastern Taiwan , fleas were more abundant in months with less rainfall and lower temperature , but the underlying mechanism still awaits investigation [47] . In fact , climatic effects on flea-borne diseases are complex and context dependent . For example , similar to murine typhus , transmission of plague also involves bacteria , fleas , and rodent hosts . While it is generally thought that fleas prosper under hot and humid conditions , Yersinia pestis , the etiologic agent of plague , can persist in arid regions ( such as Central Asia and Western USA ) , but is less likely to sustain transmission in humid tropical areas [21] . Likewise , fleas were predominantly collected in dry rather than humid regions . However , in the dry part of Reunion Island , fleas were more abundant during the hot-wet season [65] . This is akin to our finding that geographically , murine typhus tended to occur in cooler and drier areas , but seasonally murine typhus was more prevalent in the warmer season ( late spring and summer ) . Apart from fleas , climate could also influence the abundance of rodents and human behavior [21] , both of which could affect the infection risk of murine typhus . The precise relationship between rainfall , temperature and murine typhus incidence in Taiwan ( given the same distance to seaports ) is , therefore , complex and necessitates great care to disentangle it . This is further supported by the geographical heterogeneity in the importance and direction of relationship for temperature and rainfall ( Fig 6A and 6B ) . Moreover , it should be emphasized that while an association between infection risk of murine typhus and climatic variables is identified , such correlation does not definitely represent that climate does determine the risk of infection; other not-recognized variables correlated with rainfall and temperature might instead be the main determinant . Our study demonstrates that one of the costs of international trade in Taiwan might be an elevated risk of murine typhus . This can be exemplified by Kaohsiung seaport , whose container traffic ranks 13th globally ( World Shipping Council; www . worldshipping . org ) . Kaohsiung is not only a hotspot for murine typhus ( this study ) ; dengue and scrub typhus , both vector-borne diseases , are also common in this port city [45 , 66] . To prevent potential importation of exotic diseases , regulations mandated by Taiwan CDC that govern quarantine at international ports require all arriving ships to report occurrence of rodents and disease vectors on the ships . Small mammals , fleas and seroprevalence of R . typhi in rodents are also monitored constantly and the eradication of rats has been attempted in these four international seaports by Taiwan CDC [44] . Globalization has hastened the spread of infectious diseases [67 , 68] , but the burden of diseases varies geographically , and as this study has shown , regions surrounding international seaports should warrant particular surveillance . Also needed is the assessment of whether eradication programmes implemented in seaports do indeed mitigate the risks of targeted diseases .
|
Globalization has hastened the spread of infectious diseases , with seaports as hotspots for disease invasion . Transmitted by fleas infected with the rickettsia Rickettsia typhi , murine typhus occurs worldwide , but its significance as a common causative agent of illness in tropical regions remains largely neglected . Although it is recognized that murine typhus is prevalent primarily in large seaports , the significance of seaports in the occurrence of murine typhus has never been validated quantitatively . We thus investigated whether distribution of murine typhus in Taiwan was associated with international seaports . Notably , abandoned international seaports ( abandoned in the 19th century due to siltation ) were also studied to see whether the causative agent of murine typhus might still circulate around the ports even after being abandoned . We found that infection risk of murine typhus was negatively associated with distance to operating seaports but was uncorrelated with nearness to abandoned seaports . In addition , the importance of distance to operating seaports for risk of murine typhus infection varied spatially . Our study highlights elevated disease risk as a cost of international trade and suggests particular surveillance in regions surrounding major international seaports .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"typhus",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"geographical",
"locations",
"microbiology",
"vertebrates",
"animals",
"mammals",
"rickettsia",
"bacterial",
"diseases",
"fleas",
"murine",
"typhus",
"bacteria",
"bacterial",
"pathogens",
"infectious",
"diseases",
"taiwan",
"medical",
"microbiology",
"microbial",
"pathogens",
"insects",
"arthropoda",
"people",
"and",
"places",
"rodents",
"asia",
"cats",
"biology",
"and",
"life",
"sciences",
"amniotes",
"organisms"
] |
2017
|
Significance of major international seaports in the distribution of murine typhus in Taiwan
|
Schistosomiasis is a parasitic flatworm disease that infects 200 million people worldwide . The drug praziquantel ( PZQ ) is the mainstay therapy but the target of this drug remains ambiguous . While PZQ paralyses and kills parasitic schistosomes , in free-living planarians PZQ caused an unusual axis duplication during regeneration to yield two-headed animals . Here , we show that PZQ activation of a neuronal Ca2+ channel modulates opposing dopaminergic and serotonergic pathways to regulate ‘head’ structure formation . Surprisingly , compounds with efficacy for either bioaminergic network in planarians also displayed antischistosomal activity , and reciprocally , agents first identified as antischistocidal compounds caused bipolar regeneration in the planarian bioassay . These divergent outcomes ( death versus axis duplication ) result from the same Ca2+ entry mechanism , and comprise unexpected Ca2+ phenologs with meaningful predictive value . Surprisingly , basic research into axis patterning mechanisms provides an unexpected route for discovering novel antischistosomal agents .
Over a third of the world's population is estimated to be infected with parasitic worms . One of the most burdensome infections underpins the neglected tropical disease schistosomiasis ( Bilharzia ) , caused by parasitic flatworms of the genus Schistosoma . The debilitating impact of schistosomiasis results from the host's immune response to schistosome eggs , which are deposited in prolific numbers in the liver , intestine and/or bladder where they elicit granuloma formation and fibrosis [1] . Clinical outcomes span gastrointestinal and liver pathologies , anaemia , undernutrition , growth retardation , genitourinary disease and a heightened risk for co-morbidities . This burden encumbers third world economies with an annual loss of several million disability-adjusted life years [2]–[4] . The key treatment for schistosome infections is the drug praziquantel ( PZQ ) . PZQ is a synthetic tetracyclic tetrahydroisoquinoline derivative discovered over 30 years ago to confer anthelminthic activity [5]–[7] by evoking a spastic paralysis of the adult worms [8] . The low cost ( ∼$0 . 07/tablet ) yet high cure rate associated with PZQ underpins current strategies for increasing PZQ distribution to reduce the burden of schistosomiasis [9] , but obviously continued efficacy of PZQ is critical for the success of these initiatives . From a drug development perspective , it remains problematic that despite three decades of clinical use , the target of PZQ remains ambiguous and synthesized structural derivatives prove consistently less efficacious [5]–[7] , [10] , [11] . Resolution of the target and effector mechanisms of PZQ would be massively helpful for identifying new drug targets that exploit vulnerabilities within the broader PZQ interactome . Recently , we have attempted to bring fresh insight into the mechanism of action of PZQ by studying an unusual impact of this drug on regeneration of a free living planarian flatworm ( Dugesia japonica ) , a representative of a model system widely utilized by basic scientists as a model for regenerative biology [12] , [13] . This line of investigation grew from the serendipitous finding that PZQ exposure invariably caused regeneration of worms with two heads ( ‘bipolar’ ) , rather than worms with normal anterior-posterior ( ‘AP’ , head to tail ) polarity [14] . The capacity of PZQ to evoke this complete AP axis duplication was phenocopied by several Ca2+ signaling modulators , a relationship underpinned by the demonstration of PZQ-evoked Ca2+ uptake in native planarian tissue [14] , [15] . The tractability of planarians to in vivo RNAi methods allowed mechanistic interrogation of various Ca2+ entry pathways , and this approach revealed the bipolarizing efficacy of PZQ depended on the expression of neuronal voltage-operated Ca2+ channel ( Cav1 ) isoforms [14] , [15] . These observations were intriguing in the context of schistosome biology , as PZQ is well documented to cause Ca2+ entry in schistosomes [8] , [16] , [17] and PZQ has been shown to activate Ca2+ entry via modulation of a heterologously expressed schistosome Cav accessory subunit [18] , [19] . But how Ca2+ entry engages acute and chronic [20]–[22] downstream signaling pathways in either planarians or schistosomes is less clear , with resolution of this broader PZQ interactome key for identifying new druggable targets and vulnerabilities for chemotherapeutic exploitation [17] . Here , we evidence a Ca2+-dependent phenology of PZQ action between these two quite different models . We propose the same Ca2+ entry and downstream pathways are engaged by PZQ in planarians and schistosomes , and the mechanistic interrelationship underpinning these different outcomes ( death in schistosomes , axis duplication in planarians ) augers predictive value for discovery of new anti-schistosomal agents . For example , in planarians , we demonstrate the planarian AP axis duplication phenotype results from coupling of Cav1A activity to bioaminergic signaling . Modulators of regenerative polarity which impact dopaminergic and serotonergic pathways in planarians are effective against schistosomes , and reciprocally recently discovered drug leads active against schistosomes ( for example , PKC and GSK3 modulators ) regulate AP specification in planarians . As unexpected phenologs [23] , this discovery underscores the utility of basic research on axis patterning mechanisms in the tractable planarian system for the discovery of novel antischistosomal drug leads , and more broadly mechanistic insight into the signaling pathways engaged by PZQ , a key human therapeutic .
Exposure of excised trunk fragments to PZQ caused regeneration of viable , two-headed flatworms ( Figure 1A ) , an effect previously shown to relate to modulation of neuronal voltage-operated calcium ( Cav ) channels [14] , [15] ) . Given the role of Ca2+ entry in synaptic and dendritic exocytosis [24] , [25] , we hypothesized that PZQ-evoked Ca2+ entry impacted neurotransmission and thereby stem cell behavior , consistent with a ‘neurohumoral’ model for regulation of planarian stem cell proliferation proposed two decades ago [26] . To test this idea , we used loss-of-function ( in vivo RNAi ) and pharmacological methods to interrogate whether different planarian neurotransmitters mimicked the PZQ-evoked bipolarity effect . Figure 1B schematically summarizes the major neurotransmitter classes in flatworms [27]–[29] , of which neuropeptides predominate by number . A recent characterization of planarian bioactive peptides revealed >50 prohormone genes , the vast majority being neuronally expressed with over 250 discrete peptides generated from these precursors [30] . Further , bioinformatic prediction supports at least 130 planarian neuropeptide targeted G protein coupled receptors [31] . This expansive neuropeptidergic arsenal co-exists with several ‘classic’ neurotransmitter families more familiar to mammalian neurophysiologists . The largest group of these transmitters are the biogenic amines , a group of protonated amines including serotonin , histamine , catecholamines ( notably dopamine ) as well as tyramine and octopamine , two phenolamines widely used as invertebrate neurotransmitters [27] , [32] . Roles for acetylcholine ( ACh ) and amino acids ( glutamate , GABA ) are also evidenced [27] , [32] . To test the involvement of these different neurotransmitter families as PZQ effectors , we used in vivo RNAi to knockdown key enzymes involved in their synthesis . Knockdown of prohormone convertase 2 ( PC2 , [33] ) , an enzyme required for motility [34] and neuropeptide processing [30] , failed to impact the penetrance of PZQ-evoked bipolarity ( Figure 1C ) . Similarly , knockdown of glutamate decarboxylase ( GDC , to decrease planarian GABA levels [35] ) , and choline acetyltransferase ( CAT , to deplete ACh [36] ) , failed to modulate the penetrance of PZQ ( Figure 1C ) . Negative results were also obtained following RNAi of tyramine-β-hydroxylase ( TBH ) and tyrosine/histidine decarboxylase ( T/HDC ) . These data were also consistent with the outcomes of pharmacological experiments where application of the phenolamines tyramine and octopamine failed to perturb AP polarity ( Table 1 ) . In contrast , results with other biogenic amines were more intriguing – knockdown of tyrosine hydroxylase ( TH ) attenuated the ability of PZQ to evoke two-headed worms , whereas knockdown of tryptophan hydroxylase ( TPH ) increased PZQ-evoked bipolarity ( Figure 1C ) . TH is the rate-limiting enzyme of catecholamine synthesis , catalyzing the conversion of tyrosine to L-dihydroxyphenylalanine ( L-DOPA ) , whereas TPH converts tryptophan to 5-hydroxytryptophan , the first step in 5-HT synthesis . Knockdown of TH in D . japonica decreases dopamine without impacting 5-HT production [37] , while knockdown of TPH decreases 5-HT but not dopamine [38] . These RNAi results suggest that PZQ activity is mimicked by dopaminergic activity ( TH RNAi ) to promote head regeneration , and this action is opposed by serotonergic signaling ( TPH RNAi ) . On the basis of this hypothesis , we proceed to screen modulators of dopamine and 5-HT receptors: dopaminergic stimuli should phenocopy the bipolarizing activity of PZQ , while PZQ action should be opposed by serotonergic agonists . While this is a reasonable approach , care must be taken in assuming the specificity of agents established in mammal models transfers to flatworm systems . Flatworms may express more bioaminergic receptors than humans [31] , and the few flatworms receptors that have been successfully expressed and pharmacologically profiled [39] underscore the risks of assuming similar drug activities to those assigned in mammals . Keeping this caveat in mind , we nevertheless used a pharmacological approach but accrued evidence with multiple ligands and used secondary validation assays to best mitigate this problem . Below , we first describe results of drug assays assuming specificities based upon mammal data , and then we return to the issue of validating ligand specificity against particular neurotransmitter pathways . A range of compounds were screened for effects on AP polarity ( Table 1 ) , and these investigations yielded the following observations . First , the exclusion of individual neurotransmitter families on the basis of RNAi results ( Figure 1C ) received further support from pharmacological screening , as most modulators of adrenergic , GABAergic , glutaminergic , histaminergic and cholinergic pathways failed to impact regenerative polarity ( Table 1 ) . Second , bromocriptine , a potent D2 agonist in mammalian systems , produced two-headed regenerants at high penetrance ( maximal effect ∼85±5% bipolar , Figure 1D & 1E ) , with an EC50 of 220 nM compared with an EC50 of ∼40 µM for PZQ ( Figure 1F ) . Other dopaminergic modulators yielded a low , but robust , proportion of two headed worms including apomorphine ( a non-selective dopaminergic agonist in mammals ) and dopamine itself ( Figure 1D & 1E ) . Third , haloperidol , a traditional antipsychotic and known inhibitor of dopaminergic signaling in planaria [40] , blocked the bipolarizing activity of both bromocriptine and PZQ ( Figure 1F , inset ) . Fourth , 5-HT blocked head regeneration , an effect observed with 5-HT , the synthetic ligand 8-OH DPAT ( a mammalian 5-HT1A agonist ) and a serotonin-specific reuptake inhibitor ( SSRI , fluoxetine , Figure 1D & 1E ) , all of which blocked the bipolarizing effect of PZQ ( IC50 ∼147 µM , 111 nM and 230 nM , Figure 1G ) . In contrast , mianserin ( a 5-HT antagonist in flatworm [41]–[43] and mammalian systems ) yielded a small proportion of two-headed worms ( Figure 1D&E ) . Given the effects of bromocriptine , we further investigated the characteristics of bromocriptine efficacy in planaria . First , bromocriptine exhibited a similar kinetic action to that observed with PZQ ( Figure 2A ) , suggesting a similar action early in regeneration . Second , while knockdown of Cav1A attenuated PZQ-evoked bipolarity , bromocriptine-evoked bipolarity persisted in Cav1A RNAi worms ( Figure 2B ) . This surprising result is consistent with the idea that bromocriptine activation of head signaling pathways occurs downstream of Cav1A function . For example , if PZQ-evoked Ca2+ entry [15] activates neurotransmitter release , then the bipolarizing efficacy of bromocriptine should persist at downstream receptors even if Ca2+ entry is impaired . Third , given concerns about presumptions of similar pharmacological effects between mammalian and flatworm systems , we investigated whether bromocriptine exhibited affinity for dopaminergic systems in planaria by performing 3H-dopamine displacement assays . Specific 3H-dopamine binding , defined by complete displacement with cold dopamine ( IC50 = 1 . 5±0 . 5 µM ) , was inhibited by bromocriptine and other head-promoting agents ( haloperidol and apomorphine , Figure 2C ) . The extent of 3H-dopamine displacement by maximally effective concentrations of haloperidol and apomorphine was greater ( >80% of specific binding ) than observed with bromocriptine ( ∼40% of specific binding at 10 µM ) . This indicated bromocriptine may exhibit selectivity for only a subset of dopaminergic targets compared to the broader and more complete binding inhibition observed with the other agents . Finally , we investigated the impact of agents presumed to impact neurotransmitter levels ( reserpine , fluoxetine ) via HPLC . Figure 2D shows that fluoxetine ( a 5-HT reuptake inhibitor on the basis of mammalian and schistosome literature [44] , [45] ) increased 5-HT levels in regenerating planarian trunk fragments , consistent with the inhibitory effects of 5-HT ( and fluoxetine ) on head regeneration ( Figure 1G ) . In contrast , reserpine exposure depleted 5-HT in regenerating fragments ( Figure 2C ) , an opposing outcome consistent with the differential polarity effects of these drugs on head regeneration ( reserpine vs fluoxetine , Figure 1D&E ) . Collectively , these pharmacological data support the model derived from RNAi data ( Figure 1C ) where dopaminergic signaling mimics and serotonergic activity opposes PZQ action . The distinct phenotypic outcomes of dopaminergic and serotonergic modulation are also consistent with observations that these neurotransmitter networks in planarians are morphologically discrete [28] . These discoveries piqued our interest since dopaminergic and serotonergic ligands have recently emerged as hits in drug screens against various schistosome life cycle stages [46] , [47] . Figure 3A collates examples of recent drug screening data to show how efficacious drug hits are distributed relative to the functional representation of drugs screened [46] , [47] . The top three functional categories represent dopaminergic and serotonergic ligands followed by regulators of ion channel activity , notably Cav channel modulators . This triumvirate parallels the PZQ-engaged components in planarians in this study ( bioaminergics , Figure 1 ) and previously ( Ca2+ channels , [14] , [15] ) . As such , we propose the distinct phenotypes - PZQ-evoked bipolarity in planarians and PZQ-evoked toxicity against schistosomes - represent unexpected yet orthologous phenotypes ( ‘phenologs’ , [23] ) resulting from engagement of the same fundamental Ca2+-triggered interactome in each system . Although PZQ-evoked Ca2+ entry is evoked via similar mechanisms ( Cav1A ) it is harnessed in the two organisms to yield differential outcomes ( ‘death’ versus ‘axes’ ) . The utility of this phenology is its predictive value . As both outcomes derive from the same effector network , basic research on axis patterning in planarians may harbor potential for discovering new agents effective as antischistosomals . This assertion can be tested by asking whether other antischistosomals cause planarian bipolarity , and reciprocally , whether bipolarizing agents in planarians are active against schistosomes . Do other antischistosomal compounds cause planarian bipolarity ? To test this , we identified the next most prevalent category from the schistosome drug screening datasets , which was the ‘phosphorylation’ category ( Figure 3A ) . The predominant group of compounds within this category were several drugs that target protein kinase C ( PKC ) , and a couple of singleton kinase inhibitors , including one targeting glycogen synthase kinase-3 ( GSK3 ) . We investigated the role of both kinases to resolve any impact on planarian regenerative polarity ( Figure 3A ) . First , the PKC activators phorbol-12-myristate-13-acetate ( PMA ) , phorbol-12 , 13-dibutyrate ( PDB ) and oleoyl-acetyl-glycerol ( OAG ) all produced bipolar worms ( penetrance ∼5–55% respectively , Figure 3B ) , while the PKC inhibitor calphostin C [48] inhibited PZQ-evoked bipolarity ( Figure 3C ) . To complement the pharmacological data with molecular insight , we cloned several planarian PKC isoforms and diacylglycerol kinase ( DAGK ) and investigated their roles in PZQ-evoked bipolarity by RNAi . Knockdown of DAGK , which opposes PKC activity via the degradation of DAG , potentiated the penetrance of sub-maximal doses of PZQ; while RNAi of a conventional PKC isoform , but not a novel and atypical PKC , attenuated PZQ evoked bipolarity ( Figure 3C ) . The involvement of a Ca2+-regulated PKC was also consistent with the observation that the penetrance of PMA in yielding bipolar regenerants was Ca2+ dependent ( Figure 3D ) . Similarly , alsterpaullone ( ALP ) , a GSK-3 inhibitor also phenocopied PZQ in regenerative assays , producing a low frequency of two headed worms and synergistically potentiating sub-maximal doses of PZQ ( Figure 3E ) . The small molecule GSK3 agonist DIF-3 [49] displayed the opposing action , inhibiting PZQ-evoked bipolarity ( Figure 3E ) . Therefore , both these targets in the ‘phosphorylation’ category prioritized from the schistosomal screening literature ( Figure 3A ) were resolved to miscue planarian AP polarity during regeneration . Are drugs that miscue planarian regeneration deleterious to schistosomes ? To investigate this issue , schistosomules ( juvenile parasites ) were exposed to compounds first identified in planarian regenerative assays ( Figure 4A ) . Schistosomules normally exhibit a basal level of spontaneous contractile activity ( Figure 4B ) , which provides a simple phenotype for assaying drug action and paralysis , an outcome integral to the elimination of schistosome infections [46] . Bromocriptine caused a rapid paralysis of schistosomules , an effect that phenocopied the action of PZQ ( Figure 4B ) . This effect was dose-dependent ( Figure 4B&C ) . Other compounds that yielded planarian bipolarity were also found to impair schistosomule contractility , including apomorphine , mianserin and reserpine ( Figure 4B ) . In contrast , application of exogenous serotonin and other ligands that inhibited planarian head regeneration ( e . g . 8-OH DPAT and fluoxetine ) resulted in hyperactivity ( Figure 4C ) . Quantification of the action of these agents which inhibited and stimulated schistosomule activity is collated in Figures 4D&E respectively . Therefore , not only were both classes of bioaminergic compounds efficacious against schistosomules , but the dopaminergic and serotonergic ligands evoked divergent phenotypes in each model: paralysis versus hyperactivity ( schistosomules ) , compared with ‘two-headed’ versus ‘no-head’ regenerants ( planaria ) . Beyond the conservation of single genes as nodes in a signaling pathway , broader network architectures are conserved between diverse organisms . While the phenotypic outputs of these networks are diverse , their common architecture provides the mechanistic basis for predictive phenology [23] . We suggest these divergent PZQ-evoked outcomes ( death versus axes ) represent unexpected Ca2+-dependent phenologs initiated by small molecule activation of a signaling node ( Cav1A ) within a shared bioaminergic interactome ( Figure 5A ) . This conservation infers reciprocal predictive value for both discovery of new antischistosomal compounds , and reciprocally new signalling pathways impacting anterior-posterior signaling in planarians . We illustrate this principle here by highlighting de novo new compounds effective against schistosomules ( bromocriptine ) and new druggable targets ( bioaminergic signaling ) as the downstream PZQ-evoked interactome is revealed in the more tractable planarian model . PZQ engages similar pathways in these different platyhelminths such that chemical/functional genetic approaches in planarians can assist in discovering next generation antischistosomals and resolving their molecular action . This line of reasoning is analogous to a longer history of studies exploiting C . elegans for comparative insight into new drugs targeting parasitic nematodes , and this experience underscores both the utility of this approach but also the frustration in harvesting viable clinical leads from a large number of efficacious compounds in both nematode models [50] , [51] . Reciprocally , this unexpected phenology can reveal new modulators of AP patterning from the schistosome screening literature ( e . g . PKC , GSK3 ) . Such insight from schistosome life cycle drug screens will be of utility for understanding the process of in vivo stem cell differentiation and CNS regeneration in response to injury that are inherent to the remarkable regenerative prowess of planarians . Indeed , resolution of the coupling of specific neuronal Cav channels to defined neurotransmitters integrates our studies of PZQ-evoked Cav activity [14] , [15] with an older literature supporting a role for bioamines in planarian regeneration [52] . But how is small molecule activation of Cav1A in one organism deleterious , but the same Ca2+ influx process harnessed physiologically in another to regulate polarity during regeneration ? We speculate the same PZQ-evoked interactome differentially couples to these outcomes because of the different ionotropic channel portfolio supporting cellular excitability in the two organisms . Planarians express a surprisingly broad array of voltage-gated entry channels - five unique Cav channels in addition to Nav channels ( Figure 5B ) . This broad channel repertoire likely permits subfunctionalization of Cav1A activity within a broad organismal complement of voltage-gated channels in planarians to yield a physiological exploitable Cav1A dependent Ca2+ influx . In contrast , schistosomes express a more limited portfolio of voltage-sensitive channels , lacking both Nav and LVA Cav channels ( Figure 5B ) . The more limited gene repertoire of these parasites imparts a dependency and thereby vulnerability to Cav1A activity within their smaller ionotropic channel portfolio . In this context , it is intriguing that both muscle contraction and tegumental damage are Ca2+ triggered phenomena in adult schistosomes ( reviewed in [17] ) , such that Ca2+ dysregulation may serve as a common nexus predictive of in vivo antihelmintic activity . Further insight into this problem will be provided by understanding how acute Ca2+-dependent effects evoked by PZQ in different schistosome tissues regulate both acute downstream targets ( bioaminergic receptors and their second messenger coupling ) and the relevance of more chronic Ca2+ dependent transcriptional effects [20] , [22] , e . g . CamKII [21] , that have emerged from recent mRNA profiling analyses . In conclusion , exploitation of this Ca2+ dependent phenology should rekindle interest in drugs such as bromocriptine , and the druggability of their cognate bioaminergic receptors , as an avenue for resolving novel antischistosomals and modulating in vivo stem cell behavior during regeneration .
A clonal line of Dugesia japonica ( GI strain ) was maintained at room temperature and fed strained chicken liver puree once a week [53] . Regenerative assays were performed using 5 day-starved worms in pH-buffered Montjuïch salts ( 1 . 6 mM NaCl , 1 . 0 mM CaCl2 , 1 . 0 mM MgSO4 , 0 . 1 mM MgCl2 , 0 . 1 mM KCl , 1 . 2 mM NaHCO3 , pH 7 . 4 buffered with 1 . 5 mM HEPES ) as described previously [53] . Drugs were sourced as indicated in Table 1 , and used either at the highest concentrations which did not impact worm viability , or at lower concentrations if such treatments elicited an effect of maximal penetrance . Planarian regenerative phenotypes were archived using a Zeiss Discovery v20 stereomicroscope and a QiCAM 12-bit cooled color CCD camera . Total RNA was isolated from 50 intact planarians using TRIzol® and poly-A purified using a NucleoTrap mRNA mini kit . cDNA was synthesized using the SuperScript™ III First-Strand Synthesis System ( Invitrogen ) . Gene products were amplified by PCR ( LA Taq™ polymerase ) , ligated into the pGEM®-T Easy vector ( Promega ) for sequencing , and subcloned into the IPTG-inducible pDONRdT7 RNAi vector transfected into RNase III deficient HT115 E . coli . In vivo RNAi was performed by feeding [53] , and a Schmidtea mediterranea six-1 ( Smed-six-1 ) construct , which did not yield a phenotype in D . japonica , was used as a negative control . RNAi efficiencies varied between different genes , but mRNA knockdown typically ranged anywhere between 20–80% . Targeted sequences: tyrosine hydroxylase ( NCBI accession numbers AB266095 . 1 , 136–1657 bp ) , tryptophan hydroxylase ( AB288367 . 1 , 4–1623 bp ) , tyramine beta-hydroxylase ( 671–1629 bp ) , tyrosine/histidine decarboxylase ( FY934632 . 1 , 26–685 bp ) , glutamate decarboxylase ( AB332029 . 1 , 154–1937 bp ) , choline acetyltransferase ( AB536929 . 1 , 74–1175 bp ) , prohormone convertase 2 ( PC2 ( 1–2285 bp ) , Cav1A ( HQ724315 . 1 , 2229–4133 bp ) , Smed-six-1 ( AJ557022 . 1 , 1–506 bp ) . Protein kinase C ( PKC ) sequences and DAGK were cloned from planarian ESTs displaying homology to Schistosoma mansoni PKC isoforms - cPKC ( FY950278 . 1 , FY947802 . 1 , FY970060 . 1 ) , aPKC ( FY933556 . 1 , FY941429 . 1 ) , nPKC ( FY934640 . 1 ) and DAGK ( FY953983 , FY959647 . 1 , and BP187372 . 1 ) . Biomphalaria glabrata snails exposed to mirarcidia ( NMRI Puerto Rican strain of Schistosoma mansoni ) were obtained from the Biomedical Research Institute ( Rockville , MD ) and maintained at 26°C for 4 to 6 weeks . Matured cercaria were shed into aged tap water ( 40 ml ) by exposure to light ( 1 . 5 hrs ) and subsequently transformed into schistosomules [54] . Briefly , cercaria were separated from debris by filtration ( 47 µm ) and then captured onto a 25 µm filter prior to resuspension in aged tap water with an equal volume of DMEM . Cercaria tails were sheared by three rounds of vortexing ( 45 sec ) , each followed by incubation on ice ( 3 min ) prior to tail removal by Percoll column centrifugation ( 24 ml Percoll , 4 ml 10× Eagle's minimum essential medium , 1 . 5 ml penicillin-streptomycin , ml of 1M HEPES in 0 . 85% NaCl , 9 . 5 ml distilled water ) at 500 g ( 15 mins , 4°C ) . The tail-containing supernatant was discarded and the pellet-containing bodies were washed three times in DMEM ( 400 g , 10 mins ) , resuspended in modified Batch's media [55] and transformed into schistosomules ( incubation at 37°C/5% CO2 ) . For contractility assays , drugs were solubilized in DMSO and diluted in pre-warmed modified Batch's media . While detailed protocols for quantifying aspects of worm dynamics in adult worms [21] , or higher throughput screening of schistosomules [56] have been developed , the effects on schistosomule activity were simply quantified here using a custom written plugin ( wrMTrck ) in ImageJ to using resolve schistosomule body length ( major axis of an ellipse ) over time following drug exposure ( 30 min ) , just as in [39] . Videos were captured using a Nikon Coolpix 5700 camera affixed to a Nikon Eclipse TS100 microscope . Typically , for a single video ∼7–10 schistosomula were measured within the field of view ( 10× microscope objective ) over a 2 minute recording period . Data represent means for analysis of results from three independent treatments . Planarian membrane fractions were prepared by homogenizing worms on ice ( ∼1000 worms/prep ) in HEPES ( 20 mM ) supplemented with cOmplete™ protease inhibitor cocktail ( Roche ) . Cellular debris was pelleted by centrifugation ( 8000 g for 5 mins ) and the resulting supernatant was centrifuged ( 56 , 000 g for 45 mins ) to yield a pelleted membrane fraction . This material was resuspended ( 20 mM HEPES , with protease inhibitors ) to a final protein concentration of ∼5 µg/µl and stored at −80°C . Binding assays were performed on planarian membrane protein ( 50 µg ) with 26 nM 3H-dopamine ( specific activity 21 . 2 Ci/mmol , Perkin Elmer ) . Indirect binding assays were performed with various ligands ( bromocriptine , 10 µM; haloperidol , 100 µM; apomorphine , 10 µM ) in TE buffer ( 1 mM EDTA , 50 mM Tris-HCl , pH 8 . 3; final volume of 500 µl ) . Samples were incubated on ice for 15 minutes , after which time 500 µl PEG ( 30% ) and 20 µL IgG ( 25 mg/ml ) were added and samples centrifuged ( 20 , 000 g for 5 mins ) . The resulting pellet was washed ( PEG , 15% ) , centrifuged ( 20 , 000 g for 5 mins ) and solubilized in TE ( 200 µL , containing 2% Triton X-100 ) . Displacement was measured by liquid scintillation counting and nonspecific binding assessed by subtraction of values in samples incubated with cold dopamine ( 1 mM ) . All centrifugation steps were performed at 4°C . Thirty planarian trunk fragments were amputated and incubated with or without specific drugs for 24 hrs , after which time media was removed and replaced with ascorbic acid ( 300 µl , 1% m/v ) . Samples were then lysed by three successive freeze-thaw cycles and cellular debris pelleted by centrifugation ( 10 , 000 g for 5 mins ) . The resulting supernatant was then filtered ( 0 . 45 µm filter plate , Millipore ) by centrifugation ( 3 , 000 g for 10 mins ) and the filtrate ( 180 µL ) supplemented with 0 . 5M HClO4 ( 20 µl , 500 mM final concentration ) . The samples were mixed and injected by an autosampler into an Agilent 1200 HPLC apparatus , with a 5 µm , 4 . 6×150 mm Eclipse XDB C18 column attached to a Waters 2465 electrochemical detector with a glassy carbon-based electrode . The current range was set at 50 nA with a working potential of 0 . 7 V versus an in situ Ag/AgCl reference electrode . The mobile phase mixture ( 13 mg/L of the surfactant sodium octylsulfate , 170 µL/L dibutylamine , 55 . 8 mg/L Na2EDTA , 10% methanol , 203 mg/L sodium acetate anhydrous , 0 . 1M citric acid , and 120 mg/L sodium chloride ) was ran at a flow rate of 2 ml/min . The area underneath the peaks was analyzed for total amount of serotonin and dopamine . Results were normalized to sample protein concentration determined by Bradford assay ( Thermo Scientific ) . Data were analyzed using two-tailed , unpaired t-tests , and presented as mean ± standard error of the mean from at least three independent assays , except where indicated . Differences were considered significant at p<0 . 05 ( * ) , p<0 . 01 ( ** ) .
|
Schistosomiasis ( Bilharzia ) is one of the most burdensome parasitic worm infections , encumbering third world economies with an annual loss of several million disability-adjusted life years . The key treatment for schistosome infections is the drug praziquantel but the mechanism of action of this drug remains controversial hampering targeted development of next generation antischistosomal agents . Here we provide fresh insight into the signaling pathways engaged by PZQ , by resolving commonalities in the action of PZQ with the process of regenerative signaling in free-living planarian flatworms . A similar calcium-dependent network is engaged in both model systems , but with divergent phenotypic outcomes . This relationship provides predictive insight such that basic research on signaling pathways involved in tissue regeneration reveals novel drug leads for schistosomiasis , and reciprocally schistosomal drug screens reveal targets involved in regenerative signaling . We believe this phenology will be helpful for uncovering new antischistosomal drug targets by exploiting broader vulnerabilities within the PZQ interactome .
|
[
"Abstract",
"Introduction",
"Results",
"&",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"developmental",
"biology",
"molecular",
"neuroscience",
"stem",
"cells",
"small",
"molecules",
"parasitology",
"neurotransmitters",
"biology",
"cell",
"differentiation",
"zoology",
"neuroscience",
"pattern",
"formation",
"drug",
"discovery",
"evolutionary",
"developmental",
"biology"
] |
2014
|
‘Death and Axes’: Unexpected Ca2+ Entry Phenologs Predict New Anti-schistosomal Agents
|
The bacterium Burkholderia ubonensis is commonly co-isolated from environmental specimens harbouring the melioidosis pathogen , Burkholderia pseudomallei . B . ubonensis has been reported in northern Australia and Thailand but not North America , suggesting similar geographic distribution to B . pseudomallei . Unlike most other Burkholderia cepacia complex ( Bcc ) species , B . ubonensis is considered non-pathogenic , although its virulence potential has not been tested . Antibiotic resistance in B . ubonensis , particularly towards drugs used to treat the most severe B . pseudomallei infections , has also been poorly characterised . This study examined the population biology of B . ubonensis , and includes the first reported isolates from the Caribbean . Phylogenomic analysis of 264 B . ubonensis genomes identified distinct clades that corresponded with geographic origin , similar to B . pseudomallei . A small proportion ( 4% ) of strains lacked the 920kb chromosome III replicon , with discordance of presence/absence amongst genetically highly related strains , demonstrating that the third chromosome of B . ubonensis , like other Bcc species , probably encodes for a nonessential pC3 megaplasmid . Multilocus sequence typing using the B . pseudomallei scheme revealed that one-third of strains lack the “housekeeping” narK locus . In comparison , all strains could be genotyped using the Bcc scheme . Several strains possessed high-level meropenem resistance ( ≥32 μg/mL ) , a concern due to potential transmission of this phenotype to B . pseudomallei . In silico analysis uncovered a high degree of heterogeneity among the lipopolysaccharide O-antigen cluster loci , with at least 35 different variants identified . Finally , we show that Asian B . ubonensis isolate RF23-BP41 is avirulent in the BALB/c mouse model via a subcutaneous route of infection . Our results provide several new insights into the biology of this understudied species .
The Gram-negative soil- and water-dwelling bacterium B . ubonensis is a member of the Burkholderia cepacia complex ( Bcc ) [1] , a genetically related group of metabolically diverse , highly adaptable and widely dispersed environmental species [2] . The Bcc , which comprises at least 20 species , includes some members known for their ability to cause clinical disease , such as severe sepsis in the immunocompromised and progressive pulmonary disease in cystic fibrosis patients [3] . Many Bcc species are also recognised for their unique biotechnological potential , particularly in bioremediation applications and in the production of antibiotic and antifungal compounds [4] . Novel compounds produced by B . ubonensis have been proposed as potential agents in biocontrol against Burkholderia pseudomallei [5] and in biodiesel catalysis [6] . B . pseudomallei , the causative agent of the tropical infectious disease melioidosis , is frequently isolated from the same soil samples as B . ubonensis in regions where both species are endemic [7] . Melioidosis is a diagnostically challenging and often deadly disease that affects humans and many animals , and remains underdiagnosed in many regions across the globe [8] . As B . pseudomallei is not a part of the healthy human flora , the ‘gold standard’ method for melioidosis confirmation is growth of B . pseudomallei from clinical specimens . For maximum isolation of B . pseudomallei from non-sterile sites such as sputum and pus , clinical laboratories require selective culture methods such as Ashdown’s agar containing gentamicin [9] and Ashdown’s broth containing colistin [10] . These media have also been used to successfully isolate B . pseudomallei from microbiologically complex environmental samples such as soil and surface water , which would otherwise yield growth and dominance of many other species [11] . We have previously demonstrated that B . ubonensis is the most commonly co-isolated species when using B . pseudomallei enrichment methods in the melioidosis-endemic “Top End” of the Northern Territory , Australia , in part due to the indistinguishable nature of certain B . ubonensis and B . pseudomallei morphotypes [7] . In addition , it has been reported that the atypical B . pseudomallei O-antigen type B is found in 25% of B . ubonensis strains from Australia [12] , further complicating the differentiation between these species due to their immunological cross-reactivity . Little is currently known about the population biology and genomics of B . ubonensis , although a clearer picture is emerging . The first B . ubonensis isolate ( “B . uboniae” EY 3383 , isolated from soil in Ubon Ratchathani in 1989 ) was reported in 2000 [13] , and the first B . ubonensis genome ( MSMB0022 , isolated from soil in Darwin , Australia , in 2001 ) was sequenced to closure in 2015 [14] . The MSMB0022 genome encodes three circular replicons totalling ~7 . 2Mbp , which is approximately the same size as the two-chromosome B . pseudomallei genome . In Bcc species , the third replicon , a megaplasmid called pC3 ( formerly chromosome III ) , has been shown to be important for stress resistance , virulence , and antifungal and proteolytic activity in several strains [15 , 16] . This replicon is not essential for survival , with ~4% of tested Bcc isolates having spontaneously lost pC3 , and additional strains able to be cured of this replicon either by plasmid incompatibility or by removal or toxin-antitoxin systems [15] . Although pC3 loss in B . ubonensis has been achieved in vitro , pC3 loss in wild-type B . ubonensis strains has not yet been identified . Previous work has shown that Bcc species can encode for innate high-level resistance towards many clinically relevant antibiotics , including the carbapenem antibiotic meropenem [17] . Meropenem is a critical antibiotic for melioidosis therapy , being considered the treatment of choice for those with life-threatening sepsis [18 , 19] . To date , the vast majority of B . pseudomallei isolates have been fully susceptible to meropenem [20] , although recent evidence has shown that decreased susceptibility towards meropenem can occur after prolonged use of this antibiotic in melioidosis patients with severe sepsis [21] . Certain Bcc species such as B . vietnamiensis , B . cepacia and B . cenocepacia [22] , as well as B . pseudomallei [23 , 24] , exhibit high rates of intra-species recombination . This observation raises the concern that antibiotic resistance genes may spread amongst Burkholderia species in the environment and potentially to the globally important pathogen B . pseudomallei . The current study describes the first comprehensive analysis of the population biology of B . ubonensis from Australia and Asia . In addition , we identify the first B . ubonensis isolates from the Caribbean . Using large-scale comparative genome analysis , we interrogated 264 B . ubonensis genomes to better understand the geographic distribution and genetic diversity of this species , including potential loss of the pC3 megaplasmid . We also explored rates of meropenem resistance in Asian and Australian B . ubonensis strains , lipopolysaccharide ( LPS ) O-antigen cluster prevalence and diversity , and the virulence potential of an Asian B . ubonensis strain in the BALB/c mouse model .
Procedures and ethics approval for collection of the environmental specimens from which the B . ubonensis isolates were recovered has been previously described [7 , 25] . The murine challenge work was conducted according to the specific guidelines provided by the United States Department of Agriculture Animal Welfare Act under approved protocols from the Northern Arizona University IACUC ( Protocol 14–011 ) and the USA Department of Defense Animal Care and Use Review Office ( ACURO approval for HDTRA1-12-C-0066_Wagner ) . The 264 B . ubonensis isolates examined in this study originated from northern Australia ( n = 238 ) , Central Australia ( n = 4 ) , Ubon Ratchathani , Thailand ( n = 15 ) , Papua New Guinea ( PNG; n = 1 ) , and Puerto Rico ( n = 6 ) , and were obtained from samples of soil ( n = 160 ) , water ( n = 15 ) , or plant material ( n = 2 ) ( S1 Table ) . DNA was extracted using protocols optimised for B . pseudomallei [26] , and quality-checked using a NanoDrop UV spectrophotometer . Prior to WGS , all isolates were verified as B . ubonensis using the Bu550 real-time PCR [7] , which targets the conserved iron-containing redox enzyme family protein encoded by BW23_5472 on chromosome II of MSMB0022 ( also referred to as MSMB22 [14] ) . Genomic data were already publicly available for 230 of the 264 isolates [14 , 27] . For completeness , we performed paired-end sequencing of the remaining 34 isolates using a HiSeq2000 instrument ( Illumina Inc . , San Diego , CA ) at the Translational Genomics Research Institute ( Phoenix and Flagstaff; AZ , USA ) . Assemblies were performed with the Microbial Genome Assembler Pipeline ( MGAP; https://github . com/dsarov/MGAP---Microbial-Genome-Assembler-Pipeline ) , which incorporates Trimmomatic [28] , Velvet [29] , VelvetOptimiser ( https://github . com/tseemann/VelvetOptimiser ) , GapFiller [30] , PAGIT [31] and SSPACE [32] into its workflow , using the closed B . ubonensis MSMB0022 genome [14] as a reference for aligning , reordering and orientating contigs . All assemblies were quality-assessed by BLAST against phiX , with any contigs corresponding to this bacteriophage removed . Assemblies were annotated using PGAP [33] . Reference accessions for all 264 genomes are listed in S1 Table . The default settings of SPANDx v3 . 0 [34] were used to identify biallelic single-nucleotide polymorphisms ( SNPs ) from the 264 B . ubonensis genomes for phylogenetic analysis . B . ubonensis MSMB0022 was used as a reference genome for paired-end read alignment . BEDTools [35] , which is run by default in SPANDx , was used to determine gene presence/absence relative to MSMB0022 using a 1kb locus ‘window’ size . Loci were considered variable if they had ≤99% read coverage in one or more strains , and conserved otherwise . To confirm the loss of pC3 ( previously called chromosome III ) in 10 isolates and to rule out alternative replicons being present in these strains , the unmapped reads from SPANDx for each strain were assembled using MGAP . BEDTools was also used to determine LPS O-antigen type based on mapping quality against both known and novel LPS O-antigen clusters . Known clusters included B . pseudomallei K96243 ( Type A LPS; GenBank reference BX571965 . 1; coordinates 3196645–3215231 ) , B . ubonensis MSMB0057 ( Type B LPS; GenBank reference JF745807 ) , B . pseudomallei 576 ( Type B LPS; GenBank reference NZ_CP008777 . 1; coordinates 1383179–1418799 ) , B . ubonensis MSMB0122 ( Type B2 LPS; GenBank reference HQ908420 . 1 ) , B . pseudomallei MSHR0840 ( Type B2 LPS; GenBank reference GU574442 . 1 ) , B . thailandensis 82172 ( Type B2 LPS; GenBank reference JQ783347 . 1 ) and B . humptydooensis MSMB0043 ( novel LPS; GenBank reference CP013380 . 1; coordinates 971381–996024 ) . B . ubonensis type strains for determining the prevalence of novel LPS O-antigen genotypes were: A21 , BDU9 , BDU12 , BDU14 , INT1-BP158 , MSMB0022 , MSMB0054 , MSMB0063 , MSMB0083 , MSMB0103 , MSMB0268a , MSMB0609 , MSMB0742 , MSMB0782 , MSMB0827 , MSMB1058 , MSMB1137 , MSMB1172 , MSMB1173 , MSMB1178 , MSMB1189 , MSMB1206 , MSMB1304 , MSMB1471 , MSMB1517 , MSMB1586 , MSMB1591 , MSMB2105 , MSMB2123 , MSMB2166 , MSMB2180 , MSMB2207 , RF23-BP17 , and RF32-BP11 . Sequence coordinates for these LPS O-antigen clusters were extracted from MGAP-assembled genomes based on Mega BLAST analysis against the MSMB0057 O-antigen biosynthesis cluster [12] . In silico MLST was carried out on all isolates using the Bcc scheme ( http://pubmlst . org/bcc/ ) , and on 173 of the 264 B . ubonensis isolates based on the B . pseudomallei scheme ( http://pubmlst . org/bpseudomallei/ ) . Ninety-one strains could not be genotyped using the B . pseudomallei scheme as they lack the narK housekeeping locus [36] . Sequence types ( STs ) were determined from assemblies using the BIGSdb tool , which is integrated into these MLST websites [37] . ST assignments for both schemes are listed in S1 Table and are also searchable on the online databases . The maximum parsimony function of PAUP v4 . 0a153 [38] was used for phylogenetic reconstruction of genome-wide variants . The Ortho_SNP_matrix . nex output automatically generated by SPANDx was used as the PAUP input . Trees were constructed based on a heuristic search and bootstrapped using 100 replicates . FigTree ( http://tree . bio . ed . ac . uk/software/figtree/ ) was used to visualise PAUP outputs . To promote pC3 loss in vitro , phylogenetically unrelated B . ubonensis strains MSMB0782 and MSMB1215 were passaged five times on Ashdown’s agar ( 37°C for 24-48h ) , and strains INT1-BP274 and RF23-BP41 were passaged 10 times . MSMB0782 and INT1-BP274 were also subjected to five freeze/thaws ranging from -80°C to room temperature , and INT1-BP274 was passaged seven times at 42°C or room temperature . Eighteen colonies of MSMB2036 , which is the same ST as the pC3-negative strain MSMB2035 , were then examined for pC3 loss by passaging once on Luria-Bertani agar and growing at 37°C for 48h . DNA from all laboratory-passaged strains was extracted using a chelex heat soak procedure [39] and diluted 1:10 prior to PCR . pC3 detection was carried out with primers Bu_pC3_For1 ( 5’-CGATGAGCTATTCGTTCGATCT ) and Bu_ pC3_Rev1 ( 5’-AACGTGATCCGGTACAGCAC ) to generate a 52bp amplicon , using a slowdown PCR for GC-rich templates [40] . MSMB2035 was included as the pC3-negative control . All DNA was verified for quality using the Bu550 assay [7] . Etests ( bioMérieux , Baulkham Hills , NSW , Australia ) were used to determine meropenem MICs in 40 B . ubonensis strains ( S1 Table ) . This subset of strains was chosen to represent geographically and phylogenetically diverse taxa , and to identify potential MIC differences among strains of the same ST . Isolates were grown on Mueller Hinton agar for 24h at 37°C in an oxygenated environment prior to MIC assessment . The ability of B . ubonensis to cause disease via the subcutaneous ( sc ) route of infection was examined in a murine BALB/c model using a Thai environmental isolate , RF23-BP41 ( S1 Table ) , collected by Northern Arizona University in 2007 . We compared the results to sc infection with B . thailandensis type strain E264 , which is known to cause death in mice at high doses ( >106 colony forming units , or CFU ) when delivered via the intraperitoneal [41] , intranasal [42 , 43] or aerosol [44] routes . Virulence testing was performed in a similar manner as previously described [45] . After shipping , mice were acclimatised for five days before the experiment; food and water were provided ad libitum throughout the study . Mice were lightly anaesthetised with vaporised isoflurane and injected via a single 100μL sc injection in the scruff of the neck . All mice in a single cage received the same infectious dose ( B . ubonensis: 1 . 71 x 104 , 105 or 106 CFU ) . Three infection control mice were injected in an identical way , but with 100μL of sterile 1x PBS instead of bacterial culture . Mice were monitored daily for health status and euthanased on day 21 post-injection with CO2 gas followed by exsanguination .
The true global distribution of B . ubonensis is not known . To date , strains have only been reported from the environment in Wuhan , China [6] , Ubon Ratchathani , Thailand [13] , northern and Central Australia [7] , and PNG [5] . In this study , we identified B . ubonensis in the Caribbean environment for the first time , with six isolates retrieved from soil obtained from the north-central and north-eastern regions of Puerto Rico ( Juncos , Ceiba and Barceloneta ) . A recent study of soil samples in the southern United States to determine the presence of Burkholderia spp . , and particularly B . pseudomallei , did not yield a single B . ubonensis or B . pseudomallei isolate , although several other Bcc species were retrieved [40] . It is thus probable that neither B . ubonensis nor B . pseudomallei are naturally found in the environment in North America . It remains to be determined whether B . ubonensis is found in other melioidosis-endemic regions such as Africa , Central America , the Indian Ocean islands , South America or South Asia . A B . ubonensis phylogeny was reconstructed from 264 genomes derived from Australian , Thai , PNG and Puerto Rican isolates to determine the existence of a continental phylogeographic signal , a phenomenon that has been described in B . pseudomallei [23 , 46 , 47] . Based on 589 , 433 biallelic SNPs , six distinct and well-supported clades were identified . Clades II , IV , V and VI solely contained Australian B . ubonensis isolates ( n = 240 ) , whereas Clade I contained all isolates from Thailand ( n = 15 ) , the PNG isolate A21 , and two Australian strains from the tropical “Top End” region of the Northern Territory , and Clade III was comprised of the six Puerto Rican isolates ( Fig 1; S1 Fig ) . Subclades within Clade I showed that the Thai strains clustered most closely with one another ( Fig 1 ) , with A21 residing on its own branch and the two Australian strains , MSMB2035 and MSMB2036 , sharing a node with the PNG isolate . The Puerto Rican isolates share a node with the Clade IV Australian isolates ( Fig 1 ) . Due to limited availability of B . ubonensis from PNG , it could not be determined whether other PNG isolates group with A21 , although we hypothesise that PNG B . ubonensis strains will be related based on the relatively narrow genetic diversity observed in PNG B . pseudomallei populations [47 , 48] . Within Clade IV , four isolates from the arid region of Central Australia ( MSMB2166 , MSMB2167 , MSMB2185 and MSMB2186 ) , which were obtained from the same soil sample , grouped with other Australian strains , with the most closely related isolates originating from the “Top End” region . Taken together , these results demonstrate that , like B . pseudomallei , B . ubonensis populations exhibit a continental phylogeographic signal , although more samples from Asia and PNG would be needed to improve resolution of subclades within Clade I . We compared B . ubonensis MLST genotypes obtained using both the B . pseudomallei and Bcc MLST schemes with phylogenomic assignment to determine whether the STs reflected isolate relatedness on the genome level [49] , or whether homoplasy was evident among STs as has been observed with certain B . pseudomallei STs [50 , 51] . For both MLST schemes , the ST and genomic data showed excellent concordance and no evidence of ST homoplasy , with all identical STs clustering closely on the phylogeny ( Fig 1A; green box outlines ) and non-identical STs residing on separate branches . Unlike the Bcc scheme , where STs could be assigned from all genomes , STs were not able to be determined for 91 ( 35% ) isolates using the B . pseudomallei MLST scheme due to these strains lacking the “housekeeping” locus narK [36] . We identified five separate clusters within our phylogeny that lacked narK ( Fig 1A; blue branches ) . The first included all the Thai isolates ( n = 15 ) , with the remaining four comprising all Puerto Rican ( Clade III; n = 6 ) and Clade IV ( n = 14 ) isolates , plus 57 isolates within Clade VI that were isolated from various “Top End” locales . These results show that certain B . ubonensis strains cannot be fully genotyped with the B . pseudomallei MLST scheme . However , in three instances where strains could be genotyped , the B . pseudomallei scheme was superior at differentiating strains that were related yet distinct on a genomic level ( Fig 1 , red branches; S1 Table ) . MSMBs 1225 and 1559 were both ST-1187 using the Bcc scheme but were different STs using the B . pseudomallei scheme; MSMB2013 was assigned ST-1235 by both schemes but the other Bcc ST-1235 strains were found to be ST-1226 according to the B . pseudomallei scheme; and the Bcc ST-1148 strains were separated into ST-1266 and ST-1267 based on the B . pseudomallei alleles . In all cases where additional STs were found , the isolates were obtained from distinct soil samples , indicating greater resolving power of the B . pseudomallei MLST scheme in these cases . We mapped meropenem MICs for 40 strains against the genome phylogeny to ascertain whether meropenem-resistant , meropenem-intermediate or meropenem-sensitive strains belonged to a single clade . Eleven strains ( Bp8955 , Bp8958 , Bp8960 , Bp8961 , Bp8962 , Bp8964 , MSMB1162 , MSMB1471 , MSMB2166 , RF32-BP11 and RF32-BP3 ) showed high-level resistance ( ≥32 μg/mL ) towards this antibiotic , including all six Puerto Rican strains . In contrast , two Australian strains ( MSMB1215 and MSMB2152 ) exhibited the lowest MICs at 2–3 μg/mL ( Fig 2 ) . Both highly resistant and highly sensitive ( 2–6 μg/mL ) strains were found in the Asian and Australian populations , demonstrating that these phenotypes are not restricted to a certain clade and that B . ubonensis populations from these two geographic regions encode for a range of meropenem MICs ( Fig 1B ) . Although our testing was not comprehensive , we did observe similar MICs for closely related strains . For example , the closely related Thai strains RF23-BP93 , RF32-BP4 and RF32-BP6 all exhibited MICs of 24 μg/mL ( Fig 1B ) . The lack of phylogenetic congruence of high-level meropenem-resistant strains supports the hypothesis that the genetic mechanism conferring resistance is laterally transferred among strains . Alternatively , resistance may have arisen multiple times or through multiple mechanisms during the evolution of B . ubonensis due to similar environmental pressures . Many other Bcc species strains can exhibit high-level meropenem resistance [17 , 52] , indicating that this trait is not specific to B . ubonensis , although the basis for this resistance and its persistence in Bcc populations is not clear . In comparison , the highest meropenem MICs recorded for B . pseudomallei to date are ~4 μg/mL [53 , 54] , with wild-type strains consistently exhibiting MICs of 0 . 75–1 μg/mL . Unlike B . pseudomallei , where human-to-human transmission is exceptionally rare and where infections are almost always acquired from the environment [55] , Bcc species can transmit between individuals , and indeed this a major clinical issue in the management of cystic fibrosis cohorts [56] . The selective forces acting upon Bcc strains in patients receiving meropenem or other antibiotics may encourage this phenotype to persist in the population , although the lack of human B . ubonensis infections and the identification of high-level meropenem resistance in environmental samples argue against this route of selection in the context of B . ubonensis . B . pseudomallei does not encode a carbapenamase , which likely explains why high-level resistance has not been reported . However , it is conceivable that B . pseudomallei may acquire a carbapenamase whilst residing in the environment , especially from closely related species that share this niche , such as B . ubonensis or other Bcc species . Determining the molecular basis for high-level meropenem resistance in B . ubonensis and in other Bcc species should be a focus of future studies to not only promote a better understanding of resistance mechanisms in these species , but to also provide a basis for proactive monitoring of B . pseudomallei populations in the event of carbapenamase acquisition . MLST revealed that B . ubonensis is a highly diverse species . We found 128 STs among the 173 strains that could be genotyped using the B . pseudomallei MLST scheme , and 182 STs among the 264 strains based on the Bcc scheme , although these numbers underestimate diversity due to multiple related isolates being tested from single environmental specimens in our study ( S1 Table ) . Among the 33 Bcc scheme STs represented by two or more B . ubonensis isolates , 27 ( 82% ) of these STs were found within a single sample; such samples are likely to be identical or clonally related due to their physical proximity . We next examined B . ubonensis diversity within our environmental samples . Of the 51 samples where two or more B . ubonensis isolates were retrieved , 26 ( 51% ) exhibited two or more STs , revealing that multiple B . ubonensis genotypes commonly exist within single environmental samples . This result reflects similar observations made in studies examining B . pseudomallei diversity in environmental samples from Thailand [57 , 58] , B . vietnamiensis in the United States [40] , and B . cepacia genomovar III ( now known as B . cenocepacia ) in the United States , Canada and Australia [59] . Whilst isolation of multiple colonies from a single sample is a laborious endeavour , these studies reinforce the need to collect multiple isolates from individual samples to maximise capture of population diversity . Gene presence/absence analysis of the 264 B . ubonensis genomes against the MSMB0022 reference showed that 2 . 78Mbp ( 39% ) of the B . ubonensis reference genome was variably present , with the remaining 4 . 41Mbp conserved across these strains . Ten phylogenetically unrelated strains ( A21 , MSMB0312a , MSMB0668 , MSMB0705 , MSMB1080 , MSMB1509 , MSMB1520 , MSMB1809 , MSMB2035 and MSMB2108 ) failed to map reads against the entire sequence for pC3 , equating to one-third of the variable regions observed in our dataset ( Fig 3 ) . Certain closely related strains did not share this pattern: for example , MSMB2035 and MSMB2036 are clonal according to the two MLST schemes and the WGS phylogeny , yet only MSMB2035 lacked this replicon . Phylogenetic reconstruction using just pC3 as the reference showed no evidence of lateral transfer , with the topology of the tree being highly similar to the phylogenetic tree constructed for chromosomes I and II ( Fig 1 ) . This result suggests that pC3 is probably ubiquitous in B . ubonensis strains found in the environment and that it largely follows a vertical path of evolution , but , when propagated under certain conditions , segregation of this replicon can occur spontaneously; in our study , segregation occurred in 4% of strains . Agnoli and coworkers ( 2014 ) also observed that four of 110 Bcc isolates tested in their study ( 4% ) had lost pC3 , with one of these events having been confirmed to have occurred following laboratory passage [15] . In the type strain MSMB0022 , pC3 encodes for 669 genes that are involved in myriad functions ( S2 Table ) . When excluding this replicon , 1 . 86Mbp ( 26% ) of the B . ubonensis reference genome was variable among the 264 strains . The conservation of pC3 and its phylogenetic relatedness to chromosomes I and II confirms that pC3 is under strong selection pressure to be maintained in Bcc species , including B . ubonensis . However , certain growth conditions appear to encourage pC3 segregation , raising the possibility that this replicon may be a megaplasmid [60] . Based on the earlier work of Agnoli and colleagues [15 , 16] , we attempted to cure B . ubonensis strains MSMB0782 , MSMB1215 , INT1-BP274 and RF23-BP41 of pC3 by performing laboratory passage and growth under varying conditions , including multiple freeze/thaws , growth at 42°C and room temperature , or multiple passages . Despite these attempts , none were successful at segregating pC3 . To examine whether an insufficient number of colonies were being tested , we next attempted passage of 18 colonies of MSMB2036 , which is closely related to the pC3-lacking strain MSMB2035 . Four ( 22% ) colonies lost pC3 after a single passage on Luria-Bertani agar at 37°C for 48h , as observed by a lack of amplification using the Bu_pC3 primers . This finding demonstrates that , as with other Bcc species , the third replicon of B . ubonensis is not necessary for the organism’s survival , at least in a laboratory setting . It remains to be determined whether pC3 replicates independently of the two chromosomes in B . ubonensis . It has been proposed that the second ( and where applicable ) third ‘chromosomes’ found in approximately 10% of bacterial genomes are in fact ‘chromids’ , a term used to define replicons that are not strictly chromosomes or plasmids [61] . To maintain consistency with the work of Agnoli and colleagues [15 , 16] , we have chosen to refer to this replicon as a pC3 megaplasmid . At 920kb , the B . ubonensis pC3 megaplasmid is unusually large , although such size is not unprecedented , with B . cenocepacia H111 encoding a curable 1 . 04Mbp pC3 megaplasmid [16] . Larger megaplasmids have been identified in other soil- and rhizosphere-dwelling organisms including a 1 . 8Mbp linear megaplasmid identified in the actinomycete Streptomyces clavuligerus [62] , and a 1 . 59Mbp megaplasmid in Azospirillum brasilense [63] . The pC3 replicon of B . ubonensis MSMB0022 failed to be detected as a plasmid using the online PlasmidFinder and VecScreen tools; however , we found that these tools also failed to identify the B . vietnamiensis megaplasmid pBVIE01 , possibly because PlasmidFinder has been optimised for plasmid identification in Enterobacteriaceae [64] . BLAST analysis of parA and parB genes from B . vietnamiensis G4 pBVIE01 showed weak evidence of these partitioning system genes in MSMB0022 pC3 , although more solid BLAST hits were obtained with chromosome I genes . This result does not rule out the presence of plasmid maintenance loci encoded on this replicon , but rather demonstrates the difficulties in identifying genetic homology across distantly related species . Similarly , the presence of 5S , 16S and 23S ribosomal RNA-encoding genes on pC3 does not necessarily rule out this replicon as being a megaplasmid [16 , 60] . Read depth coverage analysis of pC3 showed similar depth to the two chromosomes ( e . g . MSMB0011: 108x for pC3 vs 123x for chromosome I and 124x for chromosome II ) , indicating that this megaplasmid is at a low or single copy number , a finding that is consistent with the generally low copy number of larger plasmids [65] . Earlier work has shown that 25% of Australian B . ubonensis strains possess the unusual B . pseudomallei type B LPS O-antigen [12] . Using our larger dataset , we examined LPS diversity among the 264 strains in silico . Due to insufficient contig coverage across the LPS cluster , 19 strains could not be fully genotyped using this approach; however , these strains did not possess clusters matching to other LPS types . Of the remaining 245 strains that could be genotyped , type B LPS was identified in 20 ( 8% ) . In total , 35 different LPS types were found , compared with only four LPS types among 477 global B . pseudomallei strains using the same in silico approach . The most abundant LPS type in the B . ubonensis cohort was MSMB0063 Type Novel , with 28 strains having this genotype; in contrast , eleven LPS types were seen in only a single isolate ( S1 Table ) . LPS genotypes were not restricted to particular STs or geographic regions . For example , the Thai strains RF25-BP1 and RF32-BP3 possessed an LPS cluster that was also found in Australian strains MSMB0782 , MSMB0783 , MSMB1188 , MSMB1562 , MSMB1603 , and MSMB1635 , and among these eight isolates , seven different STs were present . Our findings are consistent with the presence of similar LPS types among Burkholderia species . In addition , we show that B . ubonensis LPS is highly variable and is not associated with the genetic relatedness or geographic origin of an isolate , and would thus be a poor marker for such purposes . Unlike other Bcc species or B . pseudomallei , B . ubonensis is thought to rarely , if ever , cause disease in humans [66] , as evidenced by B . ubonensis being the only Bcc species not yet retrieved from cystic fibrosis sputum [52] . Indeed , there is only a single report of B . ubonensis being isolated from a human infection , a Thai nosocomial case ( strain LMG 24263 [1] ) . Given the absence of other reported B . ubonensis infections to date , the role of B . ubonensis as the aetiologic agent in this Thai case should be treated with scepticism; for instance , testing for the presence of known pathogens in the same clinical specimen was not stated . However , another possibility is that certain B . ubonensis strains are in fact capable of causing disease , with such cases remaining unreported due to insufficient or inaccurate differentiation of B . ubonensis from other Bcc species . To further examine the virulence potential of B . ubonensis , we inoculated BALB/c mice via sc injection using 1 . 7x 104 , 105 , and 106 CFU of the Thai strain RF23-BP41 . To our knowledge , B . ubonensis virulence has not yet been tested in the mouse model . RF23-BP41 was chosen for several reasons . First , its Thai origin maximises the probability of genetic relatedness to the putatively pathogenic LMG 24263 strain . Second , RF23-BP41 was isolated from a region where individuals ( particularly rice farmers ) regularly come into contact with soil bacteria , increasing the likelihood of successful human infection . Third , this strain demonstrated resistance towards meropenem ( MIC 16μg/mL ) , which would potentially confer a selective advantage during antibiotic treatment . Finally , this strain harbours pC3 , which has been shown to impart virulence capacity in other Bcc species [15 , 16] . Even at the highest dose of 1 . 7x106 CFU , no mice exhibited weight loss or lethargy during the 21-day challenge experiment , with their health status identical to that of the three control mice . The same result was observed in the BALB/c mice subcutaneously injected with B . thailandensis E264 at a similar dosage range [45] . Certain B . thailandensis strains are capable of infecting immunocompromised humans [67–69] , and can be lethal in murine models when administered at high doses via other routes [41–44] . In contrast , in other studies the 10-day LD50 of B . pseudomallei in BALB/c mice was ~1x103 CFU when delivered via the sc route [70] , and between 10 and 6x104 CFU when administered via the intraperitoneal route , with virulence reduced but not abolished in highly laboratory-passaged strains [71 , 72] . Other mouse model studies have shown that virulence of Bcc species can vary; for example , the epidemic B . cenocepacia strain J2315 caused universal mortality when inoculated at 103 cfu into gp91phox−/− mice via an intratracheal route , whereas other B . cenocepacia strains were less virulent and B . vietnamiensis strain R2 was avirulent [73] . Another study using intranasal inoculation of leukopaenic BALB/c mice with ~104 cfu also showed differential virulence within Bcc species , with some mice clearing their infections [74] , indicating that virulence potential varies among strains . Based on the findings of these earlier studies , pathogenicity may also vary among B . ubonensis strains . Characterising the virulence potential of other B . ubonensis strains may identify unusual pathogenic strains , although we deem this unlikely based on the lack of verified human infections caused by B . ubonensis . In consideration of the IACUC guidelines , we chose not to carry out testing of further strains using the mouse model . We acknowledge that our study only tested B . ubonensis in immunocompetent BALB/c mice via a sc route . The use of immunocompromised or immune-deficient mouse models or infection via different routes may reveal that B . ubonensis can cause disease in such cases . Bcc species carry various virulence factors that are thought to contribute to their pathogenic potential , including extracellular lipases , metalloproteases , serine proteases , flagella , pili , adhesins , toxins , siderophores and lipopolysaccharides [75] . We did not investigate the presence of virulence genes in B . ubonensis compared with other Bcc species but doing so may shed further light on its potential virulence capacity . It may be possible to use such in silico methods rather than further animal experiments to determine whether B . ubonensis is unusual compared with other Bcc species due to a lack of key virulence loci or pathways in its genome .
The metabolic diversity of Bcc species continues to spur interest in this highly adaptable group of bacteria . Our study provides important new insights into the biology of B . ubonensis , a largely neglected member of the Bcc due to its ostensibly avirulent nature . Genomic analysis of 264 B . ubonensis strains from Australia , PNG , Puerto Rico and Thailand revealed that B . ubonensis is a genetically highly diverse organism , with at least 26% of its chromosomal DNA variably present among strains . Like B . pseudomallei , B . ubonensis has a distinct phylogeographic signature that can be distinguished at the genomic level . It remains to be determined whether B . ubonensis is found on other continents . ‘Chromosome III’ encodes a ubiquitous yet apparently dispensable pC3 megaplasmid , similarly to other Bcc species , and can segregate in the laboratory setting . Like other Bcc species , we show that B . ubonensis strains exhibit variable levels of meropenem resistance . Determining the molecular mechanism underpinning high-level meropenem resistance in certain B . ubonensis strains will provide a better understanding of the potential transmission of this phenotype to the melioidosis bacterium B . pseudomallei , which frequently co-resides with B . ubonensis in the environment . Finally , using the immunocompetent BALB/c mouse model , we show that an Asian B . ubonensis strain is not likely to cause disease , providing evidence that at least some members of this species are probably avirulent in immunocompetent individuals . Further studies are needed to confirm the avirulent nature of B . ubonensis across a greater strain set using both immunocompetent and immunocompromised or immunodeficient animal models , or in silico analysis of the B . ubonensis genome to identify intact virulence determinants . The apparent non-pathogenic nature of certain B . ubonensis strains may make them amenable to large-scale biotechnological applications , such as biocontrol and biofuel production .
|
The pathogenic bacterium Burkholderia pseudomallei causes the disease melioidosis , which occurs in most tropical regions across the globe . The true burden of melioidosis is unknown but has been predicted to affect 165 , 000 people every year , resulting in 89 , 000 deaths . B . pseudomallei is easily confused with its close relative B . ubonensis as both species are frequently found in the same environmental niche and can appear phenotypically identical using serotyping and laboratory culture methods . B . ubonensis is a poorly characterised species but has recently gained interest in the research community as a potential biocontrol agent in B . pseudomallei-endemic regions , and for production of unusual and versatile biocompounds that are now being exploited for industrial applications . B . ubonensis is thought to be non-pathogenic , although other members of the B . cepacia complex to which it belongs are known for their ability to cause clinical disease that can be fatal in immunocompromised patients and people with cystic fibrosis . In this study , we investigated the biology of B . ubonensis to better understand its genetics , genomics , global distribution , virulence potential and antibiotic resistance . We show that this organism is highly genetically diverse , is avirulent in the mouse model , and can naturally encode high levels of meropenem resistance . We also identify B . ubonensis in the Caribbean for the first time , with phylogenomic analysis revealing distinct clades corresponding to geographic origin .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"discussion",
"Conclusions"
] |
[
"biogeography",
"taxonomy",
"ecology",
"and",
"environmental",
"sciences",
"medicine",
"and",
"health",
"sciences",
"burkholderia",
"pseudomallei",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"population",
"genetics",
"microbiology",
"ethnicities",
"animal",
"models",
"antibiotic",
"resistance",
"phylogenetics",
"data",
"management",
"model",
"organisms",
"phylogenetic",
"analysis",
"experimental",
"organism",
"systems",
"mammalian",
"genomics",
"population",
"biology",
"pharmacology",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"geography",
"computer",
"and",
"information",
"sciences",
"antimicrobial",
"resistance",
"medical",
"microbiology",
"microbial",
"pathogens",
"thai",
"people",
"mouse",
"models",
"phylogeography",
"evolutionary",
"systematics",
"animal",
"genomics",
"people",
"and",
"places",
"earth",
"sciences",
"genetics",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"population",
"groupings",
"burkholderia",
"genomics",
"evolutionary",
"biology",
"organisms"
] |
2017
|
Phylogeographic, genomic, and meropenem susceptibility analysis of Burkholderia ubonensis
|
Fungal biofilms are a major cause of human mortality and are recalcitrant to most treatments due to intrinsic drug resistance . These complex communities of multiple cell types form on indwelling medical devices and their eradication often requires surgical removal of infected devices . Here we implicate the molecular chaperone Hsp90 as a key regulator of biofilm dispersion and drug resistance . We previously established that in the leading human fungal pathogen , Candida albicans , Hsp90 enables the emergence and maintenance of drug resistance in planktonic conditions by stabilizing the protein phosphatase calcineurin and MAPK Mkc1 . Hsp90 also regulates temperature-dependent C . albicans morphogenesis through repression of cAMP-PKA signalling . Here we demonstrate that genetic depletion of Hsp90 reduced C . albicans biofilm growth and maturation in vitro and impaired dispersal of biofilm cells . Further , compromising Hsp90 function in vitro abrogated resistance of C . albicans biofilms to the most widely deployed class of antifungal drugs , the azoles . Depletion of Hsp90 led to reduction of calcineurin and Mkc1 in planktonic but not biofilm conditions , suggesting that Hsp90 regulates drug resistance through different mechanisms in these distinct cellular states . Reduction of Hsp90 levels led to a marked decrease in matrix glucan levels , providing a compelling mechanism through which Hsp90 might regulate biofilm azole resistance . Impairment of Hsp90 function genetically or pharmacologically transformed fluconazole from ineffectual to highly effective in eradicating biofilms in a rat venous catheter infection model . Finally , inhibition of Hsp90 reduced resistance of biofilms of the most lethal mould , Aspergillus fumigatus , to the newest class of antifungals to reach the clinic , the echinocandins . Thus , we establish a novel mechanism regulating biofilm drug resistance and dispersion and that targeting Hsp90 provides a much-needed strategy for improving clinical outcome in the treatment of biofilm infections .
In recent decades , fungal pathogens have emerged as a predominant cause of human disease , especially in immunocompromised individuals . The number of acquired fungal bloodstream infections has increased by ∼207% in this timeframe [1] , [2] , [3] . Although diverse species are capable of causing infection , a few prevail as the most prevalent cause of disease . Candida and Aspergillus species together account for ∼70% of all invasive fungal infections , with Candida albicans and Aspergillus fumigatus prevailing as the leading causal agents of opportunistic mycoses [2] . Candida species are the fourth leading cause of hospital acquired bloodstream infections in the United States with mortality rates estimated at 40% [4] , [5] . The profound economic consequences of Candida infections can be demonstrated by the ∼$1 . 7 billion spent annually on treating candidemia in the United States alone [6] . Further , A . fumigatus is the most common etiological agent of invasive aspergillosis , with a 40–90% mortality rate [7] . In patients with pulmonary disorders such as asthma or cystic fibrosis , A . fumigatus infection can cause allergic bronchopulmonary aspergillosis leading to severe complications . For these fungal species , there are numerous factors that contribute to the pathogenicity and recalcitrance of resulting infections to antifungal treatment , including the ability to evolve and maintain resistance to conventional antifungal therapy [1] . Due to the limited number of drug targets available to exploit in fungal pathogens that are absent or sufficiently divergent in the human host , the vast majority of antifungal drugs in clinical use target ergosterol or its biosynthesis . The azoles are the most widely used class of antifungal in the clinic and function by inhibiting the ergosterol biosynthetic enzyme Erg11 , causing a block in the production of ergosterol and the accumulation of the toxic byproduct 14-α-methyl-3 , 6-diol , culminating in a severe membrane stress [8] , [9] . The azoles are generally fungistatic against yeasts , including Candida species , and fungicidal against moulds , such as Aspergillus species . The fungistatic nature of the azoles towards C . albicans culminates in strong directional selection on the surviving population to evolve drug resistance [10] , [11] . In fact , high levels of azole resistance in C . albicans clinical isolates often accumulate through multiple mechanisms including: upregulation of drug efflux pumps , overexpression or alteration of Erg11 , or modification of stress response pathways that are crucial for resistance [1] , [10] , [11] , [12] , [13] . The echinocandins are the only new class of antifungal to reach the clinic in decades . They act as non-competitive inhibitors of β-1 , 3 glucan synthase , an enzyme involved in fungal cell wall synthesis [9] , resulting in the loss of cell wall integrity and a severe cell wall stress . The impact of the echinocandins is generally opposite to that of the azoles , in that they are fungicidal against yeasts and fungistatic against moulds . Resistance of C . albicans clinical isolates to the echinocandins has been reported and is often associated with mutations in the drug target [13] , [14] , [15] . An additional key factor responsible for the virulence and drug resistance of C . albicans and A . fumigatus is their tendency to form biofilms on medical devices that are highly resistant to antifungal treatment [16] , [17] , [18] , [19] , [20] . The use of such medical devices — such as venous catheters , urinary catheters and artificial joints — has dramatically risen to more than 10 million recipients per year [21] , [22] . This poses a severe clinical problem as C . albicans is the third leading cause of intravascular catheter-related infections , and has the overall highest crude mortality rate of ∼30% for device-associated infections [17] , [22] , [23] . Further , A . fumigatus infections have been reported on medical implant devices as well as on bronchial epithelial cells [17] , [24] . The inherent drug resistance of biofilms often necessitates surgical removal of the infected medical devices in order to eradicate the fungal infection . Extensive research has focused on mechanisms of drug resistance in C . albicans biofilms , and it is apparent that cells in a fungal biofilm represent an epigenetic modification of the cellular state compared to their planktonic counterparts , with changes in cellular morphology , cell-to-cell communication , and gene expression , as well as with the production of an extra-cellular matrix [16] , [18] , [20] . Multiple factors contribute to the elevated drug resistance of C . albicans biofilms . These factors include increased cell density [25] , increased expression of drug efflux pumps [26] , [27] , decreased ergosterol content [27] , elevated β-1 , 3 glucan levels in the cell wall and biofilm matrix [28] , [29] , as well as signalling mediated by protein kinase C ( PKC ) [30] and the protein phosphatase calcineurin [31] . The molecular chaperone Hsp90 regulates complex cellular circuitry in eukaryotes by stabilizing regulators of cellular signalling [32] , [33] . As a consequence , inhibiting Hsp90 disrupts a plethora of cellular processes and has broad therapeutic potential against diverse eukaryotic pathogens including the protozoan parasites Plasmodium falciparum and Trypanosoma evansi as well as numerous fungal species [34] , [35] , [36] . In the planktonic state , Hsp90 potentiates the emergence and maintenance of resistance to azoles and echinocandins in C . albicans at least in part via calcineurin [37]; Hsp90 physically interacts with the catalytic subunit of calcineurin , keeping it stable and poised for activation [38] . Recently , Hsp90 was also shown to enable azole and echinocandin resistance in C . albicans via the PKC cell wall integrity pathway [39] . Hsp90 depletion results in the destabilization of the terminal mitogen-activated protein kinase ( MAPK ) Mkc1 , providing the second Hsp90 client protein implicated in drug resistance [39] . Compromising C . albicans Hsp90 function renders drug-resistant isolates susceptible in vitro and improves the therapeutic efficacy of antifungals in a Galleria mellonella model of C . albicans pathogenesis and a murine model of disseminated candidiasis [34] . Compromising A . fumigatus Hsp90 also enhances the efficacy of echinocandins both in vitro and in the G . mellonella model of infection [34] . Notably , Hsp90 regulates not only drug resistance in C . albicans but also the morphogenetic transition between yeast and filamentous growth , a trait important for virulence [40] . Compromising Hsp90 function induces filamentation by relieving Hsp90-mediated repression of cAMP-protein kinase A ( PKA ) signalling [41] . The ability to transition between morphological states is also critical for biofilm formation and development [42] . Given that Hsp90 governs fungal morphogenesis and drug resistance in planktonic conditions , we sought to investigate if this molecular chaperone also regulates the development and drug resistance of biofilms . We discovered that genetically compromising Hsp90 function reduced but did not block biofilm maturation in vitro and had minimal impact on the ability of C . albicans to form robust biofilms in an in vivo rat catheter model , . Genetic depletion of C . albicans Hsp90 reduced biofilm dispersal , with the few dispersed cells being largely inviable . Moreover , compromising C . albicans Hsp90 function genetically or pharmacologically transformed the azole fluconazole from ineffectual to highly efficacious in eradicating biofilms both in vitro and in a rat catheter model of infection . In stark contrast to planktonic conditions , reduction of C . albicans Hsp90 levels genetically in biofilm conditions did not lead to depletion of the client proteins calcineurin or Mkc1 , suggesting that Hsp90 regulates drug resistance through distinct mechanisms in these different cellular states . Genetic depletion of Hsp90 reduced glucan levels in the biofilm matrix , providing a compelling mechanism by which Hsp90 might regulate biofilm drug resistance . Finally , in the most lethal mould , A . fumigatus , compromising Hsp90 function enhanced the efficacy of azoles and echinocandins in an in vitro model . Our results implicate Hsp90 as a novel regulator of biofilm dispersion and drug resistance , and provide strong support for the utility of Hsp90 inhibitors as a therapeutic strategy for biofilm infections caused by diverse fungal species .
Due to the key roles of Hsp90 in both morphogenesis and drug resistance under planktonic conditions [37] , [41] , we hypothesized that Hsp90 might also regulate C . albicans biofilm formation and drug resistance . First , we tested whether compromising Hsp90 function affected biofilm growth . To do this , C . albicans biofilms were cultured for 24 hours in static 96 well microtiter plates , washed to remove non-adherent cells , grown for an additional 24 hours with various concentrations of the Hsp90 inhibitor geldanamycin , and growth was quantified by metabolic activity using an XTT reduction assay [43] . The geldanamycin was added at 24 hours rather than at the initial time point as is the standard for biofilm drug studies since the initial cells are planktonic and much more susceptible to drugs than their biofilm counterparts [31] , [43]; consistent with this , initial attempts to include geldanamycin during inoculation led to a toxicity profile identical to that of planktonic cells ( data not shown ) . When geldanamycin was added at 24 hours , no significant differences in metabolic activity were observed at a variety of concentrations tested up to 100 µg/mL ( P>0 . 05 , ANOVA , Bonferroni's Multiple Comparison Test , Figure 1A ) . Thus , Hsp90 inhibitors do not compromise biofilm development . To further explore Hsp90's role in biofilm formation , we exploited a strain of C . albicans in which Hsp90 levels could be depleted by tetracycline-mediated transcriptional repression ( tetO-HSP90/hsp90Δ ) . Biofilms of the wild type and tetO-HSP90/hsp90Δ strain were cultured in static 96 well microtiter plates with or without 20 µg/mL of the tetracycline analog doxycycline from the time of inoculation . Doxycycline was included at this early point given the time required for transcriptional repression to manifest in depletion of Hsp90 , and enabled by the absence of toxicity in planktonic cells . Doxycycline-mediated transcriptional repression of Hsp90 decreased biofilm development , but did not block formation of a mature biofilm ( Figure 1B , P<0 . 01 ) . We observed comparable results when biofilms were cultured on silicon elastomer squares , and when biofilm growth was monitored by XTT reduction or by dry weight ( Figure S1A and B ) . To determine if depletion of Hsp90 prior to inoculation had a more profound effect on biofilm formation , we performed a comparable assay but in the presence or absence of doxycycline in the overnight culture . Depletion of Hsp90 prior to inoculation did not further reduce biofilm formation but rather led to a biofilm indistinguishable from the no doxycycline control ( Figure S1C ) . Although Hsp90 is essential , this dose of doxycycline causes reduced growth rate of the tetO-HSP90/hsp90Δ strain in planktonic cultures but has little effect on stationary phase cell density [41] . Western blot analysis validated that Hsp90 levels were dramatically reduced in biofilms formed by the tetO-HSP90/hsp90Δ strain when cultured in the presence of doxycycline ( Figure 1C ) . We note that when biofilms were formed under shaking conditions , the tetO-HSP90/hsp90Δ strain had reduced biofilm growth , which was exacerbated in the presence of doxycycline ( Figure S1D ) . Thus , while Hsp90's impact on biofilm development can vary , under most conditions tested compromising Hsp90 function does not block biofilm formation in vitro . In order to address the role of Hsp90 in biofilm growth in vivo , biofilm formation was examined using a rat venous catheter model of biofilm-associated candidiasis that mimics central venous catheters in patients [44] . Infection of implanted catheters with C . albicans was performed by intraluminal instillation , catheters were flushed after 6 hours , and biofilm formation was monitored with or without 20 µg/mL doxycycline after 24 hours . The tetO-HSP90/hsp90Δ strain was capable of establishing a biofilm in the rat venous catheter , as visualized by scanning electron microscopy ( Figure 2 ) . Further , transcriptional repression of HSP90 with doxycycline did not block the formation of a robust biofilm ( Figure 2 ) . These results demonstrate that compromising Hsp90 function does not impair the ability of C . albicans to form mature biofilms in vivo . As mentioned above , Hsp90 is a key regulator of the yeast to filament transition in C . albicans [41] , a process implicated in virulence and biofilm formation [42] . Therefore , we examined the architecture of geldanamycin treated biofilms cultured on silicon elastomer squares to enable imaging by confocal microscopy . Biofilms treated with geldanamycin had decreased thickness of the bottom yeast layer ( 30 µm and 45 µm versus 90 µm and 100 µm in the untreated control , P = 0 . 0237 , t-test ) without substantial change in the thickness of the upper layer of filaments ( Figure 3 ) . That a greater proportion of the biofilm thickness was occupied by filaments compared to yeast suggests that Hsp90 inhibition might lead to enhanced filamentation in biofilms . Moreover , biofilms treated with geldanamycin showed more polarized filaments extending away from the biofilm basal surface compared to the interconnected meshwork of filaments in an untreated control ( Figure 3 ) . That biofilms formed upon Hsp90 inhibition had a greater proportion of their total thickness occupied by filaments compared to yeast is consistent with Hsp90's repressive effect on filamentation in planktonic conditions . Based on our finding that C . albicans biofilms display altered morphologies upon Hsp90 inhibition , we sought to evaluate the effect of Hsp90 function on biofilm dispersion given that morphogenesis plays a critical role in this process [45] , [46] . We monitored dispersion of yeast cells using the only well validated model which involves culturing biofilms on silicon elastomer under conditions of flow [47] , [48] . When biofilms were cultured in the absence of doxycycline with the tetO-HSP90/hsp90Δ strain , the number of dispersed cells after 1 hour was 90 , 000 cells/mL and remained fairly constant over a 24 hour time period ( Figure 4A ) . In contrast , in the presence of 20 µg/mL doxycycline the number of dispersed cells was dramatically reduced to approximately 17 , 000 cells/mL throughout the 24 hours ( P = 0 . 0022 , t-test , Figure 4A ) . We confirmed that the effects of doxycycline were specifically due to transcriptional repression of HSP90 , as doxycycline had no impact on biofilm dispersal of the wild-type strain lacking the tetO promoter ( Figure S2A ) . Intriguingly , the cells that were dispersed upon reduction of Hsp90 levels had major viability defects compared to their untreated counterparts ( P = 0 . 007 , t-test ) with only 55% viable at 1 hour , 5% viable at 12 hours , and less than 1% viable at 24 hours ( Figure 4B ) . The dramatic reduction in viability was specific to the dispersed cell population with doxycycline-mediated transcriptional repression of HSP90 , as the viability of dispersed cells in the untreated control remained close to 50% even at 24 hours ( Figure 4B ) . Viability was unaffected when a wild-type strain lacking the tetO promoter was treated with doxycycline , confirming that the effects observed were due to transcriptional repression of HSP90 ( Figure S2B ) . The reduced viability upon reduction of Hsp90 levels was specific to the dispersed cell population within the biofilm , as there was only a minor defect in overall metabolic activity of the tetO-HSP90/hsp90Δ biofilms in the presence of doxycycline ( Figure 1B ) . Further , under planktonic conditions viability remained>85% when the tetO-HSP90/hsp90Δ strain was grown in the presence of doxycycline for 24 hours . Taken together , Hsp90 plays a critical role in the dispersal step of the biofilm life cycle and is crucial for survival of dispersed cells . Genetic or pharmacological compromise of Hsp90 function renders C . albicans susceptible to azoles and echinocandins under planktonic conditions [37] , [38] , [49] . Since compromising Hsp90 function pharmacologically did not impair biofilm maturation , we investigated whether inhibition of Hsp90 would alter biofilm drug resistance using the standard 96 well microtiter plate static assay that enables testing many drug concentrations . We focused on the azoles , since biofilms are notoriously resistant to this class of drugs , compromising their therapeutic utility [19] . As a positive control , a wild-type C . albicans biofilm was subjected to a gradient of concentrations of the calcineurin inhibitor FK506 in addition to a gradient of fluconazole , a drug combination with established synergistic activity against C . albicans biofilms [31] . We confirmed synergistic activity of FK506 with fluconazole by measuring metabolic activity using the XTT reduction assay ( Figure 5A ) . Biofilms were extremely susceptible to the combination of inhibitors with a calculated FIC index of 0 . 1093 , indicating potent synergy ( Table 1 ) . To determine if Hsp90 enables biofilm azole resistance , we used an equivalent experiment but with a gradient of concentrations of the Hsp90 inhibitor geldanamycin and a gradient of fluconazole . Geldanamycin exhibited potent synergy with fluconazole , dramatically reducing azole resistance at only 3 . 125 µg/mL geldanamycin . Maximal effects were observed with 12 . 5 µg/mL geldanamycin , which reduced the MIC50 of fluconazole from >1000 µg/mL to 31 . 25 µg/mL ( Figure 5A ) . Further , FIC indexes as low as 0 . 125 to 0 . 156 were calculated for the combination of fluconazole and geldanamycin confirming that inhibition of Hsp90 has a potent synergistic effect with azoles against C . albicans biofilms ( Table 1 ) . Next , we utilized the tetO-HSP90/hsp90Δ strain in order to validate that the synergistic activity of geldanamycin with fluconazole against C . albicans biofilms was indeed due to Hsp90 inhibition . Biofilms of a wild-type strain of C . albicans had a fluconazole MIC50 of over 512 µg/mL ( Figure 5B ) . Deletion of one allele of HSP90 or replacing the promoter of the sole remaining HSP90 allele with the tetracycline-repressible promoter had no impact on fluconazole resistance ( Figure 5B ) . However , upon depletion of Hsp90 by doxycycline-mediated transcriptional repression in the tetO-HSP90/hsp90Δ strain , the fluconazole MIC50 was dramatically reduced to only 8 µg/mL , a >60-fold increase in fluconazole sensitivity ( Figure 5B ) . Hence , both pharmacological and genetic evidence confirms that Hsp90 function is critical for azole resistance of C . albicans biofilms . To further dissect the mechanism by which Hsp90 regulates azole resistance of C . albicans biofilms , we repeated the drug susceptibility assay with strains lacking specific Hsp90 client proteins . Under planktonic conditions both calcineurin and Mkc1 are important Hsp90 client proteins that regulate the maintenance of azole resistance [37] , [38] , [39] . Moreover , these client proteins have previously been shown to be important for azole resistance of C . albicans biofilms [30] , [31] . We found that biofilms formed by strains lacking the catalytic subunit of calcineurin ( cna1Δ/cna1Δ ) or the terminal MAPK of the PKC cell wall integrity signalling pathway ( mkc1Δ/mkc1Δ ) had fluconazole MIC50 values of 32 µg/mL and 128 µg/mL , respectively; their fluconazole resistance levels were intermediate between the robust resistance of the wild-type parental strain and the sensitivity observed upon impairment of Hsp90 function ( Figure 5B ) . The finding that compromise of calcineurin function does not confer as severe a reduction in biofilm fluconazole resistance as compromise of Hsp90 function is intriguing in light of the fact that under all planktonic conditions tested , inhibition of calcineurin phenocopies inhibition of Hsp90 in terms of azole resistance [37] , [38] , [49] . These results suggest that calcineurin and Mkc1 may be able to partially compensate for the loss of the other client during times of azole-induced stress in a biofilm environment . Alternatively , these findings could be explained by the existence of a novel downstream effector of Hsp90 important for azole resistance of C . albicans biofilms . To further investigate the mechanisms by which Hsp90 regulates azole resistance in biofilm conditions , we examined protein levels of the client proteins calcineurin and Mkc1 upon Hsp90 depletion . Strains were cultured in RPMI medium for both planktonic and biofilm growth . Biofilms were cultured on plastic under static conditions , as with our drug studies . We previously established that under planktonic conditions genetic reduction of Hsp90 levels leads to depletion of the catalytic subunit of calcineurin ( Cna1 ) and Mkc1 [38] , [39] . Here , the tetO-HSP90/hsp90Δ strain was grown in either planktonic or biofilm conditions in the presence or absence of 20 µg/mL doxycycline for 48 hours . Under both conditions , Hsp90 levels were dramatically reduced in the presence of doxycycline ( Figure 6 ) . To monitor calcineurin levels , we used a C-terminal 6xHis-FLAG epitope tag on Cna1 in the tetO-HSP90/hsp90Δ strain . In the tagged strains , Cna1 levels were comparable under planktonic and biofilm conditions in the absence of doxycycline ( Figure 6A ) . Doxycycline-mediated reduction of Hsp90 levels led to an ∼90% reduction in Cna1 in planktonic conditions , however , Cna1 levels remained stable in biofilm conditions ( Figure 6A ) . All strains had comparable amounts of protein loaded , as confirmed with a tubulin loading control . To monitor total Mkc1 levels , we used a C-terminal 6xHis-FLAG epitope tag on Mkc1 in the tetO-HSP90/hsp90Δ strain . Mkc1 levels were comparable in the tagged strains under planktonic and biofilm conditions in the absence of doxycycline ( Figure 6B ) . As with Cna1 , doxycycline-mediated reduction of Hsp90 levels led to ∼80% reduction in Mkc1 levels in planktonic conditions , however , Mkc1 levels remained stable in biofilm conditions ( Figure 6B ) . We next addressed whether depletion of Hsp90 affected levels of activated , dually phosphorylated Mkc1 . Mkc1 was activated in all strains in the absence of doxycycline . As with total Mkc1 levels , doxycycline-mediated reduction of Hsp90 led to a reduction in levels of activated Mkc1 in planktonic conditions , however , Mkc1 remained activated in biofilm conditions . Taken together , these results suggest that Hsp90 may play different roles in client protein regulation in these distinct cellular states , and also that these client proteins may have other means of maintaining stability in a biofilm environment . Given our findings that Hsp90 client proteins remain stable in a biofilm , irrespective of Hsp90 levels , and that deletion of these client proteins does not phenocopy Hsp90 depletion in terms of biofilm azole resistance , we hypothesized that Hsp90 also regulates biofilm drug resistance through a mechanism independent of calcineurin and Mkc1 signalling . Recent studies established that glucan present in the biofilm matrix is critical for azole resistance due its capacity to sequester fluconazole , preventing it from reaching its intracellular target [29] . Consequently , we investigated whether Hsp90 affects glucan levels in the biofilm matrix . Biofilms were cultured on plastic in static conditions in the presence or absence of 20 µg/mL doxycycline for 48 hours , matrix material was harvested from biofilms with equivalent metabolic activity , and β-1 , 3 glucan levels were quantified . In the tetO-HSP90/hsp90Δ strain , the level of glucan in the biofilm matrix was ∼6 , 000 pg/mL in the absence of doxycycline ( Figure 7 ) . Transcriptional repression of HSP90 with 20 µg/mL doxycycline led to reduced glucan levels of only ∼3 , 700 pg/mL ( P<0 . 01 , ANOVA , Bonferroni's Multiple Comparison Test , Figure 7 ) . Doxycycline had no impact on matrix glucan levels of a wild-type strain lacking the tetO promoter , confirming that Hsp90 depletion leads to reduced glucan levels ( Figure 7 ) . The ∼40% reduction in matrix glucan upon Hsp90 depletion is likely to have made a major contribution to azole susceptibility , given that reduction of biofilm matrix glucan levels of ∼60% in an FKS1/fks1Δ mutant abrogates biofilm drug resistance [29] . These results provide the first link of Hsp90 to glucan production in C . albicans and mechanistic insight as to how Hsp90 regulates biofilm drug resistance . Due to the robust synergy observed between Hsp90 inhibition and fluconazole in vitro , we sought to address whether synergy was also observed in vivo in the rat venous catheter model of C . albicans biofilm infection using the tetO-HSP90/hsp90Δ strain . Addition of fluconazole alone ( 250 µg/mL ) after 24 hours of biofilm growth did not affect the biofilm formed by the tetO-HSP90/hsp90Δ strain ( Figure 8A ) . Doxycycline was delivered during both the biofilm formation and drug treatment phases , and also had no major effect on the biofilm formed by the tetO-HSP90/hsp90Δ strain ( Figure 2 ) . However , the combination of fluconazole and doxycycline destroyed the biofilm as observed by scanning electron microscopy ( Figure 8A ) . Thus , Hsp90 is required for the resistance of C . albicans biofilms to fluconazole in a mammalian host . In order to further explore the therapeutic potential of targeting Hsp90 for C . albicans biofilm infections in vivo , we explored the efficacy of combining fluconazole with an Hsp90 inhibitor structurally related to geldanamycin and in clinical development as an anti-cancer agent , 17- ( allylamino ) -17-demethoxygeldanamycin ( 17-AAG ) . Central venous rat catheters were infected with C . albicans and biofilm formation proceeded over a 24-hour period . At this point , fluconazole alone ( 250 µg/mL ) , 17-AAG alone ( 100 µg/mL ) , or the drug combination was instilled and allowed to dwell in the catheter for an additional 24 hours . Serial dilutions of the catheter fluid were then plated in order to assess viable colony forming units . We found that the combined drug treatment significantly reduced fungal burden compared to the individual drug treatments alone ( P<0 . 001 , ANOVA , Bonferroni's Multiple Comparison Test , Figure 8B ) . In fact , catheters from the animals undergoing the combination therapy were completely sterile ( Figure 8B ) . These experiments in a mammalian model provide compelling evidence that clinically relevant Hsp90 inhibitors may prove to be extremely valuable in combating C . albicans biofilm infections . We previously established that Hsp90 inhibitors increase the efficacy of the echinocandins against A . fumigatus under standard culture conditions [34] , motivating these studies to determine if Hsp90 inhibitors also affect drug resistance of A . fumigatus biofilms . After 24 hours of growth , A . fumigatus biofilms were subjected to a gradient of concentrations of the echinocandins caspofungin or micafungin , or the azoles voriconazole or fluconazole , in addition to a gradient of concentrations of the Hsp90 inhibitor geldanamycin in 96 well microtiter plates under static conditions . Metabolic activity was assessed using the XTT reduction assay after an additional 24 hours . The biofilms were completely resistant to all the antifungal drugs tested and geldanamycin individually , though the combination of geldanamycin with many of the antifungals was effective in reducing biofilm development . Geldanamycin displayed robust synergy with both caspofungin ( Figure 9A ) and micafungin ( Figure S3A ) , with an FIC value of 0 . 375 for both drugs ( Table 2 ) . Geldanamycin also enhanced voriconazole activity ( Figure 9A ) , with more potent effects observed when drugs were added to biofilms after only 8 hours of growth ( Figure S3B ) . Geldanamycin did not enhance the efficacy of fluconazole under any conditions tested ( data not shown ) . These patterns of drug synergy observed with A . fumigatus biofilms are consistent with those patterns observed with Aspergillus in planktonic conditions [37] . Next , given Hsp90's role in regulating fungal morphogenesis we explored the impact of drug treatment on morphology of A . fumigatus biofilms . Scanning electron microscopy revealed striking architectural changes of A . fumigatus biofilms upon drug treatment . The control biofilms appeared robust and healthy , however , upon Hsp90 inhibition increased hyphal and matrix production was observed ( Figure 9B ) . Treating biofilms with caspofungin alone resulted in minimal damage , however , the addition of both caspofungin and geldanamycin caused numerous burst and broken hyphae throughout the biofilm ( Figure 9B ) . Finally , voriconazole treatment resulted in a flat ribbon-like morphology , and the addition of geldanamycin induced further cell damage ( Figure 9B ) . Taken together , these results indicate that inhibition of Hsp90 induces changes in morphology of A . fumigatus biofilms , in addition to enhancing the efficacy of azoles and echinocandins against these otherwise recalcitrant cellular structures .
Our results establish a novel role for Hsp90 in dispersion and drug resistance of fungal biofilms , with profound therapeutic potential . Resistance of C . albicans biofilms to many antifungal drugs including the azoles , often necessitates surgical removal of the infected catheter or substrate demanding new therapeutic strategies . Here , we demonstrate that compromising the function of C . albicans Hsp90 blocks biofilm dispersal , potentially reducing their ability to serve as reservoirs for persistent infection ( Figure 4 ) . Further , we show that compromising Hsp90 function genetically or pharmacologically in C . albicans renders biofilms exquisitely susceptible to azoles , such that fluconazole is transformed from inefficacious to highly effective in destroying biofilms both in vitro ( Figure 5 and Table 1 ) and in a mammalian model of infection ( Figure 8 ) . Finally , in A . fumigatus we found that compromising Hsp90 function dramatically improves the efficacy of antifungals ( Figure 9 ) . Thus , inhibition of Hsp90 enhances the efficacy of antifungals against biofilms formed by the two leading fungal pathogens of humans separated by ∼1 billion years of evolution , suggesting that this combinatorial therapeutic strategy could have a broad spectrum of activity against diverse fungal pathogens . Hsp90 exerts pleiotropic effects on cellular circuitry in eukaryotes by stabilizing diverse regulators of cellular signalling [32] , [33] , [50] . Hsp90 regulates the temperature-dependent morphogenetic transition from yeast to filamentous growth in C . albicans , such that compromise of Hsp90 function by elevated temperature relieves Hsp90-mediated repression of Ras1-PKA signalling and induces filamentous growth [41] . While compromise of Hsp90 function could have impaired biofilm development by enhancing filamentous growth , we found negligible impact on biofilm development in vivo ( Figure 2 ) ; in vitro , compromise of Hsp90 function did reduce biofilm maturation under static conditions with more severe effects under shaking conditions ( Figures 1 and S1 ) . Biofilms formed in the presence of Hsp90 inhibitor had a greater proportion of their total thickness occupied by filaments compared to yeast ( Figure 3 ) , suggesting that Hsp90's role in repressing the yeast to filament transition in planktonic cells [41] is conserved in the biofilm state . Consequently , we investigated the impact of compromising Hsp90 function on dispersion , a stage of the biofilm life cycle intimately coupled to morphogenetic transitions , with the majority of dispersed cells being in the yeast form [45] , [46] . We found that compromising Hsp90 function dramatically reduces the dispersed cell population ( Figure 4 ) , consistent with previous findings with hyperfilamentous C . albicans mutants [45] , [46] . Strikingly , the majority of cells that disperse from biofilms with reduced levels of Hsp90 are inviable ( Figure 4 ) , which likely reflects an enhanced dependence of this cell population on Hsp90 . Given that the dispersed cell population is thought to be responsible for device-associated candidemia and the establishment of disseminated infection , inhibition of C . albicans Hsp90 function in individuals suffering from biofilm infections may assist in the prevention of the invasive forms of disease . In the broader sense , it is striking that depletion of Hsp90 blocks the production of yeast in C . albicans in planktonic conditions [41] as well as throughout the biofilm lifecycle , creating a constitutively filamentous program characteristic of the strictly filamentous lifestyle of the vast majority of fungi . Hsp90 potentiates the emergence and maintenance of C . albicans drug resistance through multiple client proteins . A key mediator of Hsp90-dependent drug resistance is the protein phosphatase calcineurin [37] , [38] , [49] . In planktonic cells , Hsp90 stabilizes the catalytic subunit of calcineurin , Cna1 , thereby enabling calcineurin-dependent cellular signalling required for survival of drug-induced cellular stress [38] . Hsp90 also regulates drug resistance by stabilizing the MAPK Mkc1 , thereby enabling additional stress responses important for resistance [39] . In planktonic conditions , inhibition of calcineurin phenocopies inhibition of Hsp90 reducing drug resistance of diverse mutants , though deletion of MKC1 has a less severe effect on resistance under specific conditions [37] , [38] , [39] . In biofilms , homozygous deletion of either CNA1 or MKC1 causes an intermediate increase in sensitivity to azoles compared to reduction of HSP90 levels ( Figure 5 ) . Genetic depletion of Hsp90 reduces the fluconazole MIC50 from >512 µg/mL to 8 µg/mL , whereas deletion of CNA1 reduces resistance to 32 µg/mL and deletion of MKC1 reduces resistance only to 128 µg/mL ( Figure 5 ) . Thus , both calcineurin and Mkc1 have reduced impact on azole resistance of biofilms compared to Hsp90 , suggesting differences in the Hsp90-dependent cellular circuitry between the biofilm and planktonic cellular states . Hsp90 regulates circuitry required for fungal drug resistance largely by stabilizing key regulators of cellular signalling . In planktonic conditions , reduction of Hsp90 levels leads to depletion of both Cna1 and Mkc1 [38] , [39] . In stark contrast , Cna1 and Mkc1 remain stable in biofilms , despite reduction of Hsp90 levels ( Figure 6 ) . In both planktonic and biofilm conditions , Hsp90 levels were reduced by doxycycline-mediated transcriptional repression in the tetO-HSP90-hsp90Δ strain and levels of Hsp90 were reduced sufficiently to abrogate drug resistance in both conditions . The reduced dependence of Cna1 and Mkc1 on Hsp90 in biofilms suggests that these proteins have altered stability in this cellular state . These Hsp90 client proteins may assume an alternate conformation in biofilms that is inherently more stable , or they may interact with other proteins or chaperones that confer increased stability and reduced dependence upon Hsp90 . Consistent with the possibility of altered chaperone balance in biofilm cells , the Hsp70 family member SSB1 is overexpressed six-fold in biofilms compared to their planktonic counterparts [51] . While it is possible that Hsp90 may still regulate Cna1 and Mkc1 function through a mechanism distinct from protein stability , we note that Mkc1 is still activated upon Hsp90 depletion in biofilms ( Figure 6 ) . Given Hsp90's high degree of connectivity in diverse signalling cascades , it could also affect biofilm drug resistance in a multitude of other ways , such as by regulating remodeling of the cell wall and cell membrane [27] , [28] , signalling cascades important for matrix production [29] , [52] , or the function of contact-dependent signalling molecules that initiate responses to surfaces [30] . Future studies will determine on a more global scale the impact of cellular state on Hsp90 client protein stability , and the complex circuitry by which Hsp90 regulates biofilm drug resistance . Our results suggest that Hsp90 is a novel regulator of matrix glucan levels . For C . albicans the reduction in matrix glucan levels upon Hsp90 depletion provides a mechanism by which Hsp90 might govern biofilm azole resistance . C . albicans biofilms possess elevated cell wall β-1 , 3 glucan content compared to their planktonic counterparts [28] , and matrix glucan sequesters fluconazole , preventing it from reaching its intracellular target [28] , [29] . The ∼40% reduction in matrix glucan we observed upon Hsp90 depletion ( Figure 7 ) likely contributes to reduced azole resistance , given that a reduction of matrix glucan levels of ∼60% in an FKS1/fks1Δ mutant abrogates biofilm drug resistance [29] . Hsp90 could regulate glucan levels by directly or indirectly affecting β-1 , 3 glucan synthase , Fks1 , a protein important for the production of matrix glucan and for antifungal resistance [28] , [29] . Alternatively , Hsp90 could regulate matrix production by directly or indirectly affecting Zap1 , or its downstream targets Gca1 and Gca2 , which play an important role in matrix production , likely through the hydrolytic release of β-glucan fragments from the environment [52] . We note that in A . fumigatus , inhibition of Hsp90 appears to increase matrix production ( Figure 9 ) , though glucan levels remain unknown . Future studies will dissect the molecular mechanisms by which Hsp90 regulates biofilm matrix production and if there is divergent circuitry between these fungal pathogens . This work establishes that targeting Hsp90 may provide a powerful therapeutic strategy for biofilm infections caused by the leading fungal pathogens of humans . Compromising Hsp90 function genetically or pharmacologically reduces azole resistance of C . albicans biofilms both in vitro and in the rat venous catheter model of infection ( Figures 5 and 8 ) . Importantly , inhibition of Hsp90 with 17-AAG , an Hsp90 inhibitor that has advanced in clinical trials for the treatment of cancer [53] , [54] and is synergistic with antifungals in planktonic conditions [34] , transforms fluconazole from ineffective to highly efficacious in a mammalian model of biofilm infection ( Figure 8 ) . There may in fact be a multitude of benefits of inhibiting Hsp90 in the context of C . albicans biofilm infections given a recent report that treatment of in vitro C . albicans biofilms with voriconazole induces resistance to micafungin in an Hsp90-dependent manner [55] . The therapeutic potential of Hsp90 inhibitors against fungal biofilms extends beyond C . albicans to the most lethal mould , A . fumigatus . Pharmacological inhibition of Hsp90 enhances the efficacy of both azoles and echinocandins against A . fumigatus biofilms ( Figure 9 ) . The synergy between Hsp90 inhibitors and echinocandins is more pronounced than that with azoles , consistent with findings in the planktonic cellular state [34] . Thus , targeting Hsp90 may provide a much-needed strategy to enhance the efficacy of antifungal drugs against biofilms formed by diverse fungal pathogens . Our results provide a new facet to the broader therapeutic paradigm of Hsp90 inhibitors in the treatment of infectious disease caused by fungi and other pathogenic eukaryotes . In addition to the profound effects on biofilm drug resistance and dispersal , compromising Hsp90 function enhances the efficacy of azoles and echinocandins against disseminated disease caused by the leading fungal pathogens of humans in invertebrate and mammalian models of infection [34] , [38] . Beyond enhancing antifungal activity , Hsp90 also provides an attractive antifungal target on its own given that depletion of fungal Hsp90 results in complete clearance of a kidney fungal burden in a mouse model of disseminated candidiasis [41] . Hsp90 inhibitors also exhibit potent activity against malaria and Trypanosoma infections , thus extending their spectrum of activity to the protozoan parasites Plasmodium falciparum and Trypanosoma evansi [35] , [36] . The development of Hsp90 as a therapeutic target for infectious disease may benefit from the plethora of structurally diverse Hsp90 inhibitors that have been developed , many of which are in advanced phase clinical trials for cancer treatment , with substantial promise due to the depletion of a myriad of oncoproteins upon inhibition of Hsp90 [56] . Given the importance of Hsp90 in chaperoning key regulators of cellular signalling in all eukaryotes , the challenge of advancing Hsp90 as a target for infectious disease lies in avoiding host toxicity issues . Indeed , although well tolerated in the mammalian host individually or in combination therapies [56] , Hsp90 inhibitors have toxicity in the context of an acute disseminated fungal infection [34] . This toxicity may be due to Hsp90's role in regulating host immune and stress responses during infection . Toxicity was not observed in our studies of biofilm infections in the mammalian model , perhaps owing to both the localized infection and drug delivery , suggesting that this therapeutic strategy could rapidly translate from the laboratory bench to the patients' bedside . In the broader context , the challenge for further development of Hsp90 as a therapeutic target for infectious disease lies in developing pathogen-selective inhibitors or drugs that target pathogen-specific components of the Hsp90 circuitry governing drug resistance and virulence .
All procedures were approved by the Institutional Animal Care and Use Committee ( IACUC ) at the University of Wisconsin according to the guidelines of the Animal Welfare Act , The Institute of Laboratory Animal Resources Guide for the Care and Use of Laboratory Animals , and Public Health Service Policy . Archives of C . albicans strains were maintained at −80°C in 25% glycerol . Strains were routinely maintained and grown in YPD liquid medium ( 1% yeast extract , 2% bactopeptone , 2% glucose ) at 30°C . Strains used in this study are listed in Table S1 . Strain construction is described in the Supplemental Material . Multiple in vitro assays were used to assess C . albicans biofilm growth and antifungal drug susceptibility . In the first model , biofilms were developed in 96-well polystyrene plates , as previously described [28] , [43] . Briefly , strains were grown overnight in YPD at 37°C . Subsequently , cultures were resuspended in RPMI medium buffered with HEPES or MOPS , in the presence or absence of doxycycline ( 631311 , BD Biosciences ) to a final concentration of 106 cells/mL . An aliquot of 100 µl was added to each well of a 96-well flat-bottom plate , followed by incubation at 37°C . After 24 hours , the wells were gently washed twice with phosphate-buffered saline ( PBS ) to remove non-adherent cells , and fresh medium was added with or without a gradient of geldanamycin ( ant-gl-5 , Cedarlane ) . After 24 hours , non-adherent cells were washed away with PBS and biofilm cell metabolic activity was measured using the XTT reduction assay as previously described [28] , [43] . Briefly , 90 µl of XTT ( X4251 , Sigma ) at 1 mg/mL and 10 µl phenazine methosulfate ( P9625 , Sigma ) at 320 µg/mL were added to each well , followed by incubation at 37°C for 2 hours . Absorbance of the supernatant transferred to a fresh plate was measured at 490 nm using an automated plate reader , and experiments were carried out in a minimum of 5 replicates for each strain . In the second model , biofilms were developed on silicon elastomer ( SE ) surfaces as has been described previously [57] . C . albicans wild-type cells were grown overnight in YPD medium at 30°C and diluted to an optical density at 600 nm of 0 . 5 in RPMI medium . The suspension was added to a sterile 12-well plate containing bovine serum ( B-9433 , Sigma ) -treated SE ( Cardiovascular Instrument silicon sheets; PR72034-06N ) and incubated at 37°C for 90 min at 150 rpm agitation for initial adhesion . The SE were washed with PBS , transferred to fresh plates containing either fresh RPMI medium in the absence of drug , or RPMI with 10 µg/mL geldanamycin or 20 µg/mL doxycycline . Plates were incubated at 37°C for 48 hours at 150 rpm agitation to allow biofilm formation , followed by visualization by microscopy or by monitoring biofilm growth by XTT reduction or dry weight , as previously described [31] , [43] . For obtaining cells dispersed from biofilms , C . albicans biofilms were cultured in a simple flow biofilm model , as described previously [47] , [48] . Briefly , this model involves a controlled flow of fresh medium via Tygon tubing ( Cole-Parmer , Vernon Hills , IL ) into a 15 mL polypropylene conical tube ( BD , Franklin , NJ ) holding a SE strip . Medium flow is controlled at 1 mL/minute , by connecting the tubing to a peristaltic pump ( Masterflex L/S Easy-Load II , Cole-Parmer ) . The whole apparatus is placed inside an incubator to facilitate biofilm development at 37°C . SE strips ( 1×9 cm , Cardiovascular instrument Corp , Wakefield , MA ) , were sterilized by autoclaving and pre-treated for 24 hours with bovine serum . C . albicans was grown overnight at 30°C , washed , and diluted to an optical density at 600 nm of 0 . 5 in Yeast Nitrogen base ( YNB ) medium ( BD Biosciences , San Jose , CA ) with 50 mM glucose . The SE strips were incubated with the diluted C . albicans suspension at 37°C for 90 min at 100 rpm agitation for the initial adhesion of cells . Next , the strip was inserted into the conical tube and the peristaltic pump was turned on . At various time points during biofilm development , cells released from the biofilm in the flow-through were collected from the bottom of the conical tube . The dispersed cells were enumerated by a hemocytometer to obtain cell counts and there were no differences observed in the degree of clumping or morphological state of the dispersed cells , which were in the yeast form . Viability of the dispersed cells was assessed by plating and by colony counts on YPD agar . Drug susceptibility assays were performed on biofilms formed in wells of 96-well plates . Fresh medium ( RPMI/HEPES ) and drugs were added to wells containing biofilms grown for 24 hours . Dilutions of fluconazole ( Sequoia Research Products ) were from 1000 µg/ml down to 0 with the following concentration steps in µg/ml: 1000 , 500 , 250 , 125 , 62 . 5 , 31 . 25 , 15 . 625 , 7 . 8125 , 3 . 90625 , 1 . 953125 , 0 . 9765625 . FK506 ( AG Scientific ) gradients were from 75 µg/mL down to 0 with the following concentration steps in µg/ml: 75 , 37 . 5 , 18 . 75 , 9 . 375 , 4 . 6875 , 2 . 3475 , 1 . 171875 . Geldanamycin gradients were from 100 µg/mL to 0 with the following concentration steps in µg/ml: 100 , 50 , 25 , 12 . 5 , 6 . 25 , 3 . 125 , 1 . 5625 . Drug combinations were examined alone or in combination in a checkerboard format . After incubation at 37°C for 24 hours , biofilms were washed twice with PBS and metabolic activity was measured using the XTT assay , as described above . The drug concentration associated with 50% reduction in optical density compared to the drug-free control wells ( MIC50 ) was determined . The fractional inhibitory concentration ( FIC ) was calculated as follows: [ ( MIC 50 of drug A in combination ) / ( MIC 50 of drug A alone ) ] + [ ( MIC50 of drug B in combination ) / ( MIC50 of drug B alone ) ] . Values of <0 . 5 indicates synergy , those of >0 . 5 but <2 indicate no interaction , and those of >2 show antagonism [28] . In order to evaluate biofilm formation in vivo , a rat central venous catheter infection model was employed [44] . Specific-pathogen-free Sprague-Dawley rats weighing ∼400 g were used ( Harlan Sprague-Dawley , Indianapolis , IN ) . A heparinized ( 100 U/mL ) polyethylene catheter was surgically inserted into the jugular vein and advanced 2 cm to a site above the right atrium . After the catheter was secured to the vein , the proximal end was tunneled subcutaneously to the midscapular space and externalized through the skin . The catheters were implanted 24 hours prior to inoculation with C . albicans to allow a conditioning period for deposition of host protein on the catheter surface . Infection was performed by intraluminal instillation of 500 µl of C . albicans ( 106 cells/mL ) . After 6 hours , the catheters were flushed and maintained with heparinized 0 . 85% NaCl for 24 hours to allow for biofilm formation . While one end of the catheter is open to the venous blood , most of the fluid contents remain within the catheter unless pushed into the bloodstream with additional fluid from the external end . For drug treatment studies , fluconazole ( 250 µg/mL ) , 17-AAG ( A-6880 , LC Laboratories , 100 µg/mL ) , or saline was instilled and allowed to dwell in the catheter for an additional 24 hours [28] . For doxycycline studies , doxycycline ( 20 µg/mL ) was delivered during both the biofilm formation and the drug treatment phases . At the end of the observation period , the animals were sacrificed and the catheters were removed . In order to quantify fungal biofilm formation in the catheter , the contents were drained to remove blood and non-adherent organisms . The distal 2 cm of catheter was cut from the entire catheter length and the segment was placed in 1 mL of 0 . 85% NaCl . Following sonication for 10 minutes ( FS 14 water bath sonicator and 40-kHz transducer [Fisher Scientific] ) and vigorous vortexing for 30 seconds , serial dilutions of the catheter fluid were plated on Sabouraud Dextrose Agar ( SDA ) for viable fungal colony counts . Results are expressed as the mean colony forming unit ( CFU ) per milliliter . Aspergillus fumigatus Af293 was maintained on SDA slopes at 4°C . For conidial preparation Af293 was propagated on SAB agar for 72 hours and conidia harvested in PBS containing 0 . 025% ( v/v ) Tween 20 and quantified as previously described [58] . Commercially available voriconazole ( Pfizer Pharmaceuticals , NY , USA ) , micafungin ( Astellas Pharma Inc , Ibaraki , Japan ) and caspofungin ( Merck Sharp Dohme Ltd , NJ , USA ) were used throughout this study . Each antifungal drug was prepared at stock concentrations of 10 mg/mL in sterile water and used within 24 hours of reconstitution . Af293 conidial inoculum ( 1×105 conidia/mL ) was dispensed into flat bottomed 96-well microtitre plates and incubated for 8 or 24 hours at 37°C as previously described [58] . Biofilms were gently washed twice with PBS and each antifungal agent and geldanamycin were diluted to working concentrations in RPMI , which were tested either alone or in combination in a checkerboard format . Antifungal agent dilutions were from 512 µg/ml down to 0 with the following concentration steps in µg/ml: 512 , 256 , 128 , 64 , 32 , 16 , 8 , 4 , 2 , 1 , 0 . 5 . Geldanamycin dilutions were from 100 µg/ml down to 0 with the following concentration steps in µg/ml: 100 , 25 , 12 . 5 , 6 . 25 , 3 . 125 , 1 . 5625 . The biofilms were then treated and processed as described for C . albicans . Biofilms were stained with 25 µg/mL concanavalin A–Alexa Fluor 594 conjugate ( C-11253; Molecular Probes , Eugene , OR ) for 1 hour in the dark at 37°C . Confocal scanning laser microscopy ( CSLM ) was performed with a ZeissLSM 510 upright confocal microscope using a Zeiss Achroplan 40X , 0 . 8-W objective . Stained biofilms were observed using a HeNe1 laser with an excitation wavelength of 543 nm . The Zeiss LSM Image Browser v4 . 2 software was used to assemble images into side and depth views . Artificially coloured depth view images represent cell depth using a colour gradient , where cells closest to the SE are represented in blue and the cells farthest away are represented in red . Biofilms formed in vitro were placed overnight in a fixative ( 4% formaldehyde v/v , 1% glutaraldehyde v/v in PBS ) , rinsed in 0 . 1 M phosphate buffer and air dried in desiccators . Notably , harsh dehydration steps were not performed to minimize the damage to the original biofilm structure . The samples were coated with gold/palladium ( 40%/60% ) and observed under a scanning electron microscope ( Leo 435 VP ) in high vacuum mode at 15 kV . The images were assembled using Photoshop software ( Adobe , Mountain View , CA . ) . Catheter segments were processed for scanning electron microscopy as previously described [44] . Following overnight fixation ( 4% formaldehyde , 1% glutaraldehyde in PBS ) , catheter segments were washed with PBS and treated with osmium tetroxide ( 1% in PBS ) for 30 minutes . Drying was accomplished using a series of alcohol washes followed by critical point drying . Catheter segments were mounted and gold coated . Images were obtained with a scanning electron microscope ( JEOL JSM-6100 ) in the high-vacuum mode at 10 kV . The images were assembled using Adobe Photoshop 7 . 0 . 1 . For the protein stability assay , planktonic cultures were grown in RPMI buffered with MOPS and treated as described previously [39] . For biofilm cultures , C . albicans was grown overnight in YPD medium at 30°C and diluted to an optical density at 600 nm of 0 . 5 in RPMI medium . The suspension was added to a bovine serum ( 16190; Gibco ) -treated sterile 6-well plate and incubated at 37°C for 90 minutes for initial adhesion . The plates were washed with PBS , and fresh RPMI medium was added with or without 20 µg/mL doxycycline . Plates were incubated at 37°C for 48 hours . Cells were harvested by centrifugation and were washed with sterile water . Cell pellets were resuspended in lysis buffer containing 50 mM HEPES pH 7 . 4 , 150 mM NaCl , 5 mM EDTA , 1% Triton X-100 , 1 mM PMSF , and protease inhibitor cocktail ( complete , EDTA-free tablet , Roche Diagnostics ) . Cells suspended in lysis buffer were mechanically disrupted by adding acid-washed glass beads and bead beating for 3 minutes . Protein concentrations were determined by Bradford analysis . Protein samples were mixed with one-sixth volume of 6X sample buffer containing 0 . 35 M Tris-HCl , 10% ( w/v ) SDS , 36% glycerol , 5% β-mercaptoethanol , and 0 . 012% bromophenol blue for SDS-PAGE . Samples were boiled for 5 minutes and then separated by SDS-PAGE using an 8% acrylamide gel . Proteins were electrotransferred to PVDF membranes ( Bio-Rad Laboratories , Inc . ) and blocked with 5% skimmed milk in phosphate buffered saline ( PBS ) with 0 . 1% tween . Blots were hybridized with antibodies against CaHsp90 ( 1∶10000 ) , generously provided by Brian Larsen [59] , FLAG ( 1∶10000 , Sigma Aldrich Co . ) , His6 ( 1∶10 , P5A11 , generously provided by Elizabeth Wayner ) , phospho-p44/42 MAPK ( Thr202/Tyr204 ) ( 1∶2000 , Cell Signaling ) , or against alpha-tubulin ( 1∶1000; AbD Serotec , MCA78G ) . Matrix β-1 , 3 glucan content was measured using a limulus lysate based assay , as previously described [28] , [60] . Matrix was collected from C . albicans biofilms growing in the wells of 6-well polystyrene plates with or without 20 µg/mL doxycycline for 48 hours . The method for culturing biofilms was as described above for the immune blot analysis with the exception that all reagents were glucan-free . Biofilms were dislodged using a sterile spatula , washed with PBS , sonicated for 10 minutes , and centrifuged 3 times at 4500 x g for 20 minutes to separate cells from soluble matrix material [28] , [61] . Samples were stored at -20°C and glucan concentrations were determined using the Glucatell ( 1 , 3 ) -Beta-D-Glucan Detection Reagent Kit ( Associates of Cape Cod , MA ) as per the manufacturer's directions . C . albicans: PKC1 ( 3635298 ) ; HSP90 ( 3637507 ) ; CNA1 ( 3639406 ) ; CNB1 ( 3636463 ) ; MKC1 ( 3639710 ) ; ERG11 ( 3641571 ) ; FKS1 ( 3637073 ) ; SSB1 ( 3642206 ) ; GCA1 ( 3635124 ) ; ZAP1 ( 3641162 ) .
|
Candida albicans and Aspergillus fumigatus are the most common causative agents of fungal infections worldwide . Both species can form biofilms on host tissues and indwelling medical devices that are highly resistant to antifungal treatment . Here we implicate the molecular chaperone Hsp90 as a key regulator of biofilm dispersion and drug resistance . Compromising Hsp90 function reduced biofilm formation of C . albicans in vitro and impaired dispersal of biofilm cells , potentially blocking their capacity to serve as reservoirs for infection . Further , compromise of Hsp90 function abrogated resistance of C . albicans biofilms to the most widely deployed class of antifungal , the azoles , both in vitro and in a mammalian model of catheter-associated candidiasis . Key drug resistance regulators were depleted upon reduction of Hsp90 levels in planktonic but not biofilm conditions , suggesting that Hsp90 regulates drug resistance through different mechanisms in these distinct cellular states . Reduction of Hsp90 markedly reduced levels of matrix glucan , a carbohydrate important for C . albicans biofilm drug resistance . Inhibition of Hsp90 also reduced resistance of A . fumigatus biofilms to the newest class of antifungal , the echinocandins . Thus , targeting Hsp90 provides a promising strategy for the treatment of biofilm infections caused by diverse fungal species .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"infectious",
"diseases",
"biology",
"microbiology"
] |
2011
|
Hsp90 Governs Dispersion and Drug Resistance of Fungal Biofilms
|
We have developed an efficient method to quantify cell-to-cell infection with single-cycle , replication dependent reporter vectors . This system was used to examine the mechanisms of infection with HTLV-1 and HIV-1 vectors in lymphocyte cell lines . Effector cells transfected with reporter vector , packaging vector , and Env expression plasmid produced virus-like particles that transduced reporter gene activity into cocultured target cells with zero background . Reporter gene expression was detected exclusively in target cells and required an Env-expression plasmid and a viral packaging vector , which provided essential structural and enzymatic proteins for virus replication . Cell-cell fusion did not contribute to infection , as reporter protein was rarely detected in syncytia . Coculture of transfected Jurkat T cells and target Raji/CD4 B cells enhanced HIV-1 infection two fold and HTLV-1 infection ten thousand fold in comparison with cell-free infection of Raji/CD4 cells . Agents that interfere with actin and tubulin polymerization strongly inhibited HTLV-1 and modestly decreased HIV-1 cell-to-cell infection , an indication that cytoskeletal remodeling was more important for HTLV-1 transmission . Time course studies showed that HTLV-1 transmission occurred very rapidly after cell mixing , whereas slower kinetics of HIV-1 coculture infection implies a different mechanism of infectious transmission . HTLV-1 Tax was demonstrated to play an important role in altering cell-cell interactions that enhance virus infection and replication . Interestingly , superantigen-induced synapses between Jurkat cells and Raji/CD4 cells did not enhance infection for either HTLV-1 or HIV-1 . In general , the dependence on cell-to-cell infection was determined by the virus , the effector and target cell types , and by the nature of the cell-cell interaction .
Retroviruses can infect cells as cell-free particles or by cell-to-cell transmission [1] , [2] , [3] , [4] , [5] . In the latter route of infection , specific cell-cell contacts may strongly enhance virus infection by triggering the reorganization of cytoskeletal and cell-surface protein networks to focus virus release toward clustered receptors on an apposed target cell [2] , [6] , [7] , [8] , [9] , [10] , [11] , [12] . Cell-to-cell infection would require steps in the virus infectious cycle to be integrated with events in the cell-cell adhesion process; hence , the mechanism of cell-to-cell transmission would depend on specific interactions between cell and virus proteins . HTLV-1 is a highly cell-associated virus that is most likely disseminated by cell-to-cell transmission in vivo [13] . Microscopic image analysis of HTLV-1-infected lymphocytes in close contact with uninfected cells in vitro showed aggregation and transfer of virus components at a “virological synapse” ( VS ) [14] , [15]; whether the transfer of viral proteins between cells was accompanied by provirus formation is still unknown . On the other hand , HIV-1 infection has been studied intensively , and in vitro systems have been used to examine cell-to-cell transmission of virus from infected T-cells or infected macrophages to uninfected T-cells and epithelial cells [2] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [16] , [17] , [18] , as well as the special situation where HIV-1 particles are collected on dendritic cells and transmitted to T-cells via an “infectious synapse” [19] , [20] , [21] . MLV has been shown to move between cells along filipodial bridges that connect infected and uninfected cells not in immediate proximity [22] . More recently , a variety of cell-cell communicative structures such as nanotubes , mono- and polysynapses , have been demonstrated to serve as platforms for directed HIV particle egress , transfer and endocytosis by target cells [23] . In sum , our knowledge of retrovirus biology combined with in vitro experimental data suggests that cell-to-cell transmission is an important mechanism of virus spread in vivo . Much of what we know about cell-to-cell infection is inferred from microscopic image analysis; fluorescent microscopy shows viral proteins mobilized to cell-cell contact sites and electron micrographs show virus particles localized between interacting cells [2] , [4] , [10] , [14] , [16] , [24] . Direct evidence for virus replication in the context of cell-to-cell transmission has been reported for HIV-1 by measuring reverse transcription products in infected target cells or by FACS analysis of HIV-1 protein expression in fluorescently labeled target cells [6] , [7] , [11] , [12] , [25] or by long-term video microscopy observation of HIV Gag-iGFP replication [26] . Retroviral vectors , which have greatly facilitated studies of cell-free infection , are not well suited for examining cell-to-cell infection . One problem is that reporter gene expression in the producer cells generates a strong signal , and because these cells cannot be removed entirely from newly infected target cells , they obscure infection events . Clearly , a simple method to quantify cell-to-cell infection would provide a needed functional complement to image analysis , and help to define mechanisms of cell-to-cell infection . We have solved these technical problems by constructing HIV-1 and HTLV-1 vectors that consist of a virus packaging plasmid , an Env-expression plasmid , and a replication dependent reporter vector . The design of the new reporter vectors is based on a concept described initially by Heidmann et al . [27] , which was later adapted to study retrotransposition of endogenous retroviruses [28] , mammalian LINE1 elements [29] , and yeast TY1 elements [30] . We demonstrate here that the new HIV-1 and HTLV-1 reporter vectors are ideally suited for studying cell-to-cell infection , as reporter protein expression is confined exclusively to the infected target cell . By using eYFP based transfer vectors , we track infected cells that are not multinucleated and thereby rule out cell fusion as a mechanism of viral transmission . Quantifying the transduction of the new luciferase based vectors in target cells , we show the absolute dependence of HTLV-1 transmission on cell-cell contact , cytoskeleton remodeling and Tax protein expression , while HIV infection is enhanced twofold in our cell coculture settings and has characteristics both of cell-free and cell-to-cell modes of transmission . Induction of an immunological synapse ( IS ) between Jurkat effector cells and Raji/CD4 target cells does not increase infection with either HIV-1 or HTLV-1 VLPs suggesting that cell-to-cell infection requires the formation of specialized VS .
In order to take advantage of the sensitivity and versatility of retroviral vectors for studying virus replication , and to overcome difficulties encountered with standard reporter vectors in coculture infection experiments , we constructed reporter vectors similar to those that have been used to study retrotransposition of various mobile genetic elements [28] , [29] , [30] . The new HTLV-1 and HIV-1 reporter vectors contain a reporter gene cassette in antisense orientation relative to the virus; the reporter gene is interrupted by an intron , which is oriented in the sense direction ( Fig . 1A ) . The intron , which can be spliced only from the vector mRNA , prevents expression of the reporter gene in transfected effector cells ( Fig . 1B ) . Virus-like particles ( VLPs ) are produced after transfection of effector cells with reporter vector and virus packaging vector . Infection of target cells with VLPs that contain the spliced reporter vector RNA will generate a provirus that is now capable of expressing the reporter protein ( Figure 1B ) . HTLV-1 and HIV-1 reporter vectors were made that encode either luciferase ( inLuc ) or yellow fluorescent protein ( inYFP ) genes . When transfected alone into 293T cells ( not shown ) or Jurkat cells ( Fig . 2A ) , the inLuc vectors did not express detectable levels of luciferase activity . VLPs produced from cells transfected with either HTLV-1 or HIV-1 vectors contain reporter vector mRNA of which approximately 35% has no intron ( data not shown ) . To validate these vectors and to examine the mechanisms of retrovirus transmission between cells , we developed a model coculture system with Jurkat T cells as the VLP producers . Jurkat cells are well characterized , have high transfection efficiency , and mimic virus-producing T cells . For reasons described below , Raji/CD4 cells were used as targets [31] . Jurkat cells were transfected with viral vectors and incubated for 24 h before adding an equal number of Raji/CD4 target cells and infections were quantitated by luciferase assay 48 h after the start of coculture . Transduction of Luc activity was higher with HTLV-1 vectors compared to HIV-1 vectors ( Fig . 2A ) ; this is notable because cell-free infection with HIV-1 and HTLV-1 VLPs shows the opposite ( see below , section ‘Differences in HTLV-1 and HIV-1 transmission in the coculture infections’ ) . Transduction of Luc activity was 20-fold to 50-fold lower for both viruses when Raji/CD4 target cells were absent; i . e . when VLPs were transmitted between Jurkat cells . In the absence of either an Env expression vector or a viral packaging plasmid , no reporter gene activity was detected ( Fig . 2A ) . VLP infectivity required virus budding , as cotransfection of the viral vectors with a dominant negative VPS4A expression plasmid inhibited Luc transduction . Treatment of cells with azidothymidine ( AZT ) also inhibited VLP infection , indicating the requirement for reverse transcriptase . In addition , mutations of Gag late domains or of viral protease , reverse transcriptase , and integrase genes in the viral packaging plasmids abolished Luc transduction ( data not shown ) . These results indicated that the HIV-1 and HTLV-1 vectors transduced target cells in coculture with no background from the effector cells . Furthermore , cell-cell fusion did not contribute to infection or reporter gene expression ( see Fig . S1 and Protocol S1 ) . To identify suitable target cells for coculture infection experiments , we examined Luc transduction of established cell lines and primary activated CD4+ T cells by coculture with Jurkat cells transfected with HTLV-1 vectors ( Fig . 2B ) . T-cell lines , such as Jurkat , JM , MOLT4 , SupT1 , and CEM174 , yielded similar levels of Luc activity 48 h after the start of coculture with Jurkat effector cells; H9 cells gave somewhat lower levels of Luc activity compared to the other T-cell lines . Unexpectedly , and in repeated attempts with cells from various donors , we were unable to detect Luc activity after coculture of Jurkat effector cells with activated human CD4+ T cells . In parallel experiments , the same CD4+ T cells were efficiently transduced with cell-free HIV-1 VLPs . It is presently unclear whether infection , provirus formation , or subsequent reporter gene expression are inhibited during the 48 h of coculture of activated CD4+ T cells with Jurkat effector cells; we are currently examining the reasons for this inhibitory effect . It was not surprising that HTLV-1 vectors transduced B-cell and monocyte cell lines , as it is well known that HTLV-1 infects a wide variety of cell types in vitro via Env interactions with ubiquitously expressed receptor ( s ) , GLUT-1 , NP1 , and heparan sulfate proteoglycans [32] , [33] , [34] . In general , HTLV-1 transduction of B-cell lines appeared to be higher than the T-cell lines , with Raji/CD4 cells giving the highest level of Luc activity of any cell line tested . It is not clear yet whether the high level of Luc activity detected in Raji/CD4 cells is due to specific interactions with Jurkat cells that enhance virus transmission and replication , or that Raji/CD4 cells support higher levels of reporter gene expression . In addition to the high levels Luc transduction , we chose to use Raji/CD4 cells in the following experiments for several reasons . First , B-cells are natural targets for HTLV-1 infection in vivo and in vitro [35] , [36] , [37] , [38] , [39] . Furthermore , B-cells were shown to form conjugates ( virological synapses ) with HTLV-1-infected T cells in PBMC cultures from HTLV-1-infected individuals [14] . Second , while we recognize that B-cells are not natural targets for HIV-1 infection , Raji/CD4 cells can be infected by both HTLV-1 and HIV-1 vectors for comparative analyses . Finally , Jurkat and Raji cells have been used previously to study immunological synapse formation [40] , [41] , and in experiments described below , this allowed us to determine whether forcing cells together via a superantigen-induced synapse would enhance or inhibit virus infection . To examine the role of cytoskeletal remodeling on single-cycle infection with HTLV-1 and HIV-1 vectors in the Jurkat-Raji/CD4 system , cocultures were treated with cytochalasin D ( ChD ) or jasplakinolide ( Jsp ) , which target actin polymerization/depolymerization , or with nocodazole , which inhibits tubulin polymerization . Because of the reversible effects of the inhibitors and the fact that infection is measured 48 h after the start of coculture , both effector and target cells were exposed to the drugs for the duration of the experiment ( Fig . 3 ) . We observed that there was no significant decrease in cell-free infection of inhibitor-treated Raji/CD4 target cells with HTLV-1 and HIV-1 VLPs ( data not shown ) . To control for the effects of the inhibitors on VLP production in effector cells , we measured Gag protein in cell culture supernatants at the end of the coculture experiment by HTLV-1 p19 or HIV-1 p24 ELISA . Figure 3 shows infectivity ( Luc activity ) , Gag concentration in the supernatant , and normalized infectivity ( Luc activity divided by Gag level ) relative to untreated controls . The actin inhibitors , Jsp and ChD , diminished VLP production by 45% and 14% respectively with HTLV-1 vectors ( Fig . 3A ) and by 49% and 39% with HIV-1 vectors ( Fig . 3B ) . For HIV-1 vectors , the decrease in infection resulting from Jsp and ChD treatment could be accounted for by the decrease in VLP production , because the normalized infectivity was equal to or greater than the untreated control . In contrast , the inhibition of infection with HTLV-1 vectors was significantly greater than the decrease in VLP production , and the normalized infectivity was 3% and 5% of untreated controls ( Fig . 3A ) . Nocodozole had only a modest effect on HTLV-1 VLP production ( 6% decrease ) but had a significant inhibitory effect on coculture infection ( 85% decrease ) ( Fig . 3A ) . For HIV-1 vectors , nocodazole inhibited both VLP production and infection , the normalized infectivity ( 60% compared to untreated control ) was not as severely affected as HTLV-1 . Together , the single-cycle replication data indicate that coculture infections with HTLV-1 vectors were strongly dependant on cytoskeletal remodeling , whereas infections with HIV-1 vectors were much less so in this experimental setting . These results suggest that the cell-free component of HIV-1 infection in Jurkat-Raji/CD4 co-culture predominate , but do not exclude an important role for cell-to-cell infectious transmission of HIV with other effector-target cell combinations . In contrast , HTLV-1 infection appears to require cell-cell contact . We next compared the relative efficiency of cell-free versus cell-to-cell infection with HIV-1 and HTLV-1 vectors . Although VLP titers based on Gag ELISA are very similar for HTLV-1 and HIV-1 vectors , infectious titers are quite different [42] , [43] . In order to obtain sufficiently high infectious titers of HTLV-1 VLPs for cell-free infection experiments , 293T cells were transfected with the same viral vectors and at a ratio identical to those used in transfection of Jurkat cells and luciferase transduction was normalized relative to the amount of Gag in the filtered supernatant ( Fig . 4A ) . Infections in Jurkat-Raji/CD4 cocultures were carried out as before , where infection is normalized to the amount of Gag in the supernatant 48 h after mixing ( Fig . 4A ) . For HTLV-1 vectors , infectivity was at least 4 logs higher in coculture infection compared with cell-free VLPs . In contrast , the infectivity of HIV-1 vectors was only 2-fold higher in cocultures compared with cell-free VLPs . These data are in agreement with the previous experiments and suggest a cell-to-cell mode of infection with HTLV-1 VLPs in Jurkat-Raji/CD4 cocultures , while cell free transmission constitutes a significant component of HIV-1 infection in this system . In the Jurkat-Raji/CD4 coculture system , infection is measured 48 h after cell mixing , as this is the time required for optimal reporter gene expression . It would be desirable , however , to limit the period of VLP transmission to the first few hours after the start of coculture , as one would expect there to be significant differences in the levels of infection at these early times that are related to the mode of virus transmission . This can be accomplished by blocking virus entry at various times after cell mixing with either anti-Env neutralizing antibodies or antibodies that block virus receptors . For the purposes of this experiment , it does not matter which type of antibody is used , as it is important only to block infection after a defined period of virus transmission ( Fig . S2 ) . Time-course of transmission experiments , in which virus entry and infection were blocked with antiserum to HTLV-1 Env or antibody to the CD4 receptor , are shown in Fig . 4B and 4C . Anti-HTLV-1 human serum or anti-CD4 mAb were titrated to give greater than 95% inhibition of infection; control serum or mAb did not inhibit Luc transduction and neutralization of infection was specific for each virus . When antibodies were added at the same time that cells were combined ( 0 h ) , Luc transduction was inhibited by almost 2 orders of magnitude . For HTLV-1 , levels of infection rose rapidly , increasing by about 10-fold during the first 4 h , then began to plateau after 6 h ( Fig . 4B ) . In contrast , HIV-1 infection increased only 2-fold by 4 h after the start of coculture ( Fig . 4C ) . These results indicate that the mechanisms of transmission of HTLV-1 and HIV-1 differ in this experimental system ( see discussion below ) ; for HTLV-1 VLPs , cell-to-cell spread appears to be the dominant mode of infection . HTLV-1 Tax interacts with a variety of cellular proteins to alter transcription , signal transduction , cell adhesion , and cytoskeletal remodeling [44] . To determine whether Tax has an impact on coculture infections , HTLV-1 packaging plasmids were used that contain either a wild type ( Tax+ ) or a mutated tax gene ( Tax− ) ; for HIV-1 infections , cells were cotransfected with HIV-1 vectors plus a Tax expression plasmid ( Tax+ ) or empty vector ( Tax− ) . We first examined the effects of HTLV-1 Tax expression on the time course of VLP transduction where infection was blocked with neutralizing antisera at various times after the start of coculture ( Fig . 5A and B ) . With wt HTLV-1 vectors , Luc transduction increased rapidly over the first several hours of coculture ( Fig . 5A ) . With the Tax− vector , the initial change in HTLV-1 infection was significantly diminished and the time course resembled HIV-1 infection ( Fig . 5A ) . When Tax was co-expressed with HIV-1 vectors , the initial rate of infection was increased significantly , nearly reaching the levels obtained with wild type HTLV-1 vectors ( Fig . 5B ) . These results indicate that HTLV-1 Tax is a major determinant of the difference observed in the mechanism of HTLV-1 and HIV-1 transmission here . This is consistent with known effects of Tax on the expression and activity of adhesion proteins and activation of signal transduction pathways that may cooperate to enhance virus infection and replication . To determine whether Env plays a role in determining the mechanism of virus transmission in coculture infections , we examined pseudotyped VLPs ( Fig . 5C ) . Unfortunately , HTLV-1 VLPs pseudotyped with HIV-1 Env were not infectious and HIV-1 VLPs pseudotyped with HTLV-1 Env had very poor infectivity ( Fig . S3 ) . These results may reflect differences in Gag and Env trafficking between HTLV-1 and HIV-1 , as we have previously shown that HIV-1 and HTLV-1 Gag are targeted to different plasma membrane microdomains for virion assembly in Jurkat cells [40] . However , both HIV-1 and HTLV-1 VLPs can be pseudotyped with VSV-G protein . Expression of Tax in Jurkat cells boosted infection by about 5-fold for both HTLV-1 and HIV-1 VLPs in Jurkat-Raji/CD4 cocultures ( Fig . 5C ) . However , the Tax-induced enhancement of infection was not seen when VLPs were pseudotyped with VSV-G protein , suggesting that Tax elicits specific cellular alterations that enhance infection only in combination with a particular type of Env . We also examined the effects of Tax expression in cocultures of transfected 293T cells or HeLa-P4 cells ( Fig . 5D ) . In this one-step transfection/infection coculture system , transfected cells produce VLPs that infect neighboring cells . Neither HTLV-1 nor HIV-1 VLP transduction was enhanced by Tax in 293T cells or HeLa-P4 cells ( Fig . 5D ) or in cell-free infections ( data not shown ) . Together , these data indicate that Tax-mediated enhancement of infection in cocultures is dependent on cell type and on Env . Perhaps , Tax is not only involved in the mobilization of Gag to the cell synapse , but also in establishing crosstalk between certain Env and adhesion molecules for their efficient membrane movement toward the site of cell-cell contact . It has been reported previously that HTLV-1 Tax induces homotypic aggregation in various T-cell lines [45] . To determine whether the ability of Tax to enhance infection in Jurkat-Raji/CD4 cocultures was simply due to cell-cell aggregation , we examined whether inducing a synapse between Jurkat cells and Raji/CD4 cells would affect cell-to-cell infection . Staphylococcal enterotoxin E ( SEE ) induces the formation of an immunological synapse ( IS ) by binding to both the TCR on Jurkat T cells and MHC-II on Raji B cells [46] . We used this system previously to show that HTLV-1 Gag traffics with tetraspanin-enriched plasma membrane microdomains to the IS [40] . Stable interaction between Jurkat cells and Raji/CD4 cells was assayed by flow cytometry ( Fig . 6A ) . In the absence of Tax expression or SEE treatment , about 20% of Jurkat cells fractionated with Raji/CD4 cells ( Fig . 6A , left hand panels ) . Transduction of Jurkat cells with a lentivirus vector expressing HTLV-1 Tax ( approximately 85% transduction efficiency ) before mixing with Raji/CD4 cells increased conjugate formation to 68% . When Raji/CD4 cells were pretreated with SEE and mixed with Jurkat cells , 85% of Jurkat cells were conjugated with Raji/CD4 cells ( Fig . 6A , lower right hand panel ) . Coculture of transfected Jurkat cells with Raji/CD4 cells that had been pretreated with SEE revealed that induction of an IS did not enhance HIV-1 or HTLV-1 VLP infectivity , either in the presence or absence of Tax ( Fig . 6B ) . These results are in striking contrast to the enhancement of infectivity by Tax , suggesting that Tax increases the expression or activity of specific adhesion molecules and signaling pathways necessary for efficient cell-to-cell infection .
Our aim was to develop retroviral vectors that would make it possible to directly quantify retroviral replication in cell-to-cell virus transmission experiments . Retroviral reporter vectors have greatly enhanced our understanding of the retrovirus infectious cycle , but their utility has been limited primarily to cell-free infection studies due to high levels of reporter gene expression in VLP producer cells . The new reporter vectors , referred to here as inLuc or inYFP vectors , rely on RNA splicing in the effector cell and provirus formation in the target cell to activate reporter gene expression . Considering the complex mechanisms that both HIV-1 and HTLV-1 have evolved to regulate mRNA splicing and transport , the system produces surprisingly clear results . There was no signal from the reporter vector alone in transfected cells and disrupting a viral structural or enzymatic function in the packaging vector abolished reporter protein expression . These vectors are ideally suited for studies of cell-to-cell infection in the coculture setting . In previous studies of virus transmission from a transfected or virus-infected effector cell to a target cell , infection was often inferred by measuring viral protein transfer between cells by microscopy , flow cytometry [11] , [12] , or by the formation of nascent reverse transcription products [6] , [7] . Hubner W . at al . [26] beautifully demonstrated early and late stages of HIV cell-to-cell transmission and target cell infection using 3D video microscopy of replication competent fluorescent HIV clone . However , most of the above methods require virus-infected rather than transfected effector cells due to background problems associated with transfected plasmid DNA . It is therefore difficult to analyze viral mutants and pseudotyped virions . We believe that the vectors described here will help to mitigate some of these problems and enable future quantitative studies of cell-to-cell infection . The new vectors were validated in coculture infections with transfected Jurkat cells as the effectors , because these are well characterized T-cells and have high transfection efficiency . Clearly , it will be important to examine other effector cells , particularly primary human cells , with these vectors in the future . A variety of lymphoid and monocyte cell lines were examined as targets in cocultures with Jurkat cells transfected with HTLV-1 vectors . All cell lines were susceptible to infection , but Raji/CD4 cells gave the highest levels of Luc activity . This was not due to expression of CD4 , as Raji cells gave similar levels of activity ( data not shown ) . It is possible but unlikely that the CMV promoter-driven reporter gene is more active in Raji cells compared to other cell lines , even other EBV-transformed B-cell lines . Alternatively , higher levels of Luc transduction in Raji cells may reflect a unique interaction between adhesion molecules on the surface of Raji cells and Jurkat cells that enhances cell-to-cell infection . We are currently examining these possibilities . Although Raji/CD4 cells are not natural targets for HIV-1 infection , we believe that they do provide an appropriate model system to study HTLV-1 cell-to-cell infection , as B-cells are natural targets for HTLV-1 infection in vivo and in vitro [35] , [36] , [37] , [38] , [39] and can form synapses with HTLV-1-infected T lymphocytes in vitro [14] . While it is desirable to examine cell-to-cell infection using primary T-cells as targets , we have been unable to detect Luc transduction in cocultures of transfected Jurkat T-cells and activated CD4+ T-cells . This appears to be due to a negative effect of the activated CD4+ T-cells on VLP expression in Jurkat cells during the 48 h coculture infection , as the primary CD4+ T-cells ( from various donors ) were susceptible to infection with cell-free HIV-1 VLPs ( D . M . and D . D . , unpublished observation ) . Infections carried out with HTLV-1 vectors in Jurkat-Raji/CD4 coculture had all of the characteristics expected for cell-to-cell infection . Infection was dependent on cytoskeletal remodeling , as inhibitors of actin and tubulin abolished infectivity . The time course of infection revealed rapid and efficient HTLV-1 VLP transmission immediately after mixing effector and target cells , and the difference in cell-free versus coculture infection was consistent with cell-to-cell infection for HTLV-1 . Thus , we believe that this cell culture model system will be useful for examining the cell and virus determinants of cell-to-cell infection for HTLV-1 . Relative to HTLV-1 , HIV-1 VLPs appeared to be transmitted efficiently as cell-free particles . However , the inhibitory effect of the tubulin depolymerization agent nocodazole on HIV-1 coculture infection shows a cell-to-cell component of HIV transmission in our experimental setting . The differences in kinetics of HIV-1 coculture infection versus HTLV-1 infection may also be due to the less efficient or not too rapid VS formation between HIV producer cells and target cells . Thus , early after Raji/CD4 cell addition , cell free infection may predominate , but later transmission through the VS may be favored . The release of HIV from the surface of producer cells needed for viral maturation and recent demonstration of HIV endosomal fusion [47] may influence the rate of infectious transmission . HTLV-1 Gag is found to be processed inside the cells [40] , [48] , [49] , so infectious VLPs can be readily fused with plasma membrane and quickly transfer the infection . HIV Jurkat-Raji/CD4 coculture infection displays a high resistance to neutralizing mAbs ( 25–50 µg/ml for 80–95% inhibition in kinetics experiments versus 5–10 µg/ml for cell-free infection ( DM , DD unpublished observations ) , another indication of cell mediated virus transfer . This is consistent with previous reports of HIV cell-to-cell transmission being resistant to broadly neutralizing Abs and patient serum [10] , [26] , [50] . Therefore , to estimate cell-to-cell infection of HIV per se , the cell-free component of infection should be eliminated experimentally or subtracted from a total level of infectivity measured in cell cocultures . We showed that the HTLV-1 Tax protein is a major contributor to the difference between HTLV-1 and HIV-1 infection in Jurkat-Raji/CD4 cocultures and that Tax significantly enhanced HTLV-1 cell-to-cell infection . The positive effects of Tax on HTLV-1 infection observed in the Jurkat-Raji/CD4 cocultures were not detected in 293T cells , indicating that Tax enhances infection in a cell type-specific manner . Tax is known to modulate many cellular functions , affecting the orientation and activity of the microtubule organizing center ( MTOC ) [51] , up-regulating the expression of cell adhesion molecules such as ICAM-1 [52] , and modulating signal transduction pathways [44] . Furthermore , Tax has been observed to localize to an area in infected T-cells near the interface with the target cell [51] . We suspect that multiple activities of Tax may cooperate to enhance cell-to-cell infection; examination of various Tax mutants , which are defective for specific interactions with cellular proteins , may help to identify critical Tax actions . The likelihood that Tax induces formation of a specialized cell adhesion synapse for efficient viral transmission is suggested by the result showing that a superantigen-induced immunological synapse between Jurkat cells and Raji/CD4 cells did not enhance cell-to-cell infection . Although IS has similarity with VS , the formation of the SEE-induced IS is inappropriate for VLP transmission . Recent reports demonstrating that HIV Gag preferentially forms ring [23] or wide “button” [26] like structures at VS , i . e . localizes in a peripheral , not central , supramolecular adhesion complex of synapse , further highlight the differences between IS and VS . In summary , we have developed an experimental system , which makes it possible to directly quantify cell-to-cell infection . Results obtained with this system underscore the importance of quantitative measurements to validate inferences based on microscopic observations . It will be extremely interesting to extend the experimental approach described here to other cell types and , first of all , to primary human cells , and we are optimizing transfection and coculture conditions for such experiments .
Human cell lines Jurkat E6-1 , JM , MOLT-4 , SupT1 , CEM174 , H9 , AA-2 , Ramos , U937 and THP-1 were obtained through the AIDS Reference Reagent Program , Division of AIDS , NIAID , NIH . Raji/CD4 cells [31] were from Vineet N . KewalRamani ( NCI-Frederick ) and 729B cells were from Patrick Green ( Ohio State University ) . T-cell , B-cell and monocyte cell lines were maintained in RPMI 1640 medium containing 10% fetal calf serum . Primary human CD4+ T cells were prepared from elutriated lymphocytes and grown in RPMI 1640 medium containing 10% fetal calf serum and 100 U per ml of IL-2 after activation with anti-CD28/anti-CD3 beads as described previously [53] . Human kidney 293T cells , Hela-P4 cells [54] ( Eric Freed , NCI-Frederick ) , hybridomas anti-human CD4 clone SIM . 2 and SIM . 4 , anti-HIV-1 gp120 clone 902 ( NIH AIDS Research & Reference Reagent Program ) were grown in Dulbecco's modified Eagle's medium ( DMEM ) containing 10% fetal calf serum . IgG from hybridoma supernatants were purified using protein A ( for IgG1 ) or protein G ( for IgG2b ) HiTrap columns ( GE Healthcare ) , then desaulted using PD-10 column ( GE Healthcare ) , reconstituted in Dulbecco PBS without Ca and Mg and sterilized by filtration through 0 . 45 micron low protein-binding filters ( Millipore ) . The anti-human CD3 clone UCHT1 ( unconjugated , FITC-conjugated , or APC-conjugated ) and anti-human HLA-DR clone TU36 and clone G46-6 ( conjugated with PE ) mAbs were from BD Pharmingen . Goat anti-mouse Alexa 546 secondary antibody was from Molecular Probes . Inactivated plasma from HTLV-1-positive patients was from Scripps Laboratories . Jasplakinolide was from Molecular Probes; cytochlasin D was from Calbiochem; azidothymidine and nocodozole were from Sigma . Staphylococcal enterotoxin E ( SEE ) was purchased from Toxin Technology ( Sarasota , Florida ) . The HTLV-1 packaging plasmid pCMVHT1-M expresses the full-length HTLV-1 genome and pCMVHT1M-ΔEnv expresses all HTLV-1 gene products except Env [42] . The HTLV-1 packaging plasmids , pCMVHT1M-Tax9Q and pCMVHT1M-ΔEnv-Tax9Q are Tax-minus due to a single nucleotide change creating a stop codon in place of the glutamine at Tax codon 9 . The HIV-1 packaging plasmid pCMVΔ8 . 2R expresses all HIV-1 proteins except Env [55] . pCMV-HT1Env and pCMV-VSVG express HTLV-1 Env and VSV-G protein , respectively . The HIV-1 Env expression plasmid pIIINL-4env [56] was obtained from Eric Freed ( NCI-Frederick ) . The HTLV-1 Tax expression plasmid pCMV-Tax1C was described [57] . The dominant negative form of human VPS4A protein was expressed from pGFP-VPS4A-E223Q [58] . Lentivirus vectors for transduction of HTLV-1 Tax and GFP were constructed by subcloning HTLV-1 Tax-IRES-GFP or IRES-GFP cassettes into pUCHR transfer vector to give pUCHR-TaxIRGFP and pUCHR-IRGFP , respectively . New , replication-dependent HTLV-1 reporter vectors were made from pHTC-GFPLuc [42] by first replacing the U3 region in the 5′LTR with CMV promoter , joined at the TATA box . The reporter cassette within the vector was replaced with a cassette from the plasmid pKS99gfp ( from John Moran , University of Michigan ) , containing CMV promoter , gfp gene ( with γ-globin intron ) , and TK polyA signal [29]; this cassette is oriented in the opposite direction relative to viral vector transcription , but the intron is oriented in the sense direction . Finally , the gfp gene was replaced with either luciferase or yfp genes containing a γ-globin intron to give pCRU5HT1-inLuc and pCRU5HT1-inYFP , respectively ( which we refer to as HTLV1-inLuc and HTLV1-inYFP ) . HIV-1 replication-dependent reporter vectors , pUCHR-inLuc and pUCHR-inYFP were constructed from pUCHR-GFPLuc by replacing the reporter cassette with the respective cassettes from HTLV-1 vectors described above , and are referred to here as HIV1-inLuc and HIV1-inYFP . Jurkat cells , 293T cells and Hela-P4 cells were transfected with TransIT®-Jurkat , TransIT®-293 ( Mirus ) , or FuGENE6 ( Roche ) transfection reagents , respectively , according to the manufacturers' instructions . Cell-free infection assays were performed essentially as described previously [59] . VLP concentrations were determined by HTLV-1 p19 or HIV-1 p24 antigen capture ELISA ( Zeptometrix ) . Coculture infections were initiated by transfecting Jurkat cells ( 106 cells in 1 ml ) with 0 . 6 µg of inLuc vector DNA , 0 . 4 µg of packaging plasmid DNA , and 0 . 1 µg of Env-expression plasmid DNA ( for env-minus packaging plasmids ) or empty vector DNA ( for env-positive packaging plasmids ) . After 24 h , cells were washed twice with PBS and 106 Jurkat cells were mixed with an equal number of Raji/CD4 target cells in 5 ml of medium . Cells were collected 48 h later , extracted in GLO lysis buffer ( Promega ) , and Luc activity was measured using Promega luciferase reagent and Lumat LB9501 luminometer ( Berthold ) . For antibody blocking experiments , HTLV-1-positive human plasma was heated at 56°C for 30 min . to inactivate complement and added to HTLV-1 infections at 1/100 dilution; similarly treated HIV-1-positive human plasma was used as a control for HTLV-1 infections . For HIV-1 infections , anti-gp120 902 mAb at a concentration 50 µg/ml or anti-CD4 SIM . 2 mAb 25 µg/ml was added to cells; non-blocking anti-CD4 SIM . 4 mAb at a similar concentration was used as control . Immunological synapse formation was induced by pulsing Raji/CD4 cells ( 107 cells per ml in PBS ) with 10 µg/ml of SEE for 1 hr; cells were washed 4 times with PBS , and combined with Jurkat cells at a 1∶1 ratio . Untreated or SEE-treated Raji/CD4 cells ( 107 per ml ) were mixed with an equal number of Jurkat cells in complete medium for 1 hr . After centrifugation , cells were fixed with 4% paraformaldehyde for 10 min , washed with PBS , stained with anti-CD3-FITC plus anti-HLA-DR-PE Abs for 30 min , and analyzed by flow cytometry . Cell conjugates were expressed as the percentage of double positive cells in the total Jurkat cell population . To determine the effects of HTLV-1 Tax expression on adhesion of Jurkat cells to Raji/CD4 cells , Jurkat cells were transduced with cell-free VLPs that were generated by transfecting 293T cells with pUCHR-TaxIRGFP or pUCHR-IRGFP plus pCMVΔ8 . 2R and pCMV-VSVG . At 48 hrs after transduction , Jurkat cells ( greater than 85% GFP-positive ) were washed and mixed with Raji/CD4 cells at a 1∶1 ratio for 1 hr . Cell conjugate formation was determined as described above .
|
Cell-free virus particles released from infected cells can be transmitted to target cells by diffusion or may be conveyed directly to target cells via specific intercellular contacts; the latter is referred to as cell-to-cell infection . Microscopic imaging has shown how viral proteins and virus particles move within and between cells , accumulating at sites of cell-cell contact . While we suspect that these images represent virus infection , it has been difficult to accurately quantify virus replication and provirus formation in most cell-to-cell infection experiments . Retroviral vectors that encode reporter proteins have been invaluable tools for analyzing retrovirus replication and restriction , but they have had limited utility in cell-to-cell infection studies due to high background noise resulting from reporter expression in the producer cells . We report the construction and characterization of retroviral vectors that express reporter protein exclusively in target cells and only after completing a full replication cycle . We have validated this approach and have begun to analyze cell and virus determinants for cell-to-cell infection with vectors for two human retroviruses that infect T cells . We show that the mechanism of transmission and ensuing virus replication depend on the particular virus , the effector and target cell types , and on the specific type of cell-cell interaction .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"virology/immunodeficiency",
"viruses",
"virology/viral",
"replication",
"and",
"gene",
"regulation",
"virology/host",
"invasion",
"and",
"cell",
"entry"
] |
2010
|
Quantitative Comparison of HTLV-1 and HIV-1 Cell-to-Cell Infection with New Replication Dependent Vectors
|
The functional networks of cultured neurons exhibit complex network properties similar to those found in vivo . Starting from random seeding , cultures undergo significant reorganization during the initial period in vitro , yet despite providing an ideal platform for observing developmental changes in neuronal connectivity , little is known about how a complex functional network evolves from isolated neurons . In the present study , evolution of functional connectivity was estimated from correlations of spontaneous activity . Network properties were quantified using complex measures from graph theory and used to compare cultures at different stages of development during the first 5 weeks in vitro . Networks obtained from young cultures ( 14 days in vitro ) exhibited a random topology , which evolved to a small-world topology during maturation . The topology change was accompanied by an increased presence of highly connected areas ( hubs ) and network efficiency increased with age . The small-world topology balances integration of network areas with segregation of specialized processing units . The emergence of such network structure in cultured neurons , despite a lack of external input , points to complex intrinsic biological mechanisms . Moreover , the functional network of cultures at mature ages is efficient and highly suited to complex processing tasks .
Two aspects of the cultures that are of particular interest are their structural ( anatomical ) circuitry and the interactions which take place over this circuitry , both determining the computational capacity of the underlying network . Whilst cultures are typically too dense for accurate observation of their structural connectivity , analysis of functional connectivity provides a probabilistic estimation of the relationship between distributed neuronal units [2] , thereby enabling spatio-temporal interactions between areas of the network to be measured throughout experiments . This provides a useful means to investigate the network properties of cultures , particularly since functional connectivity estimated over certain timescales may contain information about the underlying structural network [35] . Existing literature indicates that the functional network properties of cortical cultures change during maturation [36] and following stimulation [29] , [37] , [38] . However , such studies have focused on changes in the expected link-level properties such as the mean strength and metric distance of connections [36] , or the proportion of links which are strengthened or weakened following stimulation [37] , [38] . These aggregate measures capture gross changes in global connectivity , but they do not reflect the organizational features of the network , e . g . the distribution of properties amongst the neural units , or whether there are groups of neural units that are more densely connected than others . Analysis of such organizational features would reveal the architecture of the network , enabling investigation into which interactions the network could support and how the network organization changes under different experimental conditions . Importantly , by assessing the complex network properties , the relevance of results from cultures to investigations of whole-brain networks would be increased . Reports that rigorously compare culture's complex network properties under different experimental conditions are very sparse . Mature cultures were assessed in [39] and networks from cultures subject to an in vitro glutamate injury model of epileptiform activity were assessed in [34] . The utility of cultures for investigating changes in cognitive function , characterizing drug effects and modeling disease states , could be greatly extended by applying complex network statistics to quantify the influence of experimental manipulation on the network architecture . Moreover , comparison with results from in vivo networks may reveal basic organizational principles common to both . Experiments utilizing cultures can be undertaken across a range of ages , yet little is known about whether developmental changes occur in culture's complex network properties . Questions such as when and which non-random properties are present , their stability over time and the variability between cultures and their ages remain largely unanswered . The nature of such spontaneously occurring changes in a culture's functional network are important a priori knowledge for assessing experimental outcomes using complex network statistics . Moreover , by analyzing these ‘known’ conditions , a framework can be established for evaluating a variety of experimental conditions , including those resulting from embodying a culture in a closed loop system . [40] , [41] , [42] , [43] . The density at which cultures are seeded exerts an important influence on the rate of maturation . Dense cultures mature faster than their sparse equivalents , and they demonstrate bursting activity earlier in development [44] . For the purpose of the present paper , dense cultures were deemed preferable , since their use enabled network properties to be measured earlier in development than would have been possible on much sparser cultures . Additionally , to investigate changes in the functional network properties during culture maturation , maintaining consistency in plating parameters was important to minimize differences in cultures structural properties . Such differences would have complicated the analysis and interpretation of results . Therefore , cultures at a fixed density were used ( those described in [44] as ‘dense’ ) . At ∼1 , 500 to 6 , 500 cells within the ∼1 . 6 mm2 recording area of the MEA , the cells in such cultures form a monolayer . Moreover , they can be maintained for many months [45] and the density is comparable to that used by other groups ( typically ∼2 , 500–3 , 000 cells per mm2 [24] , [30] , [36] , [42] , [46] , [47] ) . The present study establishes the baseline network statistics for cultures at specified stages of development and uses them to characterize culture maturation . The topological , spatial and performance properties of functional networks captured every 7 days ( 7 to 35 days in vitro [DIV] ) were compared using a population of 10 cultures . The study is one of the first to investigate functional connectivity in an evolving complex system . Here , the evolution of network properties is a counterpart of biological processes shaping the culture's development . Since the graph-theoretic approach and use of complex network statistics is a relatively novel method for investigating functional connectivity in cultures , the key methodological decisions are described next .
The number of nodes and links for a given culture was used to calculate the edge density of its network . Figure 2 shows the expected values for each property . The mean number of nodes was relatively constant and independent of age ( P = 0 . 272 ) . In contrast , the mean number of links measured at DIVs 14 and 21 was lower than at DIVs 28 and 35 , with a strong trend towards a significant increase between the younger and older ages ( P = 0 . 074 ) . Edge density increased significantly between DIVs 14 and 21 ( P = 0 . 012 ) and showed no significant change thereafter . Statistics quoted are for the n = 5–8 cultures valid for complex network analysis ( see Materials and Methods ) . However , results were comparable when all cultures were used . Numbers of nodes and links followed a comparable trend for two different persistence thresholds ( see Figure S1 ) , indicating their robustness to threshold selection . Edge density followed different trends for the different link persistence thresholds; this was due to small differences in the numbers of nodes at each age , resulting in larger differences in edge density ( Figure S1 ) . Results presented thus far have focused on identifying changes in the network infrastructure ( via the persistent interactions between different areas [nodes] in the cultures ) . Here , the results focus upon the activity that takes place over this infrastructure . Each transient network is considered as a ‘snapshot’ of network activity , measured over a short time-scale ( duration of a network-wide burst ) and reflects interactions between different areas of the culture in this period . As per the persistent networks , the basic properties relating to network size were compared . Additionally , since there were multiple transient networks for each culture , the coefficient of variation was also analyzed ( see Materials and Methods ) . Figure 8 panel A shows the expected number of transient links as a function of culture age , panel B shows the equivalent data for number of nodes . There was a strong trend towards an increase in the mean number of transient network links ( P = 0 . 087 ) , and a strong trend towards an increase in the number of nodes ( P = 0 . 089 ) . Figure 8 panel C shows the expected coefficient of variation for the number of transient links . This was largest at DIV 21 and there was a significant increase in coefficient of variation between DIV 14 and DIV 21 ( P = 0 . 021 ) . This demonstrated that transient networks at DIV 21 varied considerably in their numbers of links , more so than at any other age . Panel D shows the equivalent data for number of nodes ( no significant difference ) .
Immature cultures ( DIV 14 ) exhibited limited interactions between neuronal units , resulting in a network of few nodes and links . The observation that at DIV 14 activity could spread rapidly between any two neuronal units ( short mean path length in Figure 3 , reflects high integration ) , but was slow to propagate network-wide ( Figure 9 ) indicated an absence of functional organization . The homogeneous node degree distribution and low clustering coefficient exemplified the poor functional differentiation between nodes , with no evidence of densely interconnected areas that could support segregation of neural processing . Together , these network properties implied a disordered spread of activity , across a random network topology , whilst the long burst propagation time indicated an inefficient structure for widespread information transfer . Since dissociated neurons were seeded randomly onto the MEA and received no external stimulation , it could be expected that their initial connectivity resulted in a random topology . Moreover , since neuron-synapse maturation is incomplete at DIV 14 [24] , [53] , it is unsurprising that the complex network properties found in mature cultures [39] were not present at this age . However , the prevalence of long-distance connections at DIV 14 ( Figure 5 and [36] ) is counter to the economy of wiring principle [54] and suggests that units are not simply making spatially convenient connections . In in vivo and ex vivo preparations the cell type and neurochemical identity have been proposed as guiding influences for connectivity [55] and there is evidence that the variety and proportions of neuron types in cortical cultures are similar to those found in vivo [25] , [27] , [56] , therefore connectivity in cultures could be similarly guided by these influences . Whilst interactions at DIV 14 were clearly unstructured , the subsequent 14 days of development represented a critical window , during which functional complexity increased ( Figure 3 ) , leading to the emergence of the small-world topology at DIVs 28 and 35 . Figures S2 , S3 and S4 demonstrate the robustness of the small-world result . We consider the possible driving forces behind this topology change to include the level of synchronization , the ratio of excitation-inhibition and the mechanism of Hebbian learning . Synchronization of culture activity can be defined over a range of timescales – from ‘synchronous busting’ [57] , where areas of the network are synchronously active ( usually to within ∼100 milliseconds ) , to precise synchronization between the spike times of two or more neural units [36] ( usually to within ∼10 ms or less ) . For the present study , the network links were derived from firing-pattern correlations and thus represent synchronization levels between neural units ( nodes ) ; the low number of nodes and links at DIV 14 reflects a low level of synchronization ( i . e . between only a few units ) , compared to a high level of synchronization ( i . e . between many units ) at DIVs 28 and 35 . Literature indicates that a low level of synchronization at DIV 14 may be due to an excitatory-inhibitory imbalance [53] . Conversely , evidence suggests that a high level of network-synchronization found in older cultures ( whereby many neural units are activated within a short time-window [24] ) is supported by a balanced excitatory-inhibitory subsystem [53] , with tight synchronization between pairs of neural units ( as observed in [36] ) arising from the activity-based refinement of synaptic connection strengths [24] , [58] . In a previous study of functional connectivity during development [36] culture properties at DIV 14 and DIVs 28–35 are in accordance with those of the present study . However , at DIV 21 [36] reported an increased level of synchronization and a dramatic change in burst properties ( compared to those at DIV 14 ) . In contrast , the present study revealed no such increase in synchronization at DIV 21 , yet burst properties were highly variable - as reflected by a highly variable number of transient links ( Figures 2 , 8 ) , and there was a highly variable burst propagation time ( Figure 9 ) . Results herein suggest a network with an uneven balance between highly and poorly interconnected areas , whereby bursts initiated from different sites ( as reported in [59] ) propagate at different rates , with little link activation regularity ( as reflected by the low link persistence at this age ) . We posit that the highly variable burst properties reported herein and in [24] , [36] point to itinerant rather than persistent synchronization at DIV 21 . Such transient synchronization effects may be averaged out by requiring multiple occurrences of correlated activity over long time-scales [58] . Therefore , our persistent networks at this age may not reflect the increased synchronization found in [36] ( where links required only a period of correlated activity during the entire recording ) . Crucially , the combination of varied burst properties and transient synchronization at DIV 21 indicates a mixture of regular and irregular activity . Modeling studies have suggested that such mixed activity constitutes optimal conditions for the emergence of a small-world topology via Hebbian learning rules and activity driven plasticity [60] . Thus , a change in the culture's spontaneous activity patterns could drive the topology transformation . Results herein and in [61] suggest that once the topology of the network has emerged , equilibrium states may exist at different time scales - from transient synchronization between subgroups of neural units at the short time-scale to regular occurrence of such transiently activated subgroups over longer time-scales . Modeling studies may provide further insight into the role of synchronization and the evolution of such equilibrium states [62] , whilst pharmacological manipulation of specific neuron sub-types could verify biological mechanisms behind activity modulation . Networks at DIVs 28 and 35 were classified as small-world , exhibiting several highly connected areas ( clusters of highly inter-connected neural units ) , alongside the ability for any two areas to interact via few intermediary connections ( short mean path length ) . Interestingly , when the network properties at DIVs 28 and 35 were compared , smaller differences were found than between earlier ages , suggesting a state of maturity [24] , [32] , [36] , [63] . The non-trivial network structure demonstrated at DIVs 28 and 35 corresponds well with previous work [39] , which concluded that mature cultures had complex network properties similar to those found in vivo . Small-world networks have an architecture which supports efficient information transfer [8] , [64] . Accordingly , our results showed a developmental reduction in burst propagation time that accompanied the emergence of cultures' small-world properties ( Figure 9 ) . Furthermore , variability of burst propagation time was lower at DIVs 28 and 35 than at younger ages . Since burst events are typically initiated from a number of sites [59] , this reduced variability suggests that burst propagation times in mature cultures are not influenced by burst source; information propagates efficiently from all parts of the network . Interestingly , a small-proportion of links at DIVs 28 and 35 were activated extremely frequently ( Figure 10 ) , suggesting that they facilitate many of the interactions; it is possible that they represent activation of the small-world ‘short cuts’ between clusters . The increasing prevalence of highly connected nodes in older cultures suggests that such hubs play a greater role in network activity as the cultures mature , perhaps indicating sources [65] , sinks , or bridges [18] , [33] for network activity . Interestingly , structural and functional hubs have recently been identified in the developing hippocampus where GABAergic interneuron hubs were found to orchestrate network synchrony [52] , firing immediately prior to network bursts . Similarities between connectivity of GABAergic interneurons in the hippocampus and neocortex [66] and suggestions that cortical cultures develop subsystems akin to those found in vivo [25] , [53] , [56] , imply that similar functional hubs may be present in the primary cortical cultures employed herein . The present study has demonstrated that networks derived from the spontaneous activity of cultures develop non-random properties despite a lack of external input . Based on these results , we draw four main conclusions . Firstly , to mitigate fluctuations in spontaneous activity , multiple network bursts should be assessed to obtain the persistent network . Secondly , the functional network of a cortical culture evolves from an initial random topology to a small-world topology; we propose this is due to a change in the culture's spontaneous activity patterns that is driven by the maturing excitatory-inhibitory balance and an increase in network-wide synchronization . Thirdly , the reduction in burst propagation time with culture maturation that accompanies the evolution of a small-world topology supports the efficient network-wide flow of information afforded by a small-world network . Lastly , the presence of hubs and increasing contribution of links with high persistence suggests a proportion of highly influential nodes and links . To the authors' knowledge , this is the first demonstration of small-world properties evolving in the functional networks of cortical neurons grown in vitro . This further supports work suggesting maturation of in vitro networks around the age of DIV 28 to 35; importantly , our results indicate that experiments which require complex network features should be undertaken from DIV 28 onwards , whilst those aiming to shape network maturation should be undertaken before DIV 28 . Moreover , the work herein further supports the use of complex network statistics to quantify network level changes resulting from different experimental conditions , and importantly it provides a benchmark against which to assess the influence of closed loop stimulation on shaping cultures network properties - a fundamental question for the work on closed loop culture embodiment . An important area for future work is to investigate the role of frequently activated nodes ( hubs ) in cultured neurons; including whether the presence of network-synchrony controlling hubs in the underlying substrate could mediate the timing and extent of functional interactions between otherwise segregated clusters , perhaps coordinating synchronous network-wide bursting . Additionally , the use of staining to identify the location and proportion of the different neuron types and sub-types , and the use of pharmacological manipulation to verify their effect on activity may help elucidate mechanisms behind the different network properties .
Data used for the present study was collected for [44] , from cultures of pre-natal ( E18 ) rat dissociated cortical neurons and glia cells , seeded onto multi-electrode arrays ( MEAs , Multi Channel Systems , Reutlingen , Germany ) . Cultures were maintained in Teflon sealed MEAs in an incubator at 5% CO2 , 9% O2 , 35°C and 65% relative humidity [44] . For the present study , ‘dense’ cultures ( estimated cell density of 2 , 500±1 , 500 per mm2 ) were used . Culture's electrical activity was recorded daily during their first 5 weeks of development . For the present study , a sample population was selected from the large number of cultures recorded , specifically , 10 cultures from 4 preparations ( plating batches ) . Cultures were arbitrarily selected from those that had recordings every 7 DIV , i . e . those which survived for the full 5 weeks and for whom none of the weekly recordings were missed . The use of multiple preparations is important as bursting patterns across preparations vary considerably [44] . Additionally , since the variation in burst properties measured at the same age ( DIV ) from different cultures ( of the same plating ) , can exceed day-to-day differences in their properties ( and inter-plating differences are significantly larger ) [44] , network properties were compared at weekly intervals . This also allows easy comparison with results from other studies [32] , [36] . Data were recorded from cultures for 30 minutes daily in the incubator used for culture's maintenance . Unit and multi-unit spontaneous spike firing was recorded from the MEA ( 8×8 array of 59 planar electrodes , each 30 µm diameter with 200 µm inter-electrode spacing [centre to centre] ) . The pre-amplifier was from Multi Channel Systems ( MCS ) , excess heat was removed using a custom Peltier-cooled platform . Data acquisition and online spike detection was performed using MEABench [67] . According to the MEA user manual ( MCS ) spike detection is reliable up to ∼100 µm from the electrode centre , beyond which spikes become indistinguishable from the background noise . Therefore , each MEA provides a grid of 59 non-overlapping 100 µm recording horizons ( once the four analogue channels and single ground electrode are removed ) . It should be noted that data recorded on each channel may be from multi-neuron activity , no attempt was made to spike sort the data as overlapping waveforms found during a burst can present problems [44] . Lastly , as recording began immediately after the cultures were transferred to the pre-amplifier , the first 10 minutes were discarded from the analysis in order to mitigate any movement induced changes in culture activity [44] , [68] . Spikes were detected online ( using MEABench ) , positive or negative excursions beyond a threshold of 4 . 5× estimated RMS noise , were classed as spikes . Their peak amplitude timestamp ( µs ) , plus electrode number were stored . For the present study , all positive amplitude spikes were removed to avoid counting spikes on both upwards and downwards phases . In cortical cultures , global bursts ( population bursts ) , characterized by an increase in culture activity across the entire MEA , are typically present from DIV 4–6 onwards [44] , but sometimes as late as DIV 14 onwards [36] . Such bursts provide a time window during which many culture interactions take place and thus a useful opportunity to assess network-wide connectivity . For the present study , global bursts were identified as an increase in the number of spikes detected per unit time , summed over all electrodes in the array: specifically ≥4 spikes per channel in 100 ms , on ≥4 channels within 250 ms; based on the SIMMUX algorithm , included as Matlab ( The MathWorks , Natick , MA , USA ) code with MEABench . Burst start was determined by the timestamp of the first spike included in the global burst , and burst end taken as the timestamp of the last spike included . To assess interactions between neural units underlying all the electrodes , global bursts in which at least 25% ( 15/59 ) electrodes registered channel bursts ( ≥4 spikes in 100 ms ) were selected . These were termed ‘network-wide’ bursts and ensured that many neural units participated in the burst ( increasing the probability that the resultant networks would have sufficient numbers of nodes for the analysis of network properties ) . Additionally , since there were typically 10 to 150 such bursts in the 20 minute recording segment used , it provided a good balance between having sufficient numbers of bursts for analysis , whilst avoiding the inclusion of ‘tiny’ bursts [44] since these may have biased results . All activity occurring from the first spike in the nw-burst to the last spike in the nw-burst ( including tonic activity from electrodes not included in the nw-burst ) was used for assessing the relationships between channel pairs . Spike occurrences were counted in 1 ms bins , this allowed a certain amount of jitter in the spike arrival times ( which could otherwise decrease the likelihood of identifying correlated activity ) . Bin size was selected based on experimentation with 1 , 5 and 10 ms bins . The 1 ms bins provided a greater separation between correlated and un-correlated channels , data not shown . Functional connectivity was assessed by correlating spike times recorded on pairs of electrodes during a network-wide burst ( as per [34] ) . This linear link analysis method assesses the probability of a spike at time t on one electrode being accompanied by a spike arriving at t±k on another electrode , where k is the allowable lag time . Spike times arriving within ±13 ms of each other were considered to be related ( under the assumption that a linear relationship between spike arrival times on pairs of electrodes indicates their coupling ) . The maximum lag time was based on speed of axonal propagation , time for synaptic transmission and the maximum distance between 2 points on the MEA . Since the firing rates recorded on each channel may be different , cross-covariance was used , this correlates deviations in firing rates from their respective means as a function of lag . Channels that had fewer than 8 spikes recorded during the burst were excluded from the cross-covariance analysis , as results from synthetic data testing showed that performing cross-covariance on vectors with fewer than 8 spikes was poor at distinguishing related vectors from independent ones ( data not shown ) . The cross-covariance function calculates the covariance of two random vectors: ( 1 ) In the case where X and Y are time-series the cross-covariance may depend on the time when it is estimated and on the lag between the time series: ( 2 ) For wide-sense stationary time series , covariance is a function of the lag only: ( 3 ) Cross covariance was calculated using the built in Matlab function xcov; specifically , each pair of channels with at least 1 ms overlap in their activity were compared from the time of the first spike on either channel to the time of the last spike on either channel . The tightness of the correlation window ( X or Y channel recording spikes ) , and requirement for overlapping activity was to mitigate the effects of long periods of quiescence and to ensure that the data were as wide-sense stationary as possible . Calculation of the cross-covariance at each lag resulted in a cross-covariance plot for each channel pair . The maximum cross-covariance value ( peak of the plot ) was used to determine whether a link between nodes was present by comparing it to a threshold as detailed next . Table S1 provides the mathematical definitions for the topological properties and complex network statistics . Basic topological properties ( related to network size ) , and complex network statistics , were calculated from the adjacency matrices ( using Matlab , with additional scripts from the Brain Connectivity Toolbox [14] ) . For each transient network , only basic topological properties were measured , complex network statistics were not calculated due to the highly variable network size and edge density ( see verification of network size and edge density ) . Instead , the mean numbers of nodes and links were calculated over all transient networks in the recording . Additionally , the coefficient of variation for number nodes and for number of links was calculated over all transient networks in the recording . The expected numbers of nodes , and links and the expected coefficients of variation were calculated over all 10 cultures . In addition to the networks' topological properties , the spatial and temporal features of the networks were also assessed; link distance was calculated as the Euclidean distance between the electrodes on the MEA , based on 200 µm centre-to-centre spacing of the electrodes . For the present study , connections between nodes up to 566 µm ( 2 electrodes ) apart were considered as ‘nearby’ and those greater than 566 µm as ‘distant’ . Link persistence was calculated using the weighted persistent network adjacency matrix ( i . e . prior to thresholding ) , normalized so that the persistence value was the percentage of transient networks in which the link was found . For both link length ( derived from the distance between connected nodes ) and link persistence , histograms were obtained over all links from all cultures at each age . Thus , for link length , a count of the number of links in each bin ( bin size = 1 electrode spacing ) was calculated for each network , this was normalized to the total number of links in the network . For link persistence , a count of the links at each persistence level ( bin size 5% ) was calculated for each network . In both cases , median bin values were obtained over all 10 cultures , therefore the histogram proportions may not always sum to 1 . To quantify the changes in link length and persistence , two further measures were assessed: for link length , the proportion of links between spatially nearby vs distant nodes was calculated for each culture , and the median of these values was used to compare results between ages; for link persistence , the contribution of persistent links was measured as the number of links in each 5% persistence category multiplied by the category persistence value ( e . g . if 20% of links were found in the 10% persistence category , the contribution was 200 ) . The link contribution counts were further binned into transient ( <25% ) and persistent ( ≥25% ) . The efficiency of information broadcast was measured as burst propagation time ( time to recruit all channels in a network-wide burst ) . This was calculated in milliseconds from the time of the first spike in the burst , until the time at which all channels participating in the burst had been recruited . Channels could be recruited to the burst whilst the burst was in progress ( i . e . sufficient channels displayed the required activity ) but once the number of channels bursting dropped below the threshold , channels could no longer be recruited . For each channel included in the burst , recruitment time was the timestamp of the first spike in the burst activity sequence . Burst propagation times were calculated for all bursts of a culture at each age and the median of these was calculated for each age . Outliers ( values <5th and >95th percentile ) were removed from the data . All statistics were obtained using SPSS version 17 . 0 ( SPSS Inc . , Chicago , USA ) . Unless otherwise specified P<0 . 05 was set as the significance level . Statistical tests for each network property were selected based on the experiment design and form of the resultant data; Checks were performed to ensure that the assumptions of each test were met . Following test selection , statistical power was verified at the 80% level ( checking that the proposed test statistic had sufficient power to detect a genuine effect [71] [typically set to a difference of 1–2 times standard deviation of the mean] , given the n numbers and variability of the data ) . For the present study , where some of the tests were applied to data with relatively low n numbers it was important to ensure that the power of each test was sufficient [72] . It was also important to ensure that the assumptions of the statistical tests were not violated ( Text S1 describes the selection and validation of statistical tests used in the present study ) . The selected tests were as follows: To check for a significant increasing or decreasing linear trend of the network properties as a function of the culture age , results for each network property were compared using a one-way ANOVA . Culture age ( DIV ) was the factor , and the network property was the dependent variable . The following properties were assessed in this manner: number of nodes , number of links , edge density , normalized mean path length , normalized clustering coefficient , small-worldness , goodness of fit ratio . In cases where a significant trend was found , Bonferroni and Tukey post-hoc tests were performed to check for significant differences between each pair of conditions , where found , the homogeneous subsets are mentioned in the results . Homogeneity of variances was tested using the Levene test . Normality was tested using the Shapiro-Wilk normality test . In cases where the sample means were not normally distributed , non-parametric tests were used . For the burst propagation times a Kruskal-Wallis test was performed on the median burst propagation times for each culture at each age , with culture age as the grouping factor and median burst propagation time as the dependent variable . For the proportion of links to nearby vs distant nodes at each culture age , a 2-tailed Wilcoxon signed rank sum test was used . To compare the contribution of persistent links at each age , Friedman's rank test was used . Lastly , for the skewness of the link length distributions , a z-test was calculated based on the skewness estimate taken over the standard error of the skewness estimate . The P value was then calculated using the online statistics analysis tool ( http://www . quantitativeskills . com/sisa/calculations/signhlp . htm , accessed November , 2010 ) .
|
Many social , technological and biological networks exhibit properties that are neither completely random , nor fully regular . They are known as complex networks and statistics exist to characterize their structure . Until recently , such networks have primarily been analyzed as fixed structures , which enable interaction between their components ( nodes ) . The present work is one of the first empirical studies investigating the adaptation of complex networks [1] . Network evolution is particularly important for applying complex network analysis to biological systems , where the evolution of the network reflects the biological processes that drive it . Here , we characterize the functional networks obtained from neurons grown in vitro . Network properties are described at seven day intervals during the neurons' maturation period . Initially , neurons formed random networks , which spontaneously reorganized to a ‘small-world’ architecture . The ‘small-world’ concept derives from the study of social networks , where it is referred to as ‘six-degrees of separation’: the connection of any two individuals by as few as six acquaintances . In brain networks , this translates to rapid interaction between neurons , mediated by a few links between locally connected clusters ( cliques ) of neurons . This architecture is considered optimal for efficient information processing and its spontaneous emergence in cultured neurons is remarkable .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biotechnology",
"neurobiology",
"of",
"disease",
"and",
"regeneration",
"signaling",
"networks",
"geometry",
"neuroscience",
"biomedical",
"engineering",
"biological",
"systems",
"engineering",
"mathematics",
"bioengineering",
"developmental",
"neuroscience",
"biology",
"neural",
"circuit",
"formation",
"systems",
"biology",
"euclidean",
"geometry",
"topology",
"computational",
"biology"
] |
2012
|
Emergence of a Small-World Functional Network in Cultured Neurons
|
Regulation of translation initiation is well appropriate to adapt cell growth in response to stress and environmental changes . Many bacterial mRNAs adopt structures in their 5′ untranslated regions that modulate the accessibility of the 30S ribosomal subunit . Structured mRNAs interact with the 30S in a two-step process where the docking of a folded mRNA precedes an accommodation step . Here , we used a combination of experimental approaches in vitro ( kinetic of mRNA unfolding and binding experiments to analyze mRNA–protein or mRNA–ribosome complexes , toeprinting assays to follow the formation of ribosomal initiation complexes ) and in vivo ( genetic ) to monitor the action of ribosomal protein S1 on the initiation of structured and regulated mRNAs . We demonstrate that r-protein S1 endows the 30S with an RNA chaperone activity that is essential for the docking and the unfolding of structured mRNAs , and for the correct positioning of the initiation codon inside the decoding channel . The first three OB-fold domains of S1 retain all its activities ( mRNA and 30S binding , RNA melting activity ) on the 30S subunit . S1 is not required for all mRNAs and acts differently on mRNAs according to the signals present at their 5′ ends . This work shows that S1 confers to the ribosome dynamic properties to initiate translation of a large set of mRNAs with diverse structural features .
Translation initiation , which ensures the formation of the first codon–anticodon interaction into the peptidyl ( P ) -site of the small ribosomal subunit in the correct frame , is the rate-limiting step of protein synthesis . Binding of the mRNA to the 30S subunit takes place at any time during the assembly of the 30S initiation complex ( 30SIC ) and the kinetics is independent of the initiation factors , relying uniquely on 30S and mRNA features ( e . g . , [1] , [2] ) . Crystal structures of the ribosome containing an unstructured mRNA and tRNA showed the mRNA path , refered as the mRNA channel , occupied by around 30 unpaired nucleotides forming numerous interactions with the 30S subunit [3]–[5] . The formation of a short duplex between the Shine–Dalgarno ( SD ) sequence of the mRNA and the 3′ end of the 16S rRNA ( aSD ) locks precisely the 5′ end of mRNA at the exit site of the mRNA channel , a specific place of the 30S known as the platform [3] , [6] . The SD/aSD interaction is sufficient for unstructured model mRNAs to bind to the 30S , however most of the natural mRNAs contain additional sequences or structure motifs in their 5′ untranslated regions ( UTRs ) that have been exploited by bacteria to regulate translation initiation [7]–[11] . Structures in the 5′UTR of mRNAs are thought to represent a kinetic barrier that could lower translation initiation rates because the 30S must disrupt first the structures it encounters in the ribosome binding site ( RBS ) to allow the mRNA to reach its decoding site [2] . Several studies have revealed that mRNA structure motifs located upstream of the initiation codon bind to the 30S in a two-step process [2] , [12]–[14] . A typical example is Escherichia coli rpsO mRNA encoding ribosomal protein ( r-protein ) S15 , which carries a pseudoknot structure within the RBS , and which is recognized by the 30S for translation and by S15 for autoregulation [15] . Structure analysis of several ribosomal complexes identified intermediates of the initiation pathway of rpsO mRNA [13] . It revealed that the pseudoknot structure is first docked on the 30S platform where it forms the SD/aSD helix and interacts with r-proteins S2 , S7 , S11 , and S18 . In a second step , the pseudoknot unfolds to promote the formation of the codon–anticodon interaction at the P-site . This activity is carried out by the ribosome , but the mechanism is yet unknown . Recent studies have shown that the 30S is endowed with an RNA helicase activity at the mRNA entry site . This helicase activity is due to the r-proteins S3 , S4 , and S5 , which unwind mRNA structures during translation elongation [16] , [17] . Are both the extremities of the mRNA channel endowed with a similar RNA unfolding activity ? In other words , is the platform of the 30S able to unfold mRNA structures to promote mRNA accommodation during translation initiation ? The protein environment of the 30S platform consists of several essential r-proteins , namely S1 , S2 , S7 , S11 , and S18 [5] , [13] , [18] . Among these proteins , S1 is an atypical r-protein because it is the largest and most acidic one that is weakly and not always associated with the 30S subunit [19] . The protein consists of six imperfect OB-fold repeats , which is an RNA-binding module specific for single-stranded regions , and is found in many proteins involved in RNA metabolism [20] . Although the structure of the protein has not yet been solved , a cryo-EM analysis suggested that the protein may adopt an elongated shape on the 30S and may bind 11 nts upstream of the SD of a model RNA [18] , [21] . E . coli r-protein S1 is essential for the translation of many mRNAs and for viability [22] . Particularly , S1 forms an essential component of the mRNA binding site for mRNAs lacking or bearing weak SD sequences [23]–[26] . Furthermore , isolated S1 is able to melt RNA duplexes or helices independently from the 30S [27]–[31] . These works led to the hypothesis that S1 would confer to the 30S an RNA melting activity to facilitate translation of structured mRNAs , although these studies were not carried out on S1 bound to the ribosome and with natural mRNAs . Finally , S1 has been implicated in many other functions [20] . The versatility of the RNA–S1 interaction and the existence of multiple OB-fold domains might explain the diverse biological functions of S1 outside or on the ribosome . In the present work , we demonstrate that r-protein S1 confers to the 30S an RNA chaperone activity , which is modulated by the ribosomal environment and essential for the binding and the accommodation of structured mRNAs into the decoding channel . We have analyzed the S1 dependence on three different mRNAs from E . coli , which all contain specific binding sites for translational repressors located close to or within the RBSs and which are repressed at the translation initiation step by various mechanisms . Using these natural mRNAs , we show that S1 on the ribosome interacts transiently with structured mRNAs and promotes a metastable folding state to create new interactions with the 30S subunit . The melting process is slow and represents most likely the rate-limiting step of translation of structured mRNAs . In contrast , an mRNA bearing optimal SD sequence and weakly structured RBS does not need S1 to form active ribosomal initiation complex . Our study reveals the mechanism of action of r-protein S1 on natural mRNAs and how S1 modulates the activity of the 30S dependent on the mRNA context .
We first monitored the effect of r-protein S1 on the formation of the 30SIC using three different natural mRNAs ( Figure 1A ) . These mRNAs have been selected because they have evolved specific structural features to be well translated and regulated at the initiation step of translation . They also all carry an unpaired SD sequence . E . coli sodB mRNA ( SD AAGGAG , ΔG −8 . 48 kcal/mol predicted for the SD/aSD helix ) , encoding superoxide dismutase , contains a weakly structured RBS [32] and the binding sites for the translational repressor RyhB sRNA and Hfq [33] . E . coli thrS mRNA ( SD UAAGGA , ΔG −5 . 96 kcal/mol ) , encoding threonyl-tRNA synthetase ( ThrRS ) , contains a bi-partite unstructured RBS interrupted by a hairpin structure recognized by ThrRS for translation repression [34] . Both RyhB and ThrRS hinder the ribosome binding to repress translation . Finally , E . coli rpsO mRNA ( SD GGAG , ΔG −5 . 85 kcal/mol ) contains a pseudoknot structure , which sequesters part of the coding sequence . Binding of r-protein S15 stabilizes the pseudoknot on the 30S platform to prevent the start codon from reaching the P-site [13] . Toeprinting assays were used to analyze the formation of a simplified 30SIC , composed of the 30S , the mRNA , and the initiator tRNA [35] . A toeprint is observed at position +16 ( +1 is the adenine of the start codon ) if the mRNA occupies the decoding channel and if the codon–anticodon interaction takes place at the P-site . To monitor the action of S1 , the assays were performed with wild-type ( WT ) 30S , S1-depleted 30S ( 30S−S1 ) ( Figure S1A ) , or the 30S−S1 complemented with purified r-protein S1 ( 30S+S1 ) . Quantification of the data showed that the 30S efficiently recognizes and accommodates sodB mRNA into the decoding channel in the presence or in the absence of S1 ( Figure 1B ) . Thus , S1 is dispensable for mRNA carrying an unstructured RBS with a strong SD sequence . Conversely , the formation of the 30SIC performed with the 30S−S1 , thrS mRNA , or rpsO mRNA are strongly perturbed , showing that S1 has a role for activating these mRNAs ( Figure 1B ) . Because thrS and rpsO mRNAs have a weak SD , we addressed the question of whether S1 would be required for the docking and/or for the accommodation process of these mRNAs by introducing an enhanced SD ( AGGAGGU , ΔG −12 . 53 kcal/mol ) to reduce the S1 dependence for mRNA docking . Translation of thrSSD and rpsOSD mRNAs was indeed significantly enhanced in vivo [36] , [37] . Formation of the 30SIC with thrSSD mRNA was similar with WT 30S and 30S−S1 , indicating that S1 becomes dispensable ( Figure 1C ) . However , for rpsOSD mRNA , the yield of 30SIC was still low when formed with 30S−S1 . Concomitantly , several other reverse transcriptase ( RT ) pauses in rpsOSD mRNA were observed when WT 30S and 30S−S1 were bound to the mRNA . These stops located at positions −5 and +10 correspond to the entrance of the SD/aSD helix and to the pseudoknot structure , respectively ( Figure 1C ) . They represent signatures of the stalled 30S pre-Initiation complex ( 30S-preIC ) in which rpsOSD mRNA binds to the 30S but remains folded onto the 30S platform [13] . We then performed filter binding assays to monitor the direct binding of 5′ end-labeled mRNAs to WT 30S or 30S−S1 in the absence of the initiator tRNA—that is , before the accommodation step . The binding saturation curves show that S1 strongly enhances the docking of WT thrS and rpsO mRNAs on the 30S , while sodB , thrSSD , and rpsOSD mRNAs bind to the 30S independently of S1 ( Figure S1B ) . The three WT mRNAs bind the WT 30S ( containing S1 ) with a similar Kd value ( around 1 µM ) , although the SD sequence of sodB mRNA is stronger than the SD sequence of thrS and rpsO mRNAs . However , the absence of S1 on the 30S strongly decreases the recognition of WT thrS and rpsO mRNAs ( Figure S1B ) . This S1-specific effect was completely alleviated when the SD was enhanced in thrSSD and rpsOSD mRNAs , and the binding affinity for the 30S increased 5-fold ( Figure S1B ) . Therefore , a strong SD sequence compensates the lack of r-protein S1 to anchor the mRNAs onto the 30S subunit . However , the ability of rpsOSD mRNA to bind the 30S independently of S1 is not sufficient for its translation because S1 is still required to promote the formation of the active 30SIC as evidenced by the toeprinting assays ( Figure 1C ) . Hence , these data indicate that S1 is directly involved in the accommodation of rpsO mRNA into the decoding channel . Together , the data show that two S1 functions can be distinguished: ( i ) promotion of mRNA binding and ( ii ) mRNA accommodation . The various activities of r-protein S1 reflect the diversity of RBS architectures . The data support the following schemes where mRNAs with weakly structured RBSs ( i . e . , sodB and thrS ) form 30SICs in a single step—that is , binding directly leads to active 30SIC formation . Instead , with structured mRNA ( i . e . , rpsO ) two distinct steps have been identified , where mRNA binding ( influenced by their SDs ) precedes its accommodation into the decoding channel . In both cases , the need of S1 for the binding is exclusively dictated by the strength of the SD sequence , whereas S1 is essential for the accommodation of structured mRNAs . Because our data suggest that r-protein S1 promotes the accommodation of rpsO mRNA on the 30S that would require unfolding of its pseudoknot structure , we used fast kinetics to analyze the structural changes of the pseudoknot on the ribosome using 2-aminopurine ( 2-AP ) modifications . The fluorescent nucleobase 2-AP , which can interact with uracil in a Watson–Crick pair or with cytosine in a wobble pair , is known to quench its fluorescence emission in a quantifiable manner [38] , depending on local changes of the RNA structure when it stacks with other bases while fluorescence increases when it is fully exposed to solvent [39] , [40] . Two modifications were introduced in a mRNA fragment encompassing the rpsO pseudoknot called psk-rpsOSD ( containing nucleotides −56 to +12 ) at the strategic positions A-40 and A-42 involved in long-range interactions of the pseudoknot structure ( Figure 2A ) . Melting of the pseudoknot on the 30S is expected to enhance fluorescence due to an increased accessibility of A-40 and A-42 towards the solvent . The kinetic of the pseudoknot melting on the 30S was analyzed by stopped-flow fluorescence experiments . The formation and stabilization of the pseudoknot structure was evidenced during the renaturation process . The addition of Mg2+ , known to greatly stabilize the pseudoknot structure [15] , causes significant quenching of the fluorescence ( Figure S2A ) . To analyze the effect of S1 isolated or 30S-bound , we added either S1 alone , 30S containing S1 , or 30S−S1 . The time course of the increase in 2-AP fluorescence as the result of the pseudoknot melting was reproducibly observed when the RNA was incubated with the WT 30S ( Figure 2B ) . Conversely , the addition of 30S−S1 to the 2-AP modified RNA only slightly changed the fluorescence emission as compared to the controls ( Figure 2B ) . Noteworthy , binding and toeprinting experiments showed that rpsOSD mRNA is well recognized by the 30S−S1 but does not form an active 30SIC ( Figures 1C and S1B ) , demonstrating that mRNA binding to the 30S is not sufficient per se to change the fluorescence . Therefore , our data indicate that the increased fluorescence is mediated through S1 due to an increased accessibility of A-40 and A-42 towards the solvent . Our data are consistent with previous findings showing that G-39 and G-41 of rpsO were highly accessible to RNase T1 in 30SIC , while these residues were not cleaved in the stalled 30S complex where the pseudoknot structure is stabilized [41] . The analysis of the stop-flow data required fitting a double exponential function , revealing two kinetic phases for the melting of the pseudoknot structure , a fast ( kfast 0 . 9 s−1 ) and a slow ( kslow 0 . 08 s−1 ) process . The kfast value corresponded to the majority of the fluorescence increase ( 73 . 3% ) . The addition of the initiator tRNA had no effect on the kinetics , suggesting that the S1-dependent melting process does not require the formation of the anticodon–codon interaction ( unpublished data ) . The same experiment performed with r-protein S1 alone shows marginally enhanced fluorescence emission . Saturation could not be attained even with long recording times , so that the fitting of the S1 curves could not be performed accurately ( Figure S2B–C ) . As a control , we demonstrate that S1 deleted of the OB-fold domains 1 , 5 , and 6 , a mutant with impaired mRNA and 30S binding ( see below ) , did not enhance the fluorescence emission ( Figure S2B–C ) . These experiments show that the RNA melting activity of S1 is strongly stimulated when the protein is bound to the 30S as compared to the isolated protein , indicating that S1 is primarily acting on the 30S subunit . Thus r-protein S1 is endowed with a 30S-stimulated melting activity that leads to the unfolding of the pseudoknot structure required for the relocation of the mRNA into the decoding channel . Because ribosomal protein S15 stabilizes the pseudoknot conformation of rpsO onto the 30S to repress its own translation , we analyzed whether r-protein S1 interferes with the regulatory function of S15 ( Figure S2D ) . Toeprinting reveals that in the absence of initiator tRNA , formation of the trapped ribosomal complex involving S15 , WT 30S , and rpsO mRNA causes several RT pauses around position +10 , corresponding to the entrance of the pseudoknot . Identical patterns were observed with 30S−S1 or 30S+S1 , indicating that the pseudoknot is stabilized by S15 regardless the presence or not of S1 on the 30S ( Figure S2D ) . Therefore , S1 did not affect the formation of the trapped complex , while in the absence of S15 , the formation of the active 30SIC was strictly dependent on S1 ( Figure 1B ) . These data illustrate that the mRNA unfolding activity of S1 can be counterbalanced by regulatory factors such as S15 , which stabilizes the mRNA in the structured form onto the 30S platform . To gain more insight into the mechanism of interactions between S1 and the pseudoknot of rpsO mRNA , we analyzed deletion mutants of S1 ( Table S1 , Figure S3A ) lacking one or more OB-fold domains based on sequence and structural information available for domains 4 and 6 [42] . To avoid possible structural heterogeneity of the rpsO mRNA fragment forming the pseudoknot , we also studied the mutant ( C-14 to G , mut psk-rpsOSD; Figure 2A ) , which was shown to exclusively form the pseudoknot structure [43] . We first show that WT S1 binds similarly to the two RNAs ( rpsOSD and mut-rpsOSD ) and that the protein concentration around 400–500 nM causes a shift of almost 50% of the 5′ end-labeled RNAs ( Figure 3A–B ) . These data were also well correlated with surface plasmon resonance ( SPR ) experiments ( Kd≈350 nM; Figure S3B ) . The contribution of each OB-fold domain in recognizing wt or mut psk-rpsOSD mRNAs was defined using gel retardation assays ( results not shown and Figure 3C , respectively ) . The deletion of domain 1 or of the two first N-terminal domains ( Δ12 ) in S1 caused a complete loss of RNA binding even at a concentration of 5 µM . The removal of domains 4 to 6 ( S1Δ4–6 ) decreased the stability by 5-fold , while the additional deletion of domain 3 ( S1Δ3–6 ) abolished mRNA binding . Deletion of domains 5 and 6 affected RNA binding only slightly . These data correlate well with the SPR experiments , which show that the truncated protein S1Δ12 interacts weakly with psk-rpsOSD ( Figure S3B ) . Taken together , these data demonstrate that the six OB-fold domains of S1 are not functionally equivalent , with the first three N-terminal domains of r-protein S1 being essential for the recognition of rpsO pseudoknot structure . Because the domains of S1 are not equivalent for RNA binding , we then analyzed the importance of each OB-fold domain for cell growth in vivo . We constructed a set of strains with the chromosomal copy of rpsA ( the gene for S1 ) carrying deletions of increasing length as well as a control allele with the kan cassette inserted downstream of WT rpsA , called rpsA1 ( Figure S4A ) . The growth of the control strain and the levels of S1 were identical to that of the WT strain ( Figure 4A–B ) . Two other mutant alleles carry either deletion of domain 6 ( rpsAΔ6 ) or of domains 5 and 6 ( rpsAΔ56 ) . The alleles rpsA1 , rpsAΔ6 , and rpsAΔ56 were obtained with high yields as haploids , indicating that they are viable ( Figure S4B ) , although the growth of the two mutant strains was slower than the WT strains ( Figure 4B ) . In addition , rpsAΔ6 and rpsAΔ56 alleles confer a cold-sensitive phenotype ( Figure S4C ) . Larger replacements such as rpsAΔ4–6 ( deletion of domains 4 to 6 ) , rpsAΔ3–6 ( deletion of domains 3 to 6 ) , and rpsAΔ2–6 ( deletion of domains 2 to 6 ) were only obtained as diploids carrying both the WT and the mutant copy of rpsA ( Figure S4D ) . We then transduced these three mutant alleles to strains transformed with the complementing plasmid pNK34 , which carries the rpsA gene under the control of an IPTG-inducible promoter . In the absence of IPTG , the strain carrying rpsAΔ4–6 was able to grow , whereas the strains carrying the larger deletions ( rpsAΔ3–6 and rpsAΔ2–6 ) did not grow , indicating that they are lethal alleles ( Figure 4C ) . In summary , the in vivo experiments showed that the successive deletions of the OB-fold domains gradually affect cell growth: the two last C-terminal domains are dispensable , the additional deletion of domain 4 still allows growth but at extremely slow rates , and the further deletion of domain 3 causes lethality . Because some of the domains of r-protein S1 were dispensable in vivo , we analyzed the implication of each OB-fold domain in binding to the 30S . The WT and mutant proteins were incubated with the 30S at a ratio of 3∶1 , and the excess was removed by size exclusion chromatography . The S1-occupancy of the 30S was quantified by Western blot and revealed that a minimal protein containing domains 1 to 3 fully retains 30S binding ( Figure 5A ) . Only the deletion of the two first N-terminal domains ( 1 and 2 ) totally abolishes 30S binding . Formation of the active 30SIC using thrS , rpsO , or rpsOSD mRNAs , the initiator tRNA , and the 30S−S1 pre-incubated with the different S1 variants was monitored by toeprinting ( Figures 5B–D and S5 ) . For thrS mRNA , which bind the 30S in a single step process and for which unfolding is not necessary , the domains 1 to 3 of S1 are essential and sufficient to promote the formation of the active 30SIC ( Figures 5B and S5B ) . Indeed , 70% of the 30SIC is formed with S1Δ4–6 , whereas the additional deletion of domain 3 causes a strong reduction to 40% . Thus , the ability of S1 to stimulate the binding step of thrS mRNA is sustained by the three first N- terminal domains of S1 . Similar data were obtained for rpsOSD ( enhanced SD ) and rpsO mRNAs ( Figures 5C–D and S5A and S5C ) where the structure of the pseudoknot needs to be unfolded on the 30S to be positioned into the decoding channel . The deletion of domains 5 and 6 only slightly affect the formation of 30SIC , while the additional deletion of domain 4 decreases the 30SIC yields to 50% and 60% for rpsO and rpsOSD , respectively . The removal of domains 3 to 6 completely abolished the formation of 30SIC for rpsO mRNA , whereas a residual signal of 30% was observed for rpsOSD . The enhanced SD compensates the lack of S1 for the binding step as demonstrated by filter binding assays ( Figure S1B ) , but a minimal core of S1 ( domain 1–3 ) is still important to promote the unfolding/accommodation second step . Noteworthy , S1Δ3–6 binds efficiently to the 30S but with impaired functions , suggesting that domains 1 to 3 are essential for all the steps including the binding of thrS and rpsO mRNAs , and the accommodation of rpsO mRNA . All in all , these data show that both 30S-dependent activities of S1 , the docking of mRNA carrying weak SD ( as for thrS and rpsO ) and the unfolding of structured mRNA and accommodation into the mRNA channel ( as for rpsO and rpsOSD ) , require the first three OB-fold domains of r-protein S1 . Hence , domains 1 to 3 constitute the minimal protein that retains most of the S1 functions with respect to structured mRNAs .
The ability of isolated r-protein S1 to unwind model RNA duplexes or helices has been well documented [2] , [27]–[31] . However , it was not yet demonstrated that S1 would be the key r-protein to unfold mRNA structures on the ribosome . In this study , we have monitored the action of r-protein S1 on the natural structured and regulated E . coli rpsO mRNA encoding r- protein S15 during the formation of the 30SIC . This mRNA carries a pseudoknot structure within the RBS , which is recognized by the 30S for translation [15] . It sequesters the beginning of the coding sequence through base pairings that need to be melted for the formation of the codon–anticodon interaction [13] , [15] . We demonstrate here that r-protein S1 and primarily its three OB-fold domains 1 to 3 are essential for the accommodation process allowing rpsO mRNA to unfold and to relocate its initiation codon into the decoding center . Using 2-AP–modified rpsO mRNA , we were able to follow the S1-dependent melting of the pseudoknot directly on the ribosome . We could also compare the S1 RNA melting activity isolated or on the ribosome ( Figure 2 ) . Using a combination of approaches , we show that the fluorescence emission does not result from the interaction of the mRNA on the 30S but is primarily due to the melting of the pseudoknot structure ( Figures 2A and 6C ) . In its natural ribosomal context , the melting activity of S1 is clearly more pronounced and is independent of the presence of the initiator tRNA . This enhanced activity on the 30S could be explained by different conformations of S1 when free in solution or anchored to the ribosome where the OB-fold domains 1 to 3 would be orientated in an optimal way to interact with rpsO mRNA . An alternative explanation is the possible contribution of other ribosomal components to the S1-dependent unfolding process . A recent single-molecule study demonstrated that isolated r-protein S1 is able to melt in a multistep process a large artificial 274 bp stem-loop structure by binding to an upstream single-stranded RNA region [31] . This elegant study showed that S1 binds to the transient open form of the helix-unpaired junction region and stabilizes the open form to promote the local melting of the base pairs . This model is consistent with our data that are obtained on a natural structured mRNA . We propose that the three first domains of S1 bind successively to the A/U-rich connecting loop next to the long-range interaction allowing S1 to bind to the transiently opened base pairs . This mechanism would then lead to pseudoknot unwinding . The rate ( 0 . 9 s−1 ) at which the pseudoknot conformational change takes place on the 30S is rather slow as compared to the rates determined for other events of the translation initiation pathway [1] . It could thus represent the rate-limiting step of the initiation of structured mRNA as it was previously proposed [44] . Our data also indicate that the initiator tRNA is not essential for the RNA melting process . However , the formation of the anticodon–codon interaction is critical to stabilize rpsO mRNA into the channel of the 30S . Using three E . coli natural mRNAs ( sodB , thrS , rpsO ) , we demonstrate that r-protein S1 acts differently according to the nature of the signals present in the 5′UTR of mRNAs to form the 30SICs . Indeed , S1 is dispensable for the formation of the initiation complex involving sodB mRNA , which contains a strong SD and a weakly structured RBS ( Figure 6A ) . In the second example , S1 is required for the docking of thrS mRNA onto the ribosome in a single step process and the replacement of its weak SD with a stronger one alleviates the requirement of S1 for the formation of 30SIC ( Figure 6B ) . Finally , S1 is required for the recruitment of rpsO mRNA through its pseudoknot structure and for the accommodation process allowing the mRNA to occupy the decoding channel ( Figure 6C ) . Noteworthy , pseudoknots were preferentially selected as strong binders of E . coli ribosomes or of free r-protein S1 , while SD-containing unstructured mRNAs were selected against S1-depleted 30S ribosomes [45] . Hence , the complexity of mRNA structure within the RBS would direct the choice of the S1 actions to promote the formation of active 30SIC ( Figure 6 ) . In addition , we show that the action of S1 can be prevented by repressor proteins such as r-protein S15 , which binds to rpsO pseudoknot and prevents its melting onto the ribosome to repress translation ( Figure 6 ) . One can predict that other translational regulatory proteins would interfere with the action of S1 onto the ribosome . This variety of mechanisms is consistent with the fact that S1 is weakly associated to the 30S subunit . In agreement with this observation , a subpopulation of ribosomes lacking S1 was suggested to co-exist in E . coli under normal growth conditions [46] . Furthermore , the overexpression of rpsA led to the dissociation of leaderless mRNAs from the ribosomes [47] . This was supported by the fact that the overproduction of S1 slightly enhanced the occupancy of the ribosomes , suggesting that the WT levels of the protein did not saturate the ribosomes [48] . Under stress conditions , subpopulations of ribosomes were recently isolated in vivo , which selectively translated leaderless mRNAs [49] , [50] . Altogether , it is tempting to speculate that the absence of S1 on the ribosome might confer selectivity for specific mRNAs with strong SDs and unstructured RBSs , such as sodB mRNA or leaderless mRNAs . Thus , S1 confers to the ribosome the ability to dynamically adapt to the sequence and structure of mRNAs , increasing ribosome plasticity . This might help the ribosome to coordinate and fine-tune the rate of protein synthesis . S1 belongs to the family of RNA-binding proteins composed of multiple RNA-binding motifs . It contains six OB ( oligonucleotide/oligosaccharide-binding ) fold domains that are connected by short linkers ( Figure 5A ) . We show here that these domains exhibit distinct but also synergistic functions . We first demonstrated that the two N-terminal domains are critical to anchor S1 onto the 30S subunit ( Figure 5A ) . Numerous studies supported the localization of S1 on the 30S platform where it makes contacts with mRNAs and r-proteins [18] , [51]–[58] . More precisely , domain 1 of S1 was shown to interact with the coiled-coil domain of r-protein S2 [59] . In addition , we show that domain 2 and to a much lesser extent domain 3 enhance binding of S1 to 30S ( Figure 5A ) . This would suggest that other ribosomal components contributed to precisely position S1 on the 30S platform so that domains 4 to 6 would be exposed to the solvent to recruit specific mRNAs at the initiation step . Domains 1 to 3 of S1 are essential and sufficient to promote the formation of active 30SIC involving either thrS or rpsO , while domain 4 exerted a stimulating effect only on rpsO , providing additional interactions required for full biological function . This is well correlated with the in vivo data since successive deletions of the OB-fold domains had an increasing effect on cell growth . Indeed , the two last C-terminal domains 5 and 6 affected growth rate in a limited way as it was previously shown [60] , while the deletion of domains 4 to 6 permitted growth at extremely slow rates and any further deletions ( Δ3–6 , Δ2–6 ) caused complete lethality ( Figure 4 ) . This effect on cell growth can be explained by the fact that the truncated proteins are still able to bind to the ribosome , while the recruitment and/or the accommodation of essential mRNAs is presumably strongly perturbed . Although domain 1 has been mainly described as the 30S binding site , we show here that this first N-terminal OB-fold domain is also critical for rpsO mRNA binding ( Figure 3C ) . Other studies revealed that various RNA substrates bind to the same surface area of a protein carrying domains 3 to 5 [61] . In addition , domain 3 with either domain 2 or domain 4 of S1 confer high affinity through cooperative contacts with RNAs [48] . Directed evolution of S1 to enhance translation of GC-rich mRNAs in E . coli selected mutations primarily in domains 3 and 4 [62] . Hence , the flexibility of the domains respective to each other might confer to S1 a high adaptability to bind a large variety of RNA substrates . The work presented here provides the notion that the six domains of S1 are not functionally equivalent , although they are structurally related with respect to a common fold . The deletion of the two last C-terminal domains of S1 had no major effects on cell growth , indicating that they are not required for translation [60] . In addition , deletion of domain 6 did not affect the translation and autoregulation of rpsA [63] . However , the absence of the C-terminal domain causes a cold-sensitive phenotype most likely due to an impaired ability to melt RNA structures stabilized at low temperature . The fact that mutations could alter the chaperone activity preferentially at low temperatures is not so surprising . Indeed , S1 r-protein does not use energy like other RNA helicases , and therefore at the permissive temperature , the thermal energy may help the protein to melt RNA secondary structures . It could also be possible that domains 5 and 6 contribute to the translation of specific mRNAs as it was previously proposed [60] , [64] . In conclusion , this study shows that r-protein S1 confers a chaperone activity to the 30S subunit that promotes the active docking and accommodation of structured mRNAs into the decoding channel . In addition , the data are indicative of a hierarchy of mRNA targets with respect to S1 recognition on the ribosome . Because S1 is essential in E . coli , phylogenetic analysis may shed light on how the S1 functions have evolved among bacteria . A phylogenetic study has been carried out on r-protein S1 based on structural signatures present within each OB-fold domain [42] . This analysis revealed that S1 from Gram-negative bacteria ( proteobacteria , chlamidiae , spirochates , bacteroides , aquificae ) , thermotogae , chloflexi , and high G+C content Gram-positive bacteria ( actinobacteria ) contained at least the four first domains , suggesting that most of the activities of S1 would be preserved in these organisms . Although the actinobacteria , such as Micrococcus luteus , contained an additional fifth domain different from E . coli S1 , M . luteus S1 was able to substitute E . coli S1 on the ribosome to translate mRNAs with weak SD [23] , [65] . Another group of bacteria including the firmicutes , tenericutes , and cyanobacteria contained shorter forms of the protein with a first N-terminal domain that differs greatly from E . coli S1 , questioning the ability of these proteins to bind to the ribosome . Two S1 homologues containing three OB-fold domains were identified in Synechococcus . One of these homologues was able to bind the ribosome and was found to be essential for the translational initiation of several mRNAs [66] . In B . subtilis , S1 protein is not essential [67]–[69] , consistent with the fact that the protein plays no major role in translation [23] , [42] , [70] . In these Gram-positive bacteria , most of the mRNAs carry a strong SD sequence , and the low G+C content of their genomes may disfavor the formation of very stable mRNA structures , which might obviate the need for S1 melting activity on the 30S . Whether these truncated forms of S1 act as RNA chaperones outside the ribosome remains to be studied . It would also be of interest to analyze how the functions of S1 have evolved , and what are the strategies used by the ribosomes to translate structured mRNAs , in the low GC content Gram-positive bacteria .
All strains and plasmids , which have been used and constructed in this study , are given in Table S1; the oligonucleotides ( oligos ) used for cloning and for mutagenesis are given in Table S2 . Experimental details for the constructions of the strains are given in the Text S1 . WT thrS , thrSSD ( −195 to +65 nts , +1 being the A of the thrS translational initiation codon ) , WT rpsO and rpsOSD ( −120 to +65 ) transcripts were prepared in vitro by T7 transcription of linearized plasmids ( see [36] for thrS and [37] for rpsO constructs ) . WT sodB mRNA ( −55 to +64 nts ) was transcribed from the PCR product on the genomic DNA of E . coli MG1655 using the appropriate oligonucleotides ( Table S2 ) . The psk and mut-psk ( −56 to +12 , +1 being the A of the rpsO initiation codon ) RNA fragments have been transcribed from linearized plasmids ( Table S1 ) . The 5′ end-labeling of dephosphorylated RNA or of the chemically synthesized RNA was performed with T4 polynucleotide kinase and [γ-32P]-ATP [71] . All RNAs were purified on 8% polyacrylamide-8 M urea slab gel electrophoresis ( PAGE ) . Before use , mRNAs were renatured as follows: incubation at 90°C for 1 min in RNase-free water , cooled in ice for 1 min , and at 25°C for 30 min in the appropriate buffer containing monovalent ions and MgCl2 . Predictions of the SD/aSD stabilities were obtained using RNAcofold of the Vienna package [72] . Wild-type and mutant rpsA genes were cloned in vectors pET23a or pDEST14 , and the plasmids were transformed into E . coli strain BL21 ( Table S1 ) . The proteins carrying six histidines at their C-terminus were purified using an affinity chromatography followed by a monoQ ( for details , see Text S1 ) . We have verified by mass spectrometry that S1 was homogeneous and was not contaminated by E . coli Hfq ( see Text S1 ) . E . coli 30S subunits were purified on sucrose gradients after dissociation of the 70S [73] . Ribosomal protein S1 was removed from the 30S using a polyU-sepharose 4B column ( Text S1 ) . The formation of a simplified translational initiation complex with mRNA ( toeprinting assays ) was done according to Fechter et al . [73] . Experimental details are given in Text S1 . RNA fragments ( psk-rpsO , G-56 to U12 ) containing two 2-AP nucleotides ( A-40 and A-42 ) were synthesized on Pharmacia Gene Assembler or Applied Biosystems instrumentations using 2′-O-TOM protected phosphoramidite nucleoside building blocks [74] . All experiments were measured on a Kintek SF-2400 stopped-flow device at 37°C . The renatured psk-rpsOSD mRNA ( 50–100 nM ) present in 20 mM Tris-HCl pH 7 . 5 , 60 mM NH4Cl , 1 mM DTT , 7 . 5 mM MgCl2 was placed in one of the two syringes just before the experiment . The r-protein S1 , 30S , 30S−S1 , or 30S/fMet-tRNA ( 1 or 2 µM ) in the same buffer were introduced in the second syringe . The protein and the 30S were incubated at 37°C for 20 min before their injection in the mixing chamber . The melting of the pseudoknot psk-rpsOSD was monitored by measuring the increment of the fluorescence signal after passing the samples through KV408 filters ( Schott ) at 405 nm , generated by the 2-APs excited at 308 nm . The kfast and kslow values obtained by double exponential fitting were obtained with the Prism Graphpad software . Purified WT and mutant proteins S1 ( 150 pmoles ) were incubated with 30S−S1 ( 50 pmoles ) for 15 min at 37°C in 20 µl of 20 mM Tris-HCl pH 7 . 5 , 60 mM KCl , 40 mM NH4Cl , 10 mM MgCl2 , 3 mM DTT , and 0 . 02 mg/ml BSA . After purification on a Superdex 200 HR 10/30 , the fractions containing the 30S or S1 were analyzed on a 4%–12% SDS-PAGE , and visualized by Western blots using antibodies against the His-tag of each S1 ( Text S1 ) . Protein S1 was pre-incubated for 15 min at 37°C in the S1 buffer containing 20 mM Tris-HCl pH 7 . 5 , 10 mM MgCl2 , 60 mM KCl , 40 mM NH4Cl , 3 mM DTT , and 0 . 02 mg/ml BSA . Complex formation was performed at 37°C for 15 min with the renatured 5′ end-labeled RNA ( 12 , 000 cpm ) and increasing concentrations of r-protein S1 in 10 µl of S1 buffer .
|
Gene expression is regulated at multiple levels , including the decision of whether or not to translate a mRNA . This phenomenon , known as translational regulation , allows rapid changes in cellular concentrations of proteins and is well suited to the adjustment of cellular growth in response to stress and environmental changes . Many bacterial mRNAs adopt structures in their 5′ untranslated regions that modulate the accessibility of the mRNA to the small ribosomal 30S subunit and so are directly involved in this regulatory process . Structured mRNAs must interact with the 30S subunit in a two-step pathway whereby the docking of a folded mRNA is followed by an accommodation step that involves unfolding of these structures . However , it is not known how the ribosome unfolds mRNA structures to promote translation initiation , nor which ribosomal factors are responsible for this activity . We demonstrate that the first three domains of ribosomal protein S1 endow the 30S subunit with an RNA chaperone activity that is essential for the binding and unfolding of structured mRNAs , allowing the correct positioning of the initiation codon for translation . However , ribosomal protein S1 is not required for all mRNAs and acts differently depending on the type of regulatory elements present in a given mRNA . In all , we have shown that ribosomal protein S1 provides an RNA-melting activity to the exit site of the 30S decoding channel and confers some plasticity on the ribosome to initiate translation of mRNAs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2013
|
Escherichia coli Ribosomal Protein S1 Unfolds Structured mRNAs Onto the Ribosome for Active Translation Initiation
|
High-throughput in vitro methods have been extensively applied to identify linear information that encodes peptide recognition . However , these methods are limited in number of peptides , sequence variation , and length of peptides that can be explored , and often produce solutions that are not found in the cell . Despite the large number of methods developed to attempt addressing these issues , the exhaustive search of linear information encoding protein-peptide recognition has been so far physically unfeasible . Here , we describe a strategy , called DALEL , for the exhaustive search of linear sequence information encoded in proteins that bind to a common partner . We applied DALEL to explore binding specificity of SH3 domains in the budding yeast Saccharomyces cerevisiae . Using only the polypeptide sequences of SH3 domain binding proteins , we succeeded in identifying the majority of known SH3 binding sites previously discovered either in vitro or in vivo . Moreover , we discovered a number of sites with both non-canonical sequences and distinct properties that may serve ancillary roles in peptide recognition . We compared DALEL to a variety of state-of-the-art algorithms in the blind identification of known binding sites of the human Grb2 SH3 domain . We also benchmarked DALEL on curated biological motifs derived from the ELM database to evaluate the effect of increasing/decreasing the enrichment of the motifs . Our strategy can be applied in conjunction with experimental data of proteins interacting with a common partner to identify binding sites among them . Yet , our strategy can also be applied to any group of proteins of interest to identify enriched linear motifs or to exhaustively explore the space of linear information encoded in a polypeptide sequence . Finally , we have developed a webserver located at http://michnick . bcm . umontreal . ca/dalel , offering user-friendly interface and providing different scenarios utilizing DALEL .
The notion that the information encoding molecular recognition could be linearly encoded in peptides has changed the way we have approached the study of protein-protein interactions both experimentally and computationally [1] . Linear information encoding recognition is typically represented using consensus sequences , i . e . motifs , which define positions and compositions of the residues contributing to both specificity and affinity of recognition [2] . There is a pressing need and great interest in deciphering the linear information encoding recognition for increasingly diverse families of domains involved in all cellular processes [3] . This information is essential to constructing both empirical and quantitative models of biochemical networks [4] . Constructing such models has broad applications to both predicting behaviors of pathways , synthesizing novel pathways to create novel biochemical processes or designing inhibitors of specific cellular processes [5] . For two decades , high-throughput in vitro screening methods based on combinatorial peptide chemistry have been utilized to explore the linear amino acid sequence information encoding recognition of linear peptides and to reveal principles that underlie their selective binding to individual domains [6] . These methods involve in vitro screening of target domains against peptide libraries to obtain large number of hits , which are then aligned to find positions and compositions of the residues involved in recognition . However , the observation of a protein binding to a peptide in vitro does not guarantee binding in vivo of the same protein to all proteins encoding the same peptide . There is also the risk of finding peptides that are not found in the cell [4 , 7] . Several factors can cause binding to differ between in vitro and in vivo measurements: in vitro detection of domain-peptide interactions are outside of their biological context; that is , binding peptides within a protein are often optimized to bind in vivo with concomitant contributions of additional contacts , cooperativity or post-translational modifications [8–10] . There is , thus , always the possibility that contextual information about the binding of a particular linear peptide to a protein is missing from in vitro binding data [6] . Moreover , in vitro methods are limited in the complexity of libraries , i . e . number and length of peptides that can be screened , thus limiting the space of linear information that can be explored . For instance , we curated the literature for experimentally validated SH3 domain interactions in yeast , and we found that more than half of the binding sites determined in vivo were not currently predicted from in vitro experiments [11] . During the past two decades , different classes of methods have been developed in attempts to address these issues . The first class of methods were designed to search specifically for annotated motifs , like those derived from the Eukaryotic Linear Motif ( ELM ) database [12–16] . For example , Stein and Aloy identified instances of peptide-mediated protein interactions of known 3D structure based on the collection of motifs in the ELM database [16] and explored individual contribution of motifs and context to global binding energy [14] , while Mooney et al . applied machine learning techniques to predict motifs based on annotated instances from the ELM database [12] . Another class of methods were designed to search for motifs displaying properties previously observed in linear binding peptides , including three-dimensional structures , intrinsic disorder , sequence conservation , and solvent accessibility [17–22] . For example , Stein and Aloy used available structures of known protein-peptide interactions and scanned protein complexes of known 3D structures to identify new peptide-mediated interactions [17] . In another example , Davey et al . used the statistical significance of sequence conservation of small stretches of residues within intrinsically disordered regions to identify putative functional motifs [18] . Another class of methods relies on statistical models , such as hidden Markov model , Gibbs sampling and Nested sampling [23–26] . For example , Bailey and Elkan used Gibbs sampling and Expectation Maximization to fit a two-component finite mixture model to describe the sequences of interest , then predicted motifs by fitting the statistical model . In another example , Dogruel et al . used a Monte Carlo inference strategy , called Nested Sampling , to build a multi-class sequence background model to find non-motif parts of sequences , then build a set of position-weight matrices to represent motifs overrepresented in the sequences considered [25] . In a different example , Nguyen et al . used hidden Markov models to include insertion/deletion/substitution events within protein sequences , then applied the model to identify short linear motifs in the budding yeast Saccharomyces cerevisiae [23] . The last class of methods were designed to search overrepresented linear motifs in proteins . For example , Edwards et al . used a probabilistic method for identifying overrepresented and evolutionary convergent motifs in proteins [19] . In another example , Kelil et al . explored motifs of any length and composition and scored their enrichment using the hypergeometric distribution [27] . Crucial to computational prediction of linear binding sites is the selection of appropriate reference sequences from which statistical inference of discovered binding sites are made . A key advance of the method we describe here are improvements in the selection of such reference sequences . Among existing methods , DILIMOT [21] developed by Neduva and Russell , and FIREPRO [22] developed by Lieber et al . score motif enrichment in sequences of interest relative to a set of reference sequences . For DILIMOT , reference sequences are taken arbitrarily from the SWISS-PROT database , while in our method the reference sequences are carefully selected from non-binding negative control proteins . These include the sequences of all proteins that are experimentally shown to bind to one or more members of a family of proteins or protein domains but not to the specific member being tested . Use of such reference sequences should enable us to distinguish between enriched functional motifs and randomly recurring ones . For FIREPRO , reference sequences are also selected from negative control proteins . However , motif enrichment is scored using mutual information , while in our method motif enrichment is scored using the cumulative hypergeometric distribution . This approach offers a significant advantage over the mutual information in that it provides a direct and exact evaluation ( i . e . based on Fisher’s demonstration [28] ) of the statistical significance of the enrichment of a motif within the sequences of interest over the reference sequences . Despite the large number of methods available to date , including all of those cited above , the exhaustive search of motifs in groups of proteins has been so far physically unfeasible . In other words , no method to date allows one to exhaustively explore the entirety of possible linear binding motifs , including flexible residues , i . e . positions with preference for multiple amino acids , within a set of proteins of interest and for each motif , calculate a score that evaluates its enrichment relative to a background model . To address these issues , we have devised a strategy to exhaustively search for linear sequence motifs shared among proteins that form direct interactions with other proteins or folded protein domains , e . g . from yeast two-hybrid screens . Our strategy can also be applied to any group of proteins of interest to identify enriched linear sequences , or to exhaustively explore the space of linear information . Our method is designed to exhaustively search the entire space of all possible motifs . The search covers all possible sequences of any length and composition within the proteins of interest . Our strategy does not require any prior knowledge about sequence or structural consensus signatures of recognition peptides and can be applied to any group of proteins expected to share sequence motifs ( e . g . because they bind to the same partner ) . The strategy presented builds on previous work [27] with several notable improvements . DALEL indeed explores preferences for multiple amino acids at each position of a motif , allows combining the background dataset with a dataset of “negative sequences” , and uses a parallel suffix tree to enable the exhaustive search of variations in motifs , which was previously unfeasible due to the associated combinatorial explosion . To test DALEL we used the SH3 domains of budding yeast Saccharomyces cerevisiae as a test system because there are numerous sources of data on SH3 domains’ interactions , both in vivo with full-length proteins and in vitro with linear peptides . In total , we manually curated 890 protein-protein interactions from the literature , between the 25 yeast SH3 domains and 361 proteins encoding a total of 1073 experimentally verified SH3 binding sites ( i . e . linear peptide segments within the proteins ) , henceforth called “known SH3 binding sites” [11] ( S1 Table ) . Using DALEL we were able to identify the majority of previously discovered SH3 binding sites . We have also identified previously unreported sites with non-canonical sequences and properties that may bind to SH3 domains in distinct ways or serve indirect ancillary roles in peptide recognition and binding . Finally , in an application of DALEL , we have discovered a remarkable relationship between SH3 domain binding motif thermodynamic , evolutionary properties and functional specificity of the proteins that have given motifs [11] . Application of our approach to other domain-linear peptide interactions may reveal similar relationships , establishing a framework for predicting functional organization of protein interactomes .
Before describing the methodology for enumerating linear motifs , we first summarize how binding specificity of the motifs for a family of proteins or protein domains is assessed . Our strategy is based on the premise that proteins known to bind to a common target domain are enriched in peptides encoding the linear information necessary to recognize that domain , while all other proteins within a cell or organism do not exhibit such enrichment . In our strategy , we exhaustively enumerate all possible consensus motifs within a set of proteins that bind to one or more types of proteins or protein domains . Our approach is closest to DILIMOT and FIREPRO , and part of its originality comes from the way we define the background model . For each target ( one member of a protein or protein domain family ) , we partition the proteome into three distinct sets . The “positives” are proteins that bind to the target . The “negatives” are proteins that do not bind to the target but do bind to at least one other member of the same family ( e . g . a different SH3 domain ) . The “background” consists of all other proteins in the proteome that bind neither to the target nor to any of the other members of its family . We calculate then two p-values for each motif using the cumulative hypergeometric distribution ( Methods ) : ( i ) pNEG quantifies motif enrichment in the positives over the negatives , and ( ii ) pBAK quantifies motif enrichment in the positives over the background . In other words , pNEG reflects specificity of the motifs for the target domain relative to other domains of the same family , while pBAK reflects specificity for the target domain relative to motifs found in the background . Thus , a motif with strong pNEG and pBAK means high specificity for the target domain . The significance of pNEG and pBAK are evaluated by calculating their z-scores , and the less significant of both z-scores is assigned to the motif ( Methods ) . Consequently , when there are not enough ( e . g . < 20 ) well-defined positives and/or negatives , the z-scores will inherently be weak and the utility of DALEL will be limited . Our strategy exploits suffix trees , which allows for enumerating motifs in sequences in time that is linear with their length and number [29] . Our algorithm searches for all possible motifs comprising any number and combination of wildcards e . g . X in consensus motif PXXP . Theoretically , the number of possible combinations of wildcards in a motif of length l equals 2l ( e . g . each position is a wildcard or not , l-times ) , hence , the number of motifs of length l in a set of protein sequences is proportional to 2l , which require a suffix tree with size O ( 2l ) to represent them all . The exponential growth of the suffix tree with motif length makes exhaustive search for motifs unfeasible . For instance , in S . cerevisiae , the suffix tree required to represent all possible motifs present in SH3 binding proteins rapidly exceeds physical memory ( S1 Fig ) . To address this problem , we devised an algorithm to exhaustively search for motifs by dividing the O ( 2l ) suffix tree into 2l smaller suffix trees that can be explored sequentially , resulting in a linear increase in memory usage with motif length ( S2 Fig ) . Such an approach also allows for faster and parallelizable searches . The first step of the algorithm consists of using a sliding window to scan the sequences of all positives for linear peptides of a desired length ( Fig 1A ) . The obtained peptides ( Fig 1B ) are passed through a set of masks , each one representing one of the possible combinations of wildcards ( Fig 1C ) . Each mask is used to find all possible motifs present in the positives and matching wildcard conformation defined by the mask ( Fig 1D ) . Then we build the suffix tree of each set of motifs ( Fig 1E ) . The number of occurrences in the positives of each motif is immediately available after the completion of the construction of the tree , making it possible to further reduce the size of the tree by removing branches corresponding to motifs with occurrences less than a desired minimum ( Fig 1F ) . This optimization contributes significantly to accelerating the next step , which is the slowest in the algorithm . Finally , we use a sliding window to scan each protein in the proteome for peptides of a desired length ( Fig 1G ) and we match the peptides with each suffix tree to obtain the number of occurrences of each motif in the negatives and the background ( Fig 1H ) . Among peptides known to bind to a common domain , only about one-third of the residues are typically essential for recognition [30 , 31] , e . g . “P” in the motif PXXP binding to SH3 domains [32] . Other residue positions can be any ( X ) or a small number ( […] ) of amino acids , e . g . [RK] in [RK]XXPXXP and PXXPX[RK] motifs binding to diverse SH3 domains [33] . They also often display correlated preferences for amino acids at distinct positions , i . e . dependence between distinct positions for their amino acids preferences [34] , e . g . [ST] in R[ST][ST]SL peptides binding to Fus1 SH3 domain [35] . The strategy for exhaustive search for linear consensus motifs consists of finding all possible variants of each motif found in the suffix tree analysis by substituting wildcards by brackets , including all possible combinations of amino acids , e . g . [IVL] or [DE] . However , exhaustive search for variations in motifs is physically unfeasible because the combinatorial space is too vast . For instance , the number of possible combinations for a single motif with 4 wildcards is on the order of 1024 . Theoretically , the number of ways of picking k amino acids is Ck20=20 ! /k ! ( 20−k ) ! ( i . e . all combinations of 1 , 2 , 3 , … , and 20 amino acids ) , and then the number of ways of picking all combinations of 1 to 20 amino acids is ∑k=120Ck20 ( e . g . PXXXP has 1 x Ck20 x Ck20 x Ck20 x 1 possible variations ) . Consequently , the total number of all possible combinations of amino acids for a motif including n wildcards is ∏n∑k=120Ck20 . Thus , with each additional wildcard the upper bound of the number of possible combinations is multiplied by ∑k=120Ck20≈106 . For this reason , we devised a strategy where we exhaustively search for all possible variants of each motif that improve its p-values ( i . e . pNEG and/or pBAK ) , by iteratively substituting each wildcard by all combinations from 1 to 20 amino acids and testing for improvement of the motif p-value with each iteration of amino acid substitution ( Fig 2 ) . The substitution of a wildcard starts with a first iteration ( Fig 2B ) in which a wildcard is substituted by each single amino acid , and substitutions that improve the p-value are retained for the next iteration ( Fig 2B . i ) . At the next iteration , for each substitution retained , we add each remaining amino acid one by one and new substitutions that improve p-values are retained for the next iteration ( Fig 2B . ii ) . Similarly , we continue at each iteration , to add remaining amino acids to each substitution that improves the p-value ( Fig 2B . iii ) . In addition , throughout the iterations , for any substitution that improves the p-value , the other wildcards in the motif are substituted in their turn in the same way ( Fig 2C ) . The goal is to exhaustively explore all possible variants of a motif that improve its p-values and that present correlated preferences for specific groups of amino acids at distinct positions . Finally , when there is no further improvement in p-value with further substitutions , the variable amino acids at a position are retained . To benchmark the method , we measured how well it could recapitulate known protein binding motifs and unforeseen motifs . Unforeseen motifs are evaluated by comparing general sequence characteristics to those of known SH3 binding sites , i . e . sequence conservation , intrinsic disorder , solvent accessibility , and predicted binding energy . For each SH3 domain , we thus selected from each of the sequences of positives , those motifs that we discovered with the best z-scores and that covered a total length comparable to that of known SH3 binding sites . The procedure yielded 377 motifs from the positives for all SH3 domains except for that of the protein Cdc25 , for which the available experimental data was insufficient ( S2 Table ) . We then calculated the overlap between these motifs and known SH3 binding sites . We found that on average , 70% of the amino acids covered by the motifs we discovered were within known SH3 binding sites , and 88% were within 10 amino acids from these sites ( Fig 3 ) . Among the 377 selected motifs , 163 ( ∼ 43% ) matched the canonical SH3 binding consensus motif PXXP , and 52 ( ∼14% ) had sequences longer than 10 amino acids . This latter result is consistent with observations showing that peptides must be longer than the consensus motif for optimal binding affinity and specificity [36] . Further , 307 motifs include positions with variable residues , among which 134 include correlated residues , supporting that cooperativity among residues in binding motifs is common [34] . We found also that on average , ∼35% of the amino acid positions in consensus motifs are wildcards , which is consistent with previous observations on SH3 binding sites [30 , 31] . We further analyzed four properties of each of the 377 selected SH3 domain binding motifs , including their binding energy , solvent accessibility , sequence conservation , and intrinsic disorder . We found that 153 of the motifs had similar properties to known SH3 binding sites for the four properties considered , but only 8 had properties that were significantly different ( Table 1 ) . We categorized the motifs we discovered into three distinct classes: class I motifs exhibit properties similar to known SH3 binding sites in terms of the four properties and have high overlap ( > 80% ) ; class II motifs are highly similar to known SH3 binding sites for the four properties but with less than 20% overlap . These appear to be bona-fide SH3 binding sites that may have been missed by previous studies . Finally , class III motifs cover sequences outside of known SH3 binding sites and display markedly different structural and evolutionary conservation properties ( Table 1 ) . Nevertheless , motifs in class III exhibit high specificity for their corresponding SH3 domains , suggesting that they may represent non-canonical binding sites or may be indirectly involved in binding . We verified if class III represents coincidental motifs that are enriched in the positives but involved in binding to other domains , as proposed by Edwards et al . [37] . For this , we searched whether the motifs we discovered matched known ELM motifs [16] . We searched also instances of these motifs in protein-protein interaction databases ( domain-peptide interactions involving binding sites matching our motifs ) , e . g . yeastgenome . org , uniprot . org , ncbi . nlm . nih . gov , and thebiogrid . org . We did not find any instance or similarity for the motifs in class III . This result suggests that there exist peptides with distinct properties that may serve ancillary roles in determining SH3 domain binding to proteins . In addition , these motifs may predict interactions that to date could not be determined with existing methods . Several studies have demonstrated the existence of non-standard SH3 binding sites [38–48] . For instance , the SH3 domain of the yeast protein Fus1 has been shown to recognize peptides belonging to the consensus motif R[ST][ST]SL [35] . To date , 25 SH3 binding sites for the Fus1 SH3 domain , including members of the R[ST][ST]SL motif and others that are not members of any consensus , have been experimentally detected in 22 different proteins . We compared our discovered motifs to standard motifs , those already reported in the literature [4] , in the prediction of binding sites recognized by the Fus1 SH3 domain ( Fig 4 , Table 2 ) . To perform the comparison , we first selected 2 motifs , R[ST]X[SW]L and RX[ST]SL , on the basis of their z-scores and because they covered a total length in the 22 target proteins at most equal to the total length of the 25 known binding sites . In fact , the goal here was to limit the number of selected motifs to the minimum , then assess our ability to predict the 25 known SH3 binding sites of FUS1 SH3 domain ( Fig 4 ) . The experimentally characterized motifs included R[ST][ST]SL , [RK]XXPXXP and PXXPX[RK] and 15 other motifs determined by phage display screening [4] . We found that the two selected motifs are present in all 22 binding proteins of Fus1 SH3 domain but standard motifs were present in only 14 . These two motifs were more statistically significant than all experimentally characterized motifs . Importantly , our motifs matched to 37 . 65% of the total length of known SH3 binding sites of the Fus1 SH3 domain , while the standard motifs , matched to only 18 . 98% . Furthermore , the proportion of our motifs found in the entire proteome and negative reference proteins was lower than that of standard motifs . Our motifs were indeed found in only 40 among the 571 negative reference proteins and in 274 among 5356 sequences from the proteome , while the standard motifs were found in 271 of the negatives and 1862 proteins in of the rest of the proteome . Thus , the motifs we discovered , R[ST]X[SW]L and RX[ST]SL , match the known motif R[ST][ST]SL , but suggest a broader range of substitutions for recognition by the Fus1 SH3 domain , which is consistent with what has been previously proposed [35] . These results suggest that our strategy has helped in redefining a standard motif involved in binding to the Fus1 SH3 domain , without any prior knowledge of their positions or signatures in proteins or what they should look like . The largest number of SH3 domain interactions have been determined in high-throughput screening studies [4 , 49 , 50] . Among the 922 unique ( i . e . one instance ) SH3 binding sites that we found in the literature ( Fig 5 ) , 904 ( ∼98% ) were identified in vitro , while only 40 ( ∼ 4% ) were identified in vivo among which only 18 ( ∼45% ) were also verified in vitro . Despite the large spectrum of binding sites discovered in vitro , they cover less than half of the peptides that we discovered based on all interactions determined in vivo , which suggests that we are still missing a significant number of SH3 binding sites in the yeast proteome . Lack of complete coverage of in vitro peptide binding data may be due to , for example , lack of adjacent sequence or of additional interactions that are required for a given peptide within a protein to bind to an SH3 domain [51] . In contrast , the 377 motifs we discovered captured 724 ( ∼80% ) SH3 binding sites among the 904 determined in vitro and of among the 40 SH3 binding sites determined in vivo , they captured 35 ( ∼87% ) . This result highlights the strength of our approach , which discovers equally well both types of sites ( Fig 5 ) . The degenerate nature of binding peptides makes them hard to detect because they are immersed in a background of mostly irrelevant peptides . For this reason , many approaches reduce the search space using a priori knowledge such as canonical motifs . For SH3 binding sites , this includes polyproline peptides that encompass the PXXP core motif or the canonical [RK]XXPXXP and PXXPX[RK] motifs [52 , 53] . Additional properties of a given sequence may be taken into account; for example , tertiary structures conforming to predefined structural templates [14 , 54 , 55] or their presence in regions that are conserved , solvent accessible , or intrinsically disordered [56 , 57] . Here we saw that with our strategy , we could accurately detect recognition peptides without using any such knowledge a priori . This enabled us to identify peptides such as the structured beta-sheet ubiquitin-like domain of UBI4 recognized by SLA1-3 SH3 domain [48] , non-canonical peptides recognized by the Fus1 SH3 domain [35] , and additional non-canonical sequences [38–48] . The absence of any predefined model for the motifs we discovered enabled us to describe , a posteriori , the biological properties of corresponding peptides within proteins . We refer here to peptide sequences within proteins that correspond to the 377 discovered motifs predicted to be involved in SH3 binding . Among predicted and known SH3 binding sites , we found that ∼20% of both of them do not belong to the PXXP core motif . Moreover , among known SH3 binding sites , ∼30% of those determined in vivo and ∼50% of those determined in vitro do not fall into any of the two canonical motifs [RK]XXPXXP and PXXPX[RK] ( Fig 6C ) . Consistent with these results , ∼75% of the SH3 binding sites that we predicted also do not match these canonical motifs . We also found that >30% of all peptides belonging to our discovered motifs do not fall in intrinsically disordered regions ( Fig 6E ) . In addition , as expected , both predicted and known sites exhibit higher sequence conservation ( Fig 6D ) , solvent accessibility ( Fig 6F ) and binding energy ( Fig 6G ) than expected . Interestingly , we found that , in comparison to their flanking regions , like predicted binding sites , known binding sites exhibit stronger contrast in sequence conservation and binding energy and weaker contrast in solvent accessibility , while protein disorder in flanking regions is found to be as high as known binding sites . The contrast between SH3 binding sites and their background is in agreement with known model of SH3 domain recognition [42 , 51] . However , these trends vary greatly among binding sites , making it hard to discriminate between binding sites and the rest of the proteins based solely on these general properties . Interestingly , when we used our motifs , we found better contrast than that obtained with general properties between known binding sites and their background ( Fig 6B ) . To assess the power of the different properties in the prediction of SH3 binding sites , we tested each property as a discriminative feature in the identification of known SH3 binding sites ( Fig 6A ) . The prediction results showed that canonical motifs are the best to discriminate SH3 binding sites compared to binding energy , intrinsic disorder , solvent accessibility , and sequence conservation ( in decreasing order of their discriminatory power ) . Overall , our results highlight three essentials points: ( i ) SH3 binding sites exhibit common physical properties and sequence conservation , however , these properties are not exclusive to these sites; ( ii ) , although SH3 binding sites exhibit common properties , we have discovered a notable number of sites that have distinct properties; ( iii ) , consequently , as utilized in other methods to search for binding sites in protein sequences , physical properties can be used as a constraint on such searches , but will bias and limit a search and could result in false predictions; for example , of sites that have the right physical properties , but sequences that are not consistent with binding to a domain . In contrast , our exhaustive and unconstrained search strategy should not likely result in any bias , limitations or false-positive results in identifying known or unforeseen linear binding motifs ( Fig 6A ) . This is because our approach takes advantage of the most representative property of the SH3 binding sites: their linear information , i . e . linear signatures . These results imply that the number of motifs involved in SH3 domain sequence recognition is larger than generally appreciated . These results highlight that there has been and remains a pressing need for new methods to explore the full complexity of this range of possible binding sites for any given family of protein or peptide binding protein domains . The exhaustive search of linear information in groups of related proteins has been so far physically unfeasible by classic approaches [12–14 , 17–26] , henceforth this is made possible by the approach we presented here . The approach we applied here , demonstrates prediction performance of SH3 binding sites that is better than approaches that use constrains such as physical properties , sequence conservation , and motifs enrichment . By capturing this large breadth of sequence properties , we were able to discover extensive correlations of physical properties and conservation of motifs and binding specificity , but most strikingly , a correlation of functional and binding specificity that links functional diversity to the chemical and thermodynamic characteristics of SH3 domain-protein interactions [11] . With extensive application of DALEL to other binding domain-protein interactions , we may be able to establish a quantitative framework for predicting functional organization of protein interactomes . Our strategy should prove an important complement to future efforts to identify linear peptide binding sites within proteins based on simple binary in vivo protein-protein interaction measurements , i . e . a critical step forward in reconstructing protein interaction networks . In addition , while DALEL does not consider motifs with wildcards of variable length , this feature could be developed in the future . We compared the results of DALEL to those of well-established algorithms on the identification of experimentally determined SH3 binding sites of the protein Grb2 . The algorithms tested include iELM [58] that identifies linear binding peptides specifically in the human proteome; MEME [24] detects ungapped linear motifs of fixed-length using finite mixture model; GLAM2 [59] detects gapped linear motifs of variable-length using local sequence alignment allowing insertions and deletions; PRATT [60] finds conserved linear motifs; DRIMUST [61] , qPMS7 [26] , NESTEDMICA [25] , FIRE-PRO [62] , DILIMOT [21] , SLIMFINDER [63] , and MOTIFHOUND [27] find enriched linear motifs in proteins ( details in S3 Table ) . The Grb2 protein is among recognition domain proteins in human that attract the most interest ( http://thebiogrid . org ) , mostly because of its implication in a large number of protein complexes and canonical cell surface receptor signalling pathways associated with normal cell growth , proliferation and differentiation , and whose component proteins are mutated in a number of cancers ( www . proteinatlas . org; www . uniprot . org ) . The great interest in this protein has helped produce abundance of validated experimental data ( http://mint . bio . uniroma2 . it ) , which makes this case “gold standard” to evaluate the prediction power of our algorithm , and other algorithms . The Grb2 protein consists of a central SH2 domain flanked by two SH3 domains ( www . rcsb . org ) . The binding specificity of the Grb2 N-terminal SH3 ( N-SH3 ) domain have been studied in detail and a consensus canonical binding motif “PXXPXR” has been identified [64] . A total of 72 binding sites within 61 different proteins were experimentally determined to bind to the Grb2 SH3-N domain ( S4 Table ) . Here , we compared our algorithm and other algorithms on the identification of the 72 binding sites of the Grb2 N-SH3 domain . For each algorithm we selected the top scoring predicted sites such that they cover about 3% of the total length of the 61 Grb2 binding proteins , a coverage that is equivalent to that of the 72 known binding sites . The coverage could , however , be smaller than 3% if the algorithm did not return enough sites . For each algorithm , we show the percentage of the total length of the 61 proteins covered by the sites identified ( Fig 7 , blue bars ) , as well as the percentage of the total length of known SH3 binding sites covered ( Fig 7 , orange bars ) . We found that the sites identified by SLIMFINDER , DILIMOT , GLAM2 , iELM , MOTIFHOUND , and DALEL , exhibit the largest overlap with experimentally characterized binding sites , ranging from about 24% to 38% , among which DALEL obtained the highest overlap of 38 . 19% . It is interesting to note that iELM performed well , possibly because it integrates properties of binding motifs specifically observed in the human proteome . Also , MEME , that detects ungapped motifs in proteins did not perform well , while its modified version GLAM2 that allows gaps performed much better , because it is better suited for finding degenerate motifs , a property that is present in the binding sites of the Grb2 SH3-N domain . PRATT , however , performed poorly on this example , suggesting that we cannot rely solely on sequence conservation to find linear binding peptides in proteins . In addition , among the algorithms that find enriched linear motifs in proteins only DILIMOT , SLIMFINDER , and MOTIFHOUND obtained satisfactory results . It is important to highlight the difference between the results obtained by DALEL and MOTIFHOUND , because both find the same motifs . In DALEL , however , the wildcards are degenerated in each motif to find positions with preferences for multiple amino acids , and also positions with correlated preferences . As a result , DALEL identified 38% but MOTIFHOUND only 28% of Grb2 SH3-N binding motifs . This difference proves the discriminative power of our approach because of our strategy for exhaustive search of sequence degeneracy in motifs . Interestingly , the class I canonical motif “PXXPXR” that was reported as recognized by the GRB2 SH3-N domain was identified by DALEL with the significant z-score of 4 . 92 and p-value of 1 . 79x10-17 . We found the motif “PXXPXR” present in just 29 among the 72 known binding sites of Grb2 SH3-N domain , thus , about 60% did not contain the motif “PXXPXR” ( S5 Table ) . Surprisingly , DALEL , identified the class II canonical motif “PXXPXK” with the very significant z-score of 5 . 35 and p-value of 2 . 02x10-18 . This is surprising because this motif was reported to be recognized by the Grb2 C-terminal SH3 domain [65 , 66] . We found this motif present in 22 among the 72 known binding sites of the Grb2 SH3-N domain . More interesting , all of the 22 known binding sites that contain the “PXXPXK” motif did not contain the “PXXPXR” motif . We thus suggest a novel binding preference of the Grb2 SH3-N domain and a possible cross-reactivity between the two SH3 domains of Grb2 protein . In addition , it shows that the binding preference of the Grb2 SH3-N domain is larger and more complex than what was previously reported , because neither the “PXXPXR” or “PXXPXK” motifs explain the binding preference of the Grb2 SH3-N domain to all the 72 known binding sites . DALEL also identified the motif “PXX[PV]XXK” with a significant z-score of 5 . 93 and a p-value of 1 . 08x10-19 . This motif was present in 42 known binding sites ( compared to 22 for “PXXPXK” ) , This result illustrates the power of our algorithm in allowing preferences for multiple amino acids at specific motif positions . We analyzed the performance of DALEL on the identification of motifs exhibiting ambiguous positions using a benchmark . The benchmark relied on 8 biological motifs from the ELM database [16] , which we planted in protein sequences of the S . cerevisiae yeast proteome . The advantage of this approach is that there is no simulated data beyond the replacement of amino acids from the original sequence . Thus , sequence features such as tandem-repeats or low complexity regions were preserved . Degenerate forms of the 8 consensus motifs were planted in 20 protein sets , using 5 , 10 , 15 or 20 occurrences , with at most one occurrence being inserted per sequence . This gave a total of 640 sets ( 8 motifs x 20 sets x 4 planted occurrences ) , in which we subsequently searched for the planted motif . The search in a sequence set was considered successful when the top motif ( s ) matched the planted sequences with a precision and recall both above 0 . 7 ( Fig 8A ) . For each motif and each number of occurrences , the fraction of successful searches over twenty sets composed of different sequences was calculated and represented in bar-plots ( Fig 8B , Material and Methods ) . We compared the performance of DALEL to two algorithms: MotifHound and SLiMFinder . We used MotifHound because it is the closest method to DALEL that does not take into account ambiguous positions , and SLiMFinder because it exhibited the highest accuracy in a previous benchmark [27] . We tested several parameter configurations for SLiMFinder and kept that yielding the best overall results ( see Methods ) . In the best configuration , SLiMFinder successfully identified six of eight motifs present in 15 occurrences and two motifs present in 10 occurrences . MotifHound detected seven out of the eight motifs inserted in 15 occurrences , and three motifs present in 10 occurrences . DALEL , however , showed a significant improvement , as all motifs were correctly identified when 15 occurrences were inserted , and seven were identified when present in only 10 occurrences . Generally , all algorithms had difficulties identifying motifs when only 5 occurrences were planted , although DALEL showed the best results with 3 motifs identified out of the eight investigated ( SLiMFinder and MotifHound both only identified one out of the eight ) . Our benchmark shows that DALEL performs well in the discovery of motifs with ambiguous positions . Datasets used for this benchmark are available as supplementary material for comparing future improvements of existing algorithms as well as new algorithms . Despite our efforts to produce bias-free datasets for the benchmark , we cannot ignore the possibility that additional unknown biases may not be addressed , perhaps impeding the performances measured here . Our benchmark largely serves the purpose of testing algorithms that were designed with similar objectives under the same controlled conditions . The artificial nature of the benchmark does , however , have advantages , e . g . it allows exploring the impact of specific parameters such as overrepresentation of a motif . On the other hand , it also limits our ability to interpret the results in the context of biological motifs discovery . Future work will address these issues .
The p-values were calculated using the cumulative hypergeometric distribution , which estimates the probability to see at least k successes in a sample of size n picked from a population of size N including a total of m successes , and defined as follows: H ( k|N , m , n ) =∑i=km ( mi ) ( N−mn−i ) ( Nn ) ( 1 ) with ( ab ) =a ! b ! ( a ! −b ! ) ( 2 ) Above , N is the size of the proteome , n the number of positives , p the number of negatives , and by defining a success as a protein comprising the motif; m , k , and q are respectively the number of successes in the proteome , the positives and the negatives . We then calculate pPRO and pNEG as follows: pPRO=H ( k|N , m , n ) andpNEG=H ( k|n+p , k+q , n ) ( 3 ) We evaluate the significance of each p-value by calculating its z-score , which provides the number of standard deviations from the average of the p-values distribution , as follows: zPRO=Z ( −log10 ( pPRO ) ) andzNEG=Z ( −log10 ( pNEG ) ) ( 4 ) withZ ( x ) =x−x¯σ ( x ) ( 5 ) The transformation here converts the p-values from a linear to logarithmic scale , which makes it possible to distinguish between extremely small p-values . The z-score shows whether a p-value is typical or atypical relative to its distribution with respect to its average , x¯ , and standard deviation , σ ( x ) . To summarise , the p-values are evaluating the probability of each motif to be enriched in the positives given its presence/absence in the negatives/background , while the z-scores are scoring the enrichment of each motif with respect to all others present in the positives . However , in cases where either the negatives set or the background set is not available , we calculate for each motif one p-value , pNEG or pBAK , and one z-score , zNEG and zBAK , depending on which set is available . These cases have been considered in the development of DALEL webserver , http://michnick . bcm . umontreal . ca/dalel/Server , to let the user decide which reference set to use , the negatives and/or the background . Binding energy was obtained using position-weight scoring matrices developed by Fernandez-Ballester et al . [67] , and available in the ADAN database [68] . For each SH3 domain in S . cerevisiae , the ADAN database provides position-weight matrices predicting the contribution of each amino acid in terms of binding and stability energy between an SH3 domain and a target motif . Protein solvent accessibility was obtained by using SABLE version 2 [69] , a program used for predicting relative solvent accessibilities of amino acid residues in proteins . In our experiments , only residues with highest confidence level ( CI = 9 ) of solvent accessibility were considered in the analysis . For an input protein sequence , highly homologous sequences are collected from a proteome reference ( i . e . here fungi proteome ) using PSI-BLAST [70] with 35% minimum homology . After that , highly similar sequences among collected homologous are filtered using CD-HIT with 95% maximum homology . After which , remaining homologous sequences including the input protein sequence are aligned using MUSCLE algorithm [71] . Finally , Rate4Site program [72] is applied on the multiple sequence alignment to compute position-specific conservation scores of the input protein sequence across diverse species . Protein disorder was determined using DISOPRED version 2 [73] . Only residues with the highest confidence level of disorder ( CI = 9 ) were considered as disordered . To carry out the benchmark , we planted curated motifs from the ELM database [16] into protein sequences from S . cerevisiae , and proceeded with their blind discovery using DALEL , SLiMFinder and MotifHound . We utilized the proteome of S . cerevisiae as background in our benchmark . We filtered out homologous sequences by choosing 1 000 sequences of 100 to 500 residues length that showed less than 50% pairwise identity over alignment of at least 50 residues . Among these 1 , 000 sequences , the average length is 270±117 amino acids and the total number of amino acids is 269 , 941 . We then randomly created 20 sets of 50 sequences within which motifs were planted . We selected 8 motifs from the ELM database , containing between 4 and 11 residues ( S4 to S11 ) and 1 to 6 ambiguous positions ( A1 to A6 , cf . Table 3 ) . First , for each ELM motif ( i . e . referred as “regular expression” , Table 3 , Fig 9A ) , we generated the exhaustive list of amino acid combinations at all ambiguous positions , totalling N combinations: In the formulas above , aai corresponds to the number of amino acids at the ith ambiguous position , while A corresponds to the number of ambiguous positions ( Fig 9B ) . Thus , each combination samples a unique arrangement of the amino acids inside the square brackets of the corresponding regular expression . Second , we sampled n degenerate combinations of each motif and planted them randomly either 5 , 10 , 15 , or 20 times in a set of 50 sequences , with at most one motif planted per sequence ( Fig 9C ) . When a motif is planted , wildcard positions take the identity of amino acids already present in the sequence . Each motif was planted into 20 independent sets of 50 sequences ( Fig 9D ) , and this process was carried out for different numbers of occurrences . Altogether , the complete benchmark dataset was composed of 640 sets of 50 sequences ( 8 ELM motifs x 20 sets of sequences x 4 planted occurrences ) . To evaluate the performance of each method , we considered the positions of the sequence dataset matched by the top-ranked ( most significant ) motifs found by each algorithm ( Fig 8A ) . Therefore , for each amino acid covered by a top-ranked motif , we assigned one of four possible prediction outcome ( TP True Positive , FP False Positive , TN True Negative , FN False Negative ) depending on whether the positions were matching the positions at which we planted the motifs: TP ( number of positions correctly detected ) , FP ( number of positions incorrectly detected ) , TN ( number of incorrect positions not detected ) , FN ( number of correct positions not detected ) . Note that here , the terms “Positive” and “Negative” used here are not related to the same terms utilized earlier in the manuscript to define binding and non-binding proteins . Top-ranked motifs were considered and matched until their sequence coverage ( number of TP+FP ) reached the number of positions to be discovered ( TP + FN ) . We then evaluated the precision of the discovery as: TP / ( TP+FP ) ) , which we required to be above 0 . 7 . Thus , we considered a motif to be successfully identified when there was an overlap of at least 70% between the positions covered by the top-ranked motifs with respect to the positions of the planted motif . We finally calculated a global “discovery accuracy” ( Fig 8B ) per motif and for each number of occurrence , by the fraction of sets in which the planted motif was successfully identified , i . e . Number of identifications divided by 20 .
|
Here we describe the first strategy for the exhaustive search of the linear information encoding protein-peptide recognition; an approach that has previously been physically unfeasible because the combinatorial space of polypeptide sequences is too vast . The search covers the entire space of sequences with no restriction on motif length or composition , and includes all possible combinations of amino acids at distinct positions of each sequence , as well as positions with correlated preferences for amino acids .
|
[
"Abstract",
"Introduction",
"Results",
"Materials",
"and",
"methods"
] |
[
"protein",
"interactions",
"chemical",
"compounds",
"organic",
"compounds",
"fungi",
"model",
"organisms",
"amino",
"acid",
"substitution",
"experimental",
"organism",
"systems",
"amino",
"acids",
"sequence",
"motif",
"analysis",
"research",
"and",
"analysis",
"methods",
"saccharomyces",
"sequence",
"analysis",
"bioinformatics",
"proteins",
"biological",
"databases",
"chemistry",
"yeast",
"biochemistry",
"sequence",
"databases",
"peptides",
"organic",
"chemistry",
"proteomes",
"database",
"and",
"informatics",
"methods",
"protein",
"domains",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"saccharomyces",
"cerevisiae",
"physical",
"sciences",
"organisms"
] |
2017
|
Exhaustive search of linear information encoding protein-peptide recognition
|
Glia are of vital importance for all complex nervous system . One of the many functions of glia is to insulate and provide trophic and metabolic support to axons . Here , using glial-specific RNAi knockdown in Drosophila , we silenced 6930 conserved genes in adult flies to identify essential genes and pathways . Among our screening hits , metabolic processes were highly represented , and genes involved in carbohydrate and lipid metabolic pathways appeared to be essential in glia . One critical pathway identified was de novo ceramide synthesis . Glial knockdown of lace , a subunit of the serine palmitoyltransferase associated with hereditary sensory and autonomic neuropathies in humans , resulted in ensheathment defects of peripheral nerves in Drosophila . A genetic dissection study combined with shotgun high-resolution mass spectrometry of lipids showed that levels of ceramide phosphoethanolamine are crucial for axonal ensheathment by glia . A detailed morphological and functional analysis demonstrated that the depletion of ceramide phosphoethanolamine resulted in axonal defasciculation , slowed spike propagation , and failure of wrapping glia to enwrap peripheral axons . Supplementing sphingosine into the diet rescued the neuropathy in flies . Thus , our RNAi study in Drosophila identifies a key role of ceramide phosphoethanolamine in wrapping of axons by glia .
Many of the essential functions of glia such as neurotransmitter metabolism , ion buffering , axon pathfinding , electrical insulation , and trophic support are conserved between vertebrates and invertebrates [1] , [2] . Among the various tasks that glia perform , the ensheathment of axons is one function that is of high clinical relevance as there are several human diseases in which the ensheathing membrane is broken down . In vertebrates , the electrical insulation is executed by specific glial subtypes , oligodendrocyte and Schwann cells , which enwrap axons with myelin , a multilayered compacted lipid-rich membrane stack [3] , [4] . Even though Drosophila glia do not generate myelin , specialized glia that ensheath individual axons or fascicles are present in flies [1] , [5] . The vital importance of glia in different organisms was clearly illustrated in cell ablation experiments . For example , when oligodendrocytes are ablated in mice , the animals become severely paralyzed and die prematurely [6] , [7] . Also in Drosophila , where glia constitute a minor cell population [8] , genetic ablation of glia induces rapid death of the flies [9] . Whereas the evolutionally conserved significance of glia is undisputed , little is known about the vital functions of glia . Here , we aimed at identifying genes in glia that are indispensable for the function of the nervous system .
In order to identify genes with essential function in glia , we performed a global in vivo glial-specific RNAi screening . As we were interested in evolutionary conserved glial functions , we restricted our analysis to all fly genes of which a human ortholog could be identified ( as provided by the VDRC ) comprising roughly 45% of all protein coding genes in the fly . Therefore , a library of 7881 RNAi lines corresponding to 6930 genes with a putative human ortholog was obtained for the screening [10] , [11] . The scheme of the screening is presented in ( Figure 1A ) . The expression of shRNA was restricted to glial cells in adult flies by using the pan-glial driver line repo-GAL4 [12] , [13] in combination with temperature sensitive ( ts ) GAL80ts under the control of the ubiquitous tubulin promoter ( tub-GAL80ts ) [14] . Crossing of virgin females ( tub-GAL80ts; repo-GAL4 ) with 2–3 males from UAS-shRNA fly lines were set at 18°C . three to five days post eclosion; male adult flies from F1 generation were shifted to 29°C to induce shRNA expression . After 10 days , RNAi lines showing lethality or climbing deficits ( motor defect ) in more than 50% of the flies were counted as primary hits . To evaluate the efficiency of the screening system , we performed a pre-screen with selected RNAi lines targeting genes essential for the viability of any cell type . For example , when UAS-nejire RNAi transgenic flies were crossed with tub-GAL80ts; repo-GAL4 virgin female flies , a drastic reduction in the lifespan was observed ( Figure 1B ) . Primary screening data indicated that 11% of the total 7881 RNAi lines used in the screening , showed lethality and 0 . 45% were scored for motor defects ( Figure 1C ) . Since the aim of our primary screen was to identify genes that result in lethality or severe motor defects when knocked down in glia , but not in other cells , we compared our hits with datasets that were obtained by using the same RNAi library but in combination with GMR-GAL4 ( Figure S1 ) . With this approach , we identified 736 candidates with possibly essential functions in glia ( Table S1 ) . Next , by using the list of primary hits and their predicted human ortholog , we performed a systematic networking analysis by Bingo online resource [15] to reveal gene ontology ( GO ) annotated biological processes . Among the 736 primary hits , we detected 306 genes with a possible function in metabolic processes and among those 79 genes with a predicted function in glial carbohydrate and lipid metabolic processes . Major carbohydrate pathways identified are glycolysis , pentose phosphate and polysaccharide metabolic pathways , while phospholipid , fatty acid and steroid metabolic process were amongst the lipid metabolic pathways . Since metabolic functions are of particular interest in mammalian glia [16] , [17] , we decided to analyze these hits further in a secondary screen . For the secondary screening , wandering third instar ( L3 ) larval peripheral nervous system ( PNS ) was chosen because its organization is less complex than that of the central nervous system ( CNS ) . It is easily accessible and renders possible visualization by light and electron microscopy [18] , [19] . Moreover , at this stage , the glial migration is complete and the terminally differentiated glia ensheath the afferent and efferent axons of the peripheral nerves [20]–[22] . Axons in each of the peripheral nerve are enwrapped by the wrapping glia , which in turn are encircled by two types of surface glia , the perineurial glia and subperineurial glia [18] , [19] . The rational of the secondary screening was to identify the essential metabolic pathway for glial ensheathment of axons . To visualize the glial membrane , a membrane tagged GFP ( UAS-mCD8-GFP ) [23] was expressed using repo-GAL4 and immunolabeling against HRP was performed to highlight the axons . For the secondary screening , the candidates from lipid and carbohydrate pathways were selected based on STRING protein association database [24] ( Figure S2 ) . Secondary screening resulted in various phenotypes such as glial swelling , axonal wrapping defect and axonal splitting in the larval PNS ( Figure S3A–C , Table S2 ) but no embryonic lethality was observed . One of the most penetrant alterations of axon-glia morphology was observed after glia-specific knockdown of lace , a subunit of the serine palmitoyltransferase . Serine palmitoyltransferase catalyzes the condensation of serine and palmitoyl-CoA to generate 3-ketosphinganine , the rate-limiting step in de novo sphingolipid synthesis [25] . Mutations in the two human subunits of the serine palmitoyltransferase are associated with hereditary sensory and autonomic neuropathy [26] . A common feature of the mutations is the loss of canonical enzyme activity and the generation of toxic lipid intermediates [27] . Glial inhibition of lace function ( repo>mCD8-GFP/lace RNAi ) resulted in glial bulging and in an alteration of the axonal packing in all eight pairs of abdominal nerves in all larval PNS examined ( n = 15 ) ( Figure 1D ) . The bulging of glia was localized to focal regions , but appeared randomly along the entire peripheral nerves with diameters ranging from 10 µm to 30 µm . Nerves of repo-GAL4/+ control flies were straight and packed in bundles with a uniform diameter of 5–8 µm ( Figure 1E ) . In contrast , knockdown of lace in neurons did not result in any visible alterations of axonal morphology ( Figure 1F ) . The average cross-section area of the nerve were similar in the knockdown ( elav-GAL4/lace RNAi ) and in the elav-GAL4/+ control flies ( Figure 1G ) Notably , the ensheathment defect was not due to a compromised blood-nerve-barrier ( Figure S4 ) as has for example been observed in null fray mutants [28] . In addition , the number of glial cells in the peripheral nerves was comparable to control ( Figure 2A , B ) . It is important to note glial cell death affects neuronal survival and results in embryonic lethality [9] . The absence of embryonic lethality and the comparable glial cell number suggested that glial cell death did not occur at the larval stage after knockdown of lace . The expression of lace in glia was confirmed by double immunolabeling of lace5 [29] ( LacZ enhancer trap line ) with anti-β-galactosidase and anti-repo in L3 larval peripheral nerves ( Figure 2C ) . In addition , by RT-PCR analysis of the fly brain and PNS we identified lace transcript in the nervous system of both male and female flies ( Figure 2D ) . Two independent RNAi lines ( Transformant ID 21803 and 110181 , VDRC ) against lace showed identical swelling and wrapping defects . In addition , we also observed in hypomorphic lace mutant ( lace2/lace5 ) axonal defasciculation ( Figure 2E ) , ruling out off-target effects of the RNAi lines . As in the lace-RNAi knockdown , the average cross-section area of the nerves was increased in the hypomorphic lace mutant animals . The mutant phenotypes appeared to be subtle , which is not surprising as complete loss of lace during the development is lethal , while this hypomorphic combination are viable even into adulthood [29] . However , 100% penetrance of the phenotype was observed both for the RNAi knockdown ( n = 16 ) and the hypomorphic mutants ( n = 16 ) . Importantly , the lace mutant phenotype was rescued by expressing UAS-lace specifically in the glial cells ( repo-GAL4 ) , pointing to an essential function of lace in glia ( Figure 2E , F ) . Next , we analyzed , in which of the different glial subtype lace was required . Glia subtype specific GAL4 drivers were used to silence lace function . A phenotype was only observed when lace was depleted in wrapping glia ( Nrv2-GAL4 ) ( Figure 3A ) . Quantification of GFP signal intensity revealed that membrane area was significantly reduced as compared to control ( Nrv2>mCD8GFP/lace RNAi versus Nrv2>mCD8GFP ) ( Figure 3C ) . Similar results were observed when the Nrv2>mCD8-mcherry driver line was used to knockdown lace in the wrapping glia ( Figure S5 A , E ) . In contrast , knockdown of lace in the two other glial subtypes , the subperineurial ( gliotactin-GAL4 ) and the perineurial ( NP6293-GAL4 ) , did not lead to any visible changes ( glial swellings or decrease in the GFP signal intensity ) in glia or in axons ( –G ) , suggesting a predominant role of lace in the encapsulation of peripheral nerves . To examine the ultrastructure in more detail we performed electron microscopy . Electron micrographs clearly showed that knockdown of lace in glia ( repo/lace RNAi ) severely impaired axonal enwrapping compared to control ( repo/+ ) ( Figure 3B ) . Notably , also in the non-swollen regions ( A2–A3 segment ) of the nerve much less glial processes covered the axons . Quantification demonstrated a significant increase in the number of completely unwrapped axons in this region . We observed a similar phenotype , when lace was knocked down in wrapping glia ( Nrv2/lace RNAi ) . Again , a clear increase in the completely unwrapped axons was detected as compared to the controls ( Figure 3D ) . Importantly , TUNEL assay could not detect any apoptotic glial nuclei ( Nrv2>laceRNAi ) suggesting that loss of axonal ensheathment is not because of dying wrapping glial cells ( Figure 3E ) . Together , these results indicate that sphingolipids or intermediates of the sphingolipid pathway are necessary for membrane expansion of wrapping glia . In order to search for the specific sphingolipid ( SL ) species required by the wrapping glia to mediate axonal ensheathment , a genetic dissection study was performed by expressing RNAi against all known SL metabolic enzymes selectively in glia [30] . Out of 12 genes , we found that knockdown of Spt-I [31] , schlank [32] , Des1 [33] and Pect [34] in glia ( repo>mCD8-GFP/RNAi ) with two different RNAi lines ( except Des1 due to unavailability ) phenocopied the glial swelling and axonal defasciculation as observed upon loss of lace function ( Figure 4A , S6 ) . Interestingly , all four genes that show 100% penetrance ( Table S3 ) are known to be involved in the biosynthesis of ceramide-phosphoethanolamine ( CerPE ) ( Figure 4B ) . The specificity of the effect was demonstrated by the absence of any visible phenotype after neuronal specific knockdown of Spt-I , schlank , Des1 and Pect . ( Figure 4C , D ) . Additionally , glial specific knockdown of different ceramide derivative synthesizing enzymes ( GlcT1 , CGT , CerK ) [35] , [36] and PE synthesizing enzyme ( bbc ) [34] did not show any visible defects of axon or glial morphology ( Figure S7 ) . Moreover , when Spt-I , schlank , Des1 or Pect were knocked down specifically in wrapping glia , wrapping defects similar to the lace phenotype were observed ( Figure 5A ) . The quantification revealed that the GFP signal intensity was significantly reduced in all four experiments as observed after lace knockdown ( Figure 5C , S8 ) . Ultrastructural analysis by transmission electron microscopy also showed that wrapping glia failed to extend their membrane around the axons ( Figure 5B ) ; and consequently there was an increase of the completely unwrapped axons ( Figure 5D ) . Hence , our data strongly suggests an essential function of glial CerPE in axonal ensheathment by wrapping glia . In order to analyze whether knockdown of lace , schlank , Des1 and Pect resulted in depletion of CerPE levels , we performed a detailed lipidomics analysis of the nervous system . This is particularly important in RNAi studies targeting enzymes , because residual enzyme activity due to inefficient RNAi-silencing is often sufficient for their function . Since the nervous system of Drosophila only contains 10% of glia , we expressed the RNAi both in neurons and glia using repo-GAL4 and elav-GAL4 drivers to deplete the enzymes in the entire nervous system . L3 larval brain and peripheral nerves were dissected and lipidomics analysis was performed with high-resolution shotgun mass spectrometry [37] . Importantly , our lipidomics analysis confirmed that knockdown of lace , schlank , Des1 and Pect reduced CerPE levels significantly , whereas triacylglycerol ( TAG ) and diacylglycerol ( DAG ) and sterol levels were unaltered . Ceramide levels were reduced upon downregulation of lace and Des1 , whereas knockdown of Pect lead to increased ceramide levels consistent with its function as a phosphoethanolamine cytidylyltransferase . Phosphatidylcholines ( PC ) and Phosphatidylethanolamine ( PE ) levels were slightly changed possibly due to compensatory mechanisms ( Figure 6A–D ) . To test the functional consequences of lace downregulation , we performed paired electrode recordings from abdominal nerves ( Figure 7A ) and determined the spike propagation velocities of afferent and efferent units ( Figure 7B ) . We found that afferent spike propagation velocities are mildly decreased in repo>mCD8-GFP/lace RNAi mutants compared to repo>mCD8-GFP/+ controls ( median velocity smaller by 10 . 4% ) , whereas efferent spike propagation velocities remain unchanged ( Figure 7C ) . The apparent reduction of afferent spike propagation velocities was confirmed by bootstrapping ( Figure S9 ) , which revealed that the medians of the velocity distributions obtained for lace RNAi flies and controls are significantly distinct ( p<0 . 05 ( two-tailed ) , for efferent units p>0 . 4 ) . Next , we tested whether it was possible to rescue the morphological phenotype induced by knockdown of lace in glia by supplementing sphingosine ( re-converted to ceramide by condensation with a fatty-acylCoA catalyzed by the various ceramide synthases ) into the diet of the flies . Indeed , the phenotype of glia-specific knockdown of lace was efficiently rescued by the exogenous addition of sphingosine ( 300 µM ) to the food ( Figure 8 ) . Double Immunolabelling of glia and neuronal membrane reveals that the glial bulging and axonal unpacking was rescued upon addition of sphingosine to the diet ( Figure 8A ) . Orthogonal projections ( Figure 8B ) and the quantification demonstrated the rescue of the neuropathy like phenotype in flies ( Figure 8C ) . We , furthermore , observed with the ultrastructural analysis that the glial enwrapment defect was recovered upon sphingosine addition to the diet ( Figure 8D–E ) . Quantitative analysis of the peripheral nerves using the confocal and electron microscopy showed that the oral administration of sphingosine can restore the enwrapping defect and the neuropathy-like phenotype ( Figure 8C , E ) . Sphingolipids have both structural and signalling functions in cells . CerPE is a relatively low abundant lipid constituting only around 1% of the total fly lipidome . Interestingly , CerPE appears to be enriched in the fly brain membrane lipidome ( 4% ) ( Figure 4A ) [37] , [38] . In mammals , CerPE is only found in trace amounts , since sphingolipids are in general built on ceramide phosphatidylcholine in higher organisms . There are different possibilities of how CerPE could exert its function in glia . CerPE might be required for signal transduction pathways that control membrane synthesis in wrapping glia . Recently , a mutation in egghead , an enzyme that extends the glycosphingolipids ( GSLs ) in flies , causes the proliferation and overgrowth of subperineurial glia mediated by aberrant activation of phosphatidylinositol 3-kinase-Akt pathway [39] . CerPE may also increase the packing density of the lipids in the membrane , thereby helping to build up an efficient barrier for the electrical insulation of the axons . In vertebrates , a related sphingolipid , galactocylceramide , is critical for the formation of an insulating myelin sheath in oligodendrocytes . Galactosylceramide and/or its sulphated form are required for the tight sealing of the glial paranodal membrane to the axon [40] . Interestingly , mice lacking ceramide synthase 2 [41] , a vertebrate homolog of schlank , have myelination defects . Alterations of enzyme function or enzyme deficiencies do not only result in a reduction in the amount of an essential product , but can also lead to the accumulation of a toxic intermediate , or the production of a toxic side-product For example , mutations in human serine palmitoyltransferase result in a loss of normal enzyme function causing a shift in the substrate specificity , which increase the accumulation of atypical , toxic lipid products [27] . Thus , gain-of-toxic-function is another possibility of how knockdown of lace may cause the axonal ensheathment defects . Interestingly , supplementing sphingosine to the diet restored the ability of wrapping glia to extend their membrane around axons . How diets affect the distribution of lipids in cells and thereby modulate biological processes will be an important question for future investigations . Drosophila is an ideal system to pursue such studies because of the short life span and the powerful genetics , which enable rapid and detailed analysis . In summary , our current study illustrates that a large-scale screen in Drosophila , in combination with concomitant morphological and electrophysiological analysis has the potential to dissect the basic mechanisms of neuron-glia communication . Detailed knowledge of neuron-glia interactions is a pre-requirement for the rational design of treatment strategies for neuropathies or other diseases in the future .
Following lines were used in the study w[*]; P{w[+mC] = nrv2-GAL4 . S}3 ( in the text referred to as Nrv2-GAL4 ) [42] , w[*]; P{rl82-GAL4}/CyO ( referred to as Gliotactin-GAL4 ) [43] , y[*] w[*]; P{w[+mW . hs] = GawB}Bsg[NP6293]/CyO , P{w[-] = UAS-lacZ . UW14}UW14 ( referred to as NP6293-GAL4 ) ( DGRC , Japan ) [44] , UAS-mCD8-GFP , UAS-stinger-GFP ( kindly provided by Christian Klämbt ) , lace5 , UAS-lace-HA ( kindly provided by Thomas Hummel ) w[1118]; P{w[+mC] = GAL4} repo/TM3 , Sb ( referred to as repo-GAL4 ) , w[*]; P{w[+mC] = tubP-GAL80[ts]}20; TM2/TM6B , Tb ( referred to as tub-GAL80ts ) , P{w[+mW . hs] = GawB}elav[c155] ( referred to as elav-GAL4 ) ( Bloomington Stock Center ) . For the screening , we generated a fly line ( w; tub-GAL80ts; repo-GAL4/TM3 , Sb ) referred to as tub-GAL80ts; repo-GAL4 by combining glial specific driver repo-GAL4 with ubiquitously expressed temperature-sensitive allele of GAL80ts . The RNAi library with predicted human orthologs was provided by VDRC based on common database ( status October 2007 ) . For the secondary screening and further morphological analysis , two different RNAi lines were obtained ( GD and KK library ) . Two lace mutant lines Adh[n7] lace[2] cn[1] vg[1]/CyO ( referred to as lace2 ) , y[1] w[67c23]; P{w[+mC] = lacW}[45]lace[k05305]/CyO ( referred to as lace5 ) ( Bloomington Stock Center ) . Flies were provided with standard cornmeal-agar-yeast food if it is not mentioned otherwise . To deplete the mature glial cells tub-GAL80ts; repo-GAL4 flies were crossed with UAS-nejire RNAi , and OregonR ( negative control ) . 3–4 days post-eclosion , adult males with the respective combination of GAL4-driver , UAS-transgene and GAL80ts were shifted to 29°C ( 10–15 flies per vial ) . Numbers of dead flies were counted daily . After 2–3 days , fresh fly food was provided during the assay . At least 50 flies per genotype were used in the assay . Log Rank Test ( Mantel-Cox ) was used for statistical significance . Drosophila L3 stage larva were dissected in 1× PBS ( 137 mM NaCl , 12 mM Phosphate , 2 . 7 mM KCl , pH 7 . 4 ) and the PNS fixed with Bouin's fixative solution ( HT10132 , Sigma-Aldrich , Germany for 3 min . Then the tissue was permeabilized with 1× PBT ( 0 . 1% Triton-X in PBS ) solution for 15 min . For blocking ( 1 hour ) and antibody dilutions 10% goat serum ( G9023 , Sigma-Aldrich , Germany ) was used . Primary antibodies GFP ( A11122 , Invitrogen , Germany ) , anti-HRP-Cy3 ( 123-165-021 , Dianova , Germany ) , antiHRP-alexa647 ( 123-605-021 , Dianova , Germany ) , anti-repo ( 8D12 , DSHB , University of Iowa , USA ) were used with 1∶1000 , 1∶200 , 1∶200 , 1∶20 dilutions , respectively . Primary antibodies were incubated overnight whereas secondary antibodies anti-rabbit alexa-488 ( A-11008 ) , mouse alexa-647 ( A-21236 , both from Invitrogen , Germany ) mouse-Cy3 ( 115-165-14 , Dianova , Germany ) were used with 1∶200 dilutions for 2 hours . After washing 3 times with PBT , larval mouth part was removed and fillets were mounted in Vectashield ( H-1000 , VectorLab , USA ) . Approximately 200 µm nerve segments were imaged from A3 or A4 body wall segment . Five nerve widths were measured approximately after every 40 µm along the length of the nerve . Every five measurements of each nerve were considered as an ordered quintuplet ( d1 , d2 , d3 , d4 , d5 ) [46] . These five values were used to estimate average cross-sectional area of the nerve with the following equation: This estimated cross-sectional area of the nerve was calculated by considering the volume of the nerve same as that of a cylinder . At least 5–7 nerves per animals were used to measure this A-value . A-values from each animal were averaged and the mean of these average values were compared between control and lace knockdown . For the analysis of wrapping glia defects , Nrv2-GAL4 line was crossed with different UAS-shRNA lines . Images of L3 larval stage PNS were taken for both control and treated groups with exactly the same settings of the confocal microscope ( Zeiss , LSM 510 ) . Quantification of the intensity was performed using ImageJ software ( NIH , USA ) . The intensity of the signal was measured as the mean grey value per square micrometer . L3 larval PNS was imaged with Zeiss confocal microscope ( LSM510 ) having 40× water-immersion objective . z-stacks images with optical section of 0 . 5 µm were taken and digital projections of the stack and optical orthogonal section was analyzed using Zeiss LSM image browser software . ImageJ was used for the image processing . Drosophila larval brain and PNS was dissected and total RNA was isolated using Macherey Nagel ( Germany ) RNA isolation kit according to the manufacturer protocol . 1 µg of RNA was used for cDNA synthesis using SuperScript III First-Strand synthesis kit ( 18080-051 , Invitrogen , Germany ) . 1 µg RNA , 1 µl of 50 µM oligo ( dT ) 20 , 1 µl of 10 mM dNTP mix and sterile water were mixed to make up volume to 10 µl . The mixture was incubated at 65°C for 5 min and then cooled down to 4°C . A Reverse Transcriptase mix ( RT mix ) was prepared by mixing 2 µl 10× RT buffer , 4 µl 25 mM MgCl2 , 2 µl 0 . 1 M DTT , 1 µl RNAseOUT and 1 µl Superscript III RT . All of the reagents were provided in kit . RT mix was added to pre-cooled RNA-mix and incubated for 50 min at 50C . The reaction was stopped by increasing the temperature to 85°C for 5 min . 1 µl RNAse H was added and incubated for 20 min at 37°C to cleave remaining RNA . The mixture was cooled to 4°C and cDNA samples were stored at −20°C . cDNA samples were used as template to do a normal semi-quantitative PCR . 2 µl cDNA ( 1∶10 dilution ) , 0 . 3 µl of each primer , 2 . 5 µl of 25 mM MgCl2 , 1 µl 10 mM dNTP , 10 µl 5× GoTaq flexi reaction buffer , 0 . 2 µl of Go-Taq flexi DNA polymerase ( M8301 , Promega , Germany ) were mixed and sterile water was added to make final volume 50 µl . Primers for elav are 5′- CGCACAAACCTTATTGTCAACTAC-3′ and 5′-AATTTTACCACTATGGGGTCTGTG-3′ . Primers for lace are 5′-TTCGACGGCGATTCTGGAAC-3′ and 5′-CAGAGCAATAACCTCGGGCAAA-3′ . For all our experiments , we chose the very late third instar larva namely , wandering L3 larva that stopped eating and climbed away from the food . All larva for the analysis were collected 6 days after egg laying at 25°C . Larval fillets were fixed with a mixture of 4% paraformaldehyde and 2 . 5% glutaraldehyde in 0 . 1 M PBS for 4 hours at room temperature . The fillets were washed with PBS and then subjected to post-fixation with 1% osmium tetroxide for 1 hour at 4°C . Next , the post-fixed fillets were dehydrated and stained with a mixture of freshly prepared 1 . 5% uranyl acetate and 1 . 5% tungstophosphoric acid . After completion of dehydration process , the fillets were embedded in Epon . Then the silver sections ( from A2–A3 regions ) were cut and contrasted with 4% uranyl acetate followed by 0 . 3% lead citrate . Multiple sections were cut , contrasted and imaged for every genotype . The sections were imaged with a LEO EM912 Omega electron microscope ( Carl Zeiss , Germany ) and the digital micrographs were obtained with an on-axis 2048X2048 CCD camera ( Proscan GmbH , Germany ) . Paired suction electrode recordings of spontaneous spiking activity ( 20°–23°C , HL3 saline [45] ) obtained at the same abdominal nerve in fillet dissection were band-pass filtered ( 100–3000 Hz ) and simultaneously sampled at a rate of 20 kHz . Waveform templates generated with Spike II ( Cambridge Electronics , UK ) were used as triggers for averaging both electrode signals . Electrode tips were placed on approximately the same z-level and photographed to assess their tip-to-tip distance . Action potential speed distributions were compared by bootstrapping using 10000 repetitions to reveal the 95% confidence intervals of the medians in the two experimental groups . L3 larval brain and peripheral nerves were dissected and then processed for lipid isolation and mass-spectrometry analysis as described before [37] . Briefly , for each replicate , 5 brains were homogenized in 150 mM ammonium bicarbonate using a pestle attached to a cordless motor . For each knockdown experiment , samples were collected from the crossing of three different parents . Sample volume was adjusted to 200 µl . For absolute quantification , internal standards were added to control for lipid-class dependent differences in extraction and ionization . The internal standard mix contained PG 17∶0 and TAG 36∶0 , 10 pmol; Cer 17∶0 , GalCer 12∶0 and DAG 17∶0 , 20 pmol; PA 17∶0 , 25 pmol; PS 17∶0 and PC 18∶3 , 40 pmol; PI 17∶0 , CerPE 12∶0 , Chol-D7 and PE 17∶0 , 50 pmol . Lipids were extracted using a modified Folch extraction protocol [47]: 265 µl of methanol was added to the aqueous phase and vortexed; 730 µl of chloroform was added and the samples were vortexed for 1 h; after centrifugation , the organic phase was collected and dried under vacuum to avoid lipid oxidation . The whole extraction procedure including sample preparation was performed at 4°C in order to prevent lipid degradation . All lipid standards were purchased from Avanti Polar Lipids ( Alabaster , USA ) . Solvents were purchased from Sigma-Aldrich ( Taufkirchen , Germany ) . Mass spectrometric analyses were performed on a QExactive instrument ( Thermo Fisher Scientific , Germany ) equipped with a robotic nanoflow ion source TriVersa NanoMate ( Advion BioSciences , Ithaca , USA ) using chips with the diameter of spraying nozzles of 4 . 1 µm . The ion source was controlled by Chipsoft 8 . 3 . 1 software . Ionization voltages were +1 . 25 kV and −0 . 9 kV in positive and negative modes , respectively; backpressure was set at 0 . 95 psi in both modes . The temperature of ion transfer capillary was 200°C; tube voltages were 90 V ( MS+ ) and −150 V ( MS− ) . Acquisitions were performed at the mass resolution Rm/z 200 = 140 000 . AGC control was set at 106 ions and maximum injection time was set to 50 ms . Dried total lipid extracts were re-dissolved in 100 µl of chloroform: methanol 1∶2 . For the analysis , 10 µl of samples were loaded onto 96-well plate ( Eppendorf , Germany ) of the TriVersa NanoMate ion source and sealed with aluminium foil . Each sample was analyzed for 4 min in positive ion mode where PE , PC PC-O , TAG , CerPE and DAG were detected and quantified . This was followed by an acquisition in negative ion mode for 5 min where PA , PI , PS , PG , PE , PEO- , Cer , HexCer were detected and quantified . Sterol quantification method was performed as described elsewhere [37] . Briefly , dried samples were sulfated with sulfur trioxide pyridine complex in pyridine ( Sigma-Aldrich , Germany ) , sonicated and incubated at room temperature . Then barium acetate ( Sigma-Aldrich , Germany ) was added , samples sonicated and incubated 10 min at room temperature and then 1 hour at 4°C . Sulfated sterols were quantified in MS mode on QExactive mass spectrometer using cholesterol-D7 ( Avanti Polar Lipids , USA ) as internal standard . Lipids were identified by LipidXplorer software [48] by matching the m/z of their monoisotopic peaks to the corresponding elemental composition constraints . Mass tolerance was 5 p . p . m and intensity threshold was set according to the noise level reported by Xcalibur software ( Thermo Scientific , Germany ) . In order to check the integrity of blood nerve barrier , third instar larva were injected with dextran conjugated Rhodamine dye using standard procedure [49] , [50] . 10 kD dextran conjugated Rhodamine was injected in the third instar larval abdominal cavity using a FemtoJet express microinjecting device ( Eppendorf ) . For the injections , glass micropipettes prepared with a P-97 pipette puller ( Sutter Instrument ) from glass tubes ( thin wall , 3 inches , 1-mm diameter; World Precision Instruments ) were used . Injection was monitored using a dissection microscope ( Leica MZ6 ) . The successful injections were monitored using a stereomicroscope ( Zeiss ) . After the injections , animals were kept in small Petridis with standard fly food so that they can eat and move . 20–30 min post injection live animals were placed under the microscope after a brief cold shock . Confocal images of larval ventral cord and peripheral nerves were acquired using a Zeiss 510 Confocal microscope . Terminal deoxynucleotidyl transferase- mediated biotinylated UTP nick end labeling ( TUNEL ) was performed to detect the wrapping glial apoptosis . In situ cell death detection kit ( Roche ) was used for the assay and was performed according to the manufacturer protocol .
|
Glia are essential for the function of any nervous system . The number of glia correlates with the complexity of the nervous system . Important functions of glia include maintaining ionic homeostasis , supporting neurotransmission , and insulating axons to speed up nerve conduction . The biomedical relevance of glia is highlighted by an increasing number of neurological diseases , in which glia appear to play an essential role , ranging from neuropathies to schizophrenia . Here , we performed a global in vivo glial-specific RNAi screen of evolutionary conserved genes in Drosophila . With this approach , we identified 736 candidate genes that resulted in lethality or motor deficits when knocked down specifically in glia . One essential pathway identified was ceramide phosphoethanolamine biosynthesis , which was found to be important for wrapping glia to extend their membrane around axons of the peripheral nerve . Our study illustrates that a large-scale screen in Drosophila , in combination with morphological analysis is able to dissect the basic mechanism of neuron-glia communication and identify candidate genes in human neuropathies .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
A Global In Vivo Drosophila RNAi Screen Identifies a Key Role of Ceramide Phosphoethanolamine for Glial Ensheathment of Axons
|
Comparative genomics of multiple related species is a powerful methodology for the discovery of functional genomic elements , and its power should increase with the number of species compared . Here , we use 12 Drosophila genomes to study the power of comparative genomics metrics to distinguish between protein-coding and non-coding regions . First , we study the relative power of different comparative metrics and their relationship to single-species metrics . We find that even relatively simple multi-species metrics robustly outperform advanced single-species metrics , especially for shorter exons ( ≤240 nt ) , which are common in animal genomes . Moreover , the two capture largely independent features of protein-coding genes , with different sensitivity/specificity trade-offs , such that their combinations lead to even greater discriminatory power . In addition , we study how discovery power scales with the number and phylogenetic distance of the genomes compared . We find that species at a broad range of distances are comparably effective informants for pairwise comparative gene identification , but that these are surpassed by multi-species comparisons at similar evolutionary divergence . In particular , while pairwise discovery power plateaued at larger distances and never outperformed the most advanced single-species metrics , multi-species comparisons continued to benefit even from the most distant species with no apparent saturation . Last , we find that genes in functional categories typically considered fast-evolving can nonetheless be recovered at very high rates using comparative methods . Our results have implications for comparative genomics analyses in any species , including the human .
We evaluate both well-known methods for gene identification as well as several metrics that we have developed . These metrics are briefly summarized here and in Table 1 , while we provide full implementation details in the Methods section . Most initial efforts at comparative gene identification used a single informant genome to support the annotation of a target genome [15] , [23]–[29] . We selected several metrics that capture the essential properties of coding sequence evolution that they observe: the KA/KS ratio [30] , [31] and the Codon Substitution Frequencies ( CSF ) score [5] observe biases towards synonymous and other conservative codon substitutions; the Reading Frame Conservation ( RFC ) score observes the strong bias of indels within coding regions to be multiples of three in length [4] , [32]; TBLASTX measures the genome-wide significance of protein sequence similarity [33]; finally , a baseline sequence conservation metric simply measures the percent nucleotide identity between the target and informant sequences . We also selected several metrics that use multi-species alignments: the dN/dS test observes biases towards synonymous codon substitution using a statistical test based on maximum likelihood phylogenetic algorithms [34]–[36]; the multi-species CSF and RFC scores use ad hoc strategies to efficiently combine their respective pairwise scores; lastly , a baseline multi-species sequence conservation metric measures the largest fraction of species having the same nucleotide in each column ( plurality ) , averaged across the alignment . We also included several single-sequence metrics in our benchmarks to compare them to the comparative methods . Since previous studies have benchmarked many single-sequence metrics extensively [1]–[3] , we chose only a representative set here: the Fourier transform measures the strength of the three-base periodicity in coding sequences [37]; codon bias observes the unequal usage of synonymous codons , resulting in part from how different synonymous codons affect translation efficiency [38]; interpolated context models ( ICMs ) are generative probabilistic models that observe reading frame-dependent biases in the frequencies of k-mers in coding sequences , simultaneously for several different k-mer sizes [39]; lastly , Z curve observes reading frame-dependent biases in k-mer frequencies using a discriminative approach based on Fisher linear discriminant analysis [2] . To benchmark the discriminatory power of each of these metrics , we assembled a test set consisting of 10 , 722 known protein-coding exons ( from 2 , 734 genes ) in the fruit fly Drosophila melanogaster , and 39 , 181 random intergenic regions with the same length and strand distribution ( see Methods ) . These provide an ideal setting in which to evaluate genome-wide comparative genomics methods given the high quality of the FlyBase gene annotations [5] and the recent sequencing of ten Drosophila genomes [21] , [22] , in addition to D . melanogaster [40] and D . pseudoobscura [41] . We extracted each of these regions from two different sets of whole-genome sequence alignments of the twelve fly genomes [22] , one generated by MULTIZ [42] , which uses local alignments of high-similarity regions , and the second generated by the Mercator orthology mapper ( C . Dewey and L . Pachter ) and MAVID sequence aligner [43] , based on the identification of orthologous segments in each genome by conserved gene order ( synteny ) . For each metric , we scored all the 49 , 903 regions in our test set ( 10 , 722 exons and 39 , 181 non-coding regions ) and then measured its ability to correctly classify them as coding or non-coding . We used four-fold cross-validation to train and apply the metrics that require training data . We evaluated the performance of each metric by examining receiver-operator characteristic ( ROC ) curves showing its sensitivity and specificity at different score cutoffs . ( Here and throughout this paper , we use the term specificity as it is defined in binary classification problems: the fraction of true negatives that are correctly classified as negative . This differs from the common usage of the term in the gene prediction field to refer to the fraction of the examples classified as positive that are true positives . Additionally , we use the term false positive rate to mean 1-Specificity , or the fraction of true negatives incorrectly classified as positive . ) Based on the ROC curve for each metric , we also computed two different summary error measures , to facilitate comparing the performance of different metrics and methodological choices:
We first compared the overall performance of the metrics ( Figure 1 ) . All of the metrics we evaluated demonstrated high classification performance , but some general trends were apparent . The comparative metrics ( using the MULTIZ alignments of all twelve fly genomes ) generally outperformed the single-sequence metrics ( except for the baseline sequence conservation metric ) . For example , the best comparative metric resulted in 24% lower error than the best single-sequence metric ( 0 . 050 MAE for the dN/dS test vs . 0 . 065 for Z curve ) . Different metrics were preferable at different sensitivity/specificity tradeoffs . For example , the CSF and dN/dS metrics achieved the highest specificity ( 99 . 9% for CSF ) even at fairly high sensitivities ( 85 . 2% ) . RFC tended towards higher sensitivity and lower specificity than CSF and dN/dS . We also compared the pairwise metrics , using the best pairwise informant ( D . ananassae; we investigate different pairwise informants below ) , and found similar trends ( Figure S1 ) . For example , CSF and KA/KS performed comparably , showing the highest specificity , while RFC tended towards higher sensitivity and lower specificity . TBLASTX performed substantially worse than KA/KS , CSF , and RFC , but it was still better than our baseline conservation metric . Notably , none of the pairwise comparative metrics outperformed the best single-sequence metric ( Z curve ) according to MAE and AAC error , and they exhibited generally lower sensitivity . CSF and KA/KS were , however , able to achieve higher specificity at a moderate sensitivity tradeoff . For example , at 80% sensitivity , CSF had a nearly ten-fold lower false positive rate than Z curve ( 0 . 15% and 1 . 39% ) ; the specificity of CSF exceeded Z curve at less than 85% sensitivity , compared to 93% sensitivity at Z curve's MAE point . We next assessed each metric's discriminatory power for different sequence length categories ( Figure 1C ) . All of the metrics performed better on longer sequences than shorter sequences . Single-sequence metrics performed comparably or slightly better than comparative methods for long sequences ( >240 nt ) , but comparative methods strongly outperformed single-sequence metrics on shorter sequences . For example , in the length range of 181–240 nt ( which includes the median exon length ) the best comparative metric resulted in 51% lower error than the best single-sequence metric ( 0 . 027 MAE for the dN/dS test and 0 . 056 MAE for Z curve ) . In the shorter length range of 121–180 nt , the best comparative metric resulted in 60% lower error than the best single-sequence metric ( 0 . 029 MAE for CSF and 0 . 073 MAE for Z curve ) . Different comparative methods were also preferred at different lengths . For example , CSF strongly outperformed the dN/dS test on the shortest sequences ( ≤60 nt ) , while they performed comparably on longer sequences . While each of the metrics we studied exhibited unique performance characteristics , some measure similar fundamental lines of evidence , and thus may tend to err on the same examples . We investigated the independence of the metrics , indicated by how differently they rank the exons in our test set , using a dimensionality reduction technique called multidimensional scaling ( MDS; see Methods ) . This analysis led to a two-dimensional visualization shown in Figure 2A , in which each point represents one of the metrics and the distance between the points approximately represents their dissimilarity . We found that the dN/dS test and CSF behaved very similarly , while RFC was clearly distinct . The sequence conservation metric was separate from each of these , while TBLASTX clustered with CSF and dN/dS . The four single-sequence metrics formed two additional clusters distinct from the comparative metrics . These findings agree with intuition: CSF and the dN/dS test both observe the distinctive biases in codon substitutions in protein-coding sequences , while RFC observes patterns of insertions and deletions that are essentially orthogonal to codon substitutions , and the single-sequence metrics observe compositional biases and periodicities that are ignored by the comparative metrics . The relative independence of several of the metrics suggests that combining them could lead to higher performance . We selected five metrics representing each of the MDS clusters ( CSF , RFC , sequence conservation , Z curve , and codon bias ) and combined them using cross-validated linear discriminant analysis ( LDA ) . As expected , the hybrid metric outperformed any of its inputs: by MAE error , the LDA hybrid resulted in 27% lower error than its best input metric ( 0 . 040 MAE for LDA vs . 0 . 055 for CSF ) . The hybrid metric demonstrated much higher sensitivity than any of its input metrics ( Figure 2B ) , and higher specificity than all of the input metrics except CSF . We obtained almost identical results using a second hybrid metric based on a linear support vector machine instead of LDA . Thus , although CSF and the dN/dS test remain the methods of choice for the highest specificity , the hybrid metrics achieved higher overall performance . We next investigated how strongly the performance of the comparative methods depends on genome sequence alignments . We compared the above results , based on MULTIZ local similarity-based alignments , with the corresponding results based on the synteny-anchored Mercator/MAVID alignments . Overall , the two alignments led to highly concordant results , with similar trends in the performance of the metrics relative to each other and across different sequence lengths . There were , however , some notable differences in their absolute levels of performance . We expected the local alignment approach to give higher sensitivity than the synteny-anchored alignments , since it should be better able to align exons that have undergone rearrangements [45] . Indeed , we found that MULTIZ tended to align more species for each region ( Figure S2 ) and led to higher sensitivity than the Mercator/MAVID alignments ( e . g . 90% vs . 87% for CSF at 99% specificity , with 85% of exons detected in both alignments; Figure S3 ) . Conversely , we expected the synteny-anchoring approach used by Mercator/MAVID to give higher specificity than the local alignment approach of MULTIZ , since it may generate fewer spurious non-orthologous alignments [45] . However , we found that while the Mercator/MAVID alignment could lead to slightly higher specificity , it did so only at disproportionate sensitivity tradeoffs . For example , with the baseline sequence conservation metric , specificity using the Mercator/MAVID alignments exceeded that of the MULTIZ alignments only at lower than 58% sensitivity ( compared to 80% sensitivity at the MULTIZ-based MAE point ) . Similarly , with RFC , specificity resulting from the Mercator/MAVID alignments was greater only at lower than 63% sensitivity ( compared to 92% MAE sensitivity ) . Overall , the Mercator/MAVID alignments led to somewhat lower sensitivity without a clear specificity advantage , and this was reflected in worse MAE and AAC error statistics ( Figure S3 ) . We therefore focused on the MULTIZ alignments for the remainder of our analysis . We note , however , that the Mercator/MAVID alignments did allow detection of some exons not detected in the MULTIZ alignments ( ∼2% of all exons ) . More generally , these empirical observations could be highly dependent on parameter settings of the genome alignment programs , and further investigation of these strategies is required . To investigate which species are the most and least effective informants for gene identification , we evaluated each pairwise comparative metric using informant genomes at increasing evolutionary distance from D . melanogaster . We applied each metric to pairwise alignments of D . melanogaster with D . erecta , D . ananassae , D . pseudoobscura , D . willistoni , and D . grimshawi , each representing various clades within the genus Drosophila ( Figure 3 ) . We found that D . ananassae was overall the most effective informant , outperforming other species on most metrics . However , inspection of the corresponding ROC curves often revealed a more complex situation , with multiple species showing similar performance , and sometimes higher for certain sensitivity/specificity tradeoffs . For example , with KA/KS , D . ananassae and D . willistoni performed comparably , with D . ananassae leading to slightly higher sensitivity and D . willistoni leading to slightly higher specificity ( Figure 4A ) . Similarly , with RFC , closely related species led to slightly higher sensitivities , and more distant species led to slightly higher specificities ( Figure S4 ) . Hence , while D . ananassae was overall the most effective informant , it did not robustly outperform the other pairwise informants we studied . The only exception was D . erecta , the most closely related to D . melanogaster of the species we studied . D . erecta was consistently less informative than the others , leading to the lowest overall classification performance on most of the pairwise metrics . To investigate more distant species for which we lacked whole-genome alignments , we also applied TBLASTX to the genomes of the mosquito [46] and honeybee [47] . We found that these species led to much worse performance than the Drosophila species as informants for D . melanogaster ( Figure 4B ) . We conclude that a broad range of species within the genus Drosophila ( outside of the melanogaster subgroup ) make effective pairwise informants for gene identification in D . melanogaster , while the mosquito and honeybee , the next most closely related species with fully sequenced genomes , are likely to be too distant for this application . These findings are consistent with a previous smaller-scale study of comparative gene identification power in flies [14] , and previous theoretical and simulation studies suggesting that , while some mathematically optimal distance may exist , species at a broad range of phylogenetic distances should be comparably effective informants for identifying exons and other conserved elements [13] , [15] . We next investigated the effectiveness of increasing numbers of informant species on the metrics that can use multiple informants . We evaluated each metric using subsets of the available species corresponding to increasingly broad clades within the genus Drosophila ( see phylogeny in Figure 3 ) : the melanogaster subgroup ( 5 species including D . melanogaster ) , the melanogaster group ( 6 species ) , the melanogaster and obscura groups ( 8 species ) , the subgenus Sophophora ( 9 species ) , and finally all 12 species of the genus Drosophila . We found that for each of the metrics we benchmarked in this way , discriminatory power tended to increase as additional informant species were used ( Figure 5A ) . In contrast to our previous pairwise analysis , in which the most distant Drosophila informants led to similar or slightly worse performance than closer species , adding informants at increasing distances led to a clear trend in higher classification performance . The dN/dS test , RFC , and the sequence conservation metric each showed a smooth progression of increasing performance with each successively larger group of informant species . For example , starting from the four informants within the melanogaster subgroup , the dN/dS test achieved an MAE of 0 . 103 . With the addition of each successive group of informants , the MAE was reduced relatively by 35% , 43% , 48% , and finally by 52% . CSF showed a similar trend through the subgenus Sophophora , but did not clearly benefit from the subsequent addition of the final three informants of subgenus Drosophila . In all cases , the improvement with multiple species was most pronounced for short exons ( Figure 5B ) . With a sufficient number of informants , the multi-species metrics surpassed single-sequence metrics according to MAE ( Figure 5C ) . This also stands in contrast to our pairwise analysis , in which no informant enabled any comparative metric to outperform the best single-sequence metric ( Z curve ) . CSF exceeded the performance of Z curve once we used at least six species ( ≥1 . 3 sub/site ) , dN/dS with at least eight species ( ≥1 . 9 sub/site ) , and RFC , using its simplistic vote-tallying scheme , with all twelve species ( 4 . 1 sub/site ) . The baseline sequence conservation metric never outperformed Z curve , although its performance also increased with additional species . ( We note that while these results show that a certain number of informants is sufficient , they do not imply that they are all necessary to achieve some level of performance; removing informants that contribute very little independent branch length might not substantially reduce performance . ) In most cases , the four informants of the melanogaster subgroup together yielded worse performance than pairwise analysis with the best pairwise informant , D . ananassae . In contrast , all of the informant clades that combined D . ananassae with more distant species led to better performance than any pairwise analysis . This affirms our earlier conclusion , based on a pairwise analysis with D . erecta , that the species within the melanogaster subgroup are sub-optimal informants for the metrics we studied , presumably because they are too closely related to D . melanogaster . Indeed , the neutral distance of D . ananassae from D . melanogaster is 1 . 0 substitutions per neutral site , while the total independent branch length provided by the four melanogaster subgroup informants is only 0 . 4 sub/site . It is well-known that genes in certain categories of biological function tend to be faster-evolving [41] , [46]–[48] . We lastly investigated whether comparative metrics therefore systematically fail to distinguish such genes from non-coding regions . We obtained Gene Ontology ( GO ) annotations [49] , [50] for each of the 2 , 734 genes comprising our test set . For each of the 192 GO terms represented by at least thirty genes in our test set , we determined the fraction of those genes with at least one exon scoring above a stringent cutoff ( “detected genes” ) . We found that all of the functional categories we investigated had very high detection rates ( Table S1 ) . For example , with a CSF cutoff corresponding to 85% exon sensitivity and 99 . 9% specificity using all twelve fly genomes , the overall fraction of detected genes was 92% , and the detection rates surpassed 90% for all but two functional categories: serine-type endopeptidase activity ( 89% detected genes ) and its superset , serine-type peptidase activity ( 86% ) . Serine proteases play key roles in insect innate immunity , and some likely evolve under positive selection [46] , [51] , [52] . Several other categories that intuition suggests might relate to more rapidly evolving genes , however , were not problematic , including immune response ( 94% ) , gametogenesis ( 95% ) and G-protein coupled receptor activity ( 100% ) . Instead , comparative metrics had the most difficulty detecting genes of unknown function . Three GO terms indicating unknown function ( unknown cellular component , molecular function , and biological process ) had only 67% , 61% , and 60% detected genes . In fact , of the genes that were not detected at this cutoff , 85% were of unknown function or lacked any GO term , compared to 49% of all the genes in our dataset . These trends held for all of the comparative metrics and cutoffs we investigated ( Table S1 ) . Overall , these results indicate that comparative methods using the twelve fly genomes were able to detect the vast majority of genes in all of the functional categories we investigated ( which were represented by at least 30 genes in our dataset; a larger sample might reveal more specific functional categories that are , in fact , very difficult for comparative methods to detect ) . They had much greater difficulty detecting genes of unknown function , which may be under less selective constraint overall [14] , [21] but could also include a higher proportion of incorrect or spurious annotations [5] . Interestingly , Z curve , a single-sequence metric , also showed much lower sensitivity to genes of unknown function ( Table S1 ) , suggesting that these genes , if they are correctly annotated , tend to be unusual in several ways .
Using a variety of different methods , we found that species ranging from 1 . 0–1 . 4 substitutions per neutral site from D . melanogaster are comparably effective informants for pairwise gene identification , with slight preference given to the closer end of this range . This “optimal” range might extend both towards closer species ( between D . erecta and D . ananassae ) and towards more distant species ( between D . grimshawi and A . gambiae ) , but these distances were not explored in the currently sequenced genomes . This range is comparable to the distance from human of the opossum ( 0 . 8 sub/site ) , chicken ( 1 . 1 sub/site ) , and lizard ( 1 . 3 sub/site ) , suggesting that species more distant than the eutherian mammals ( the farthest of which are less than 0 . 5 sub/site; Figure 3 ) may prove to be excellent informants for human gene identification . Moreover , our study showed that comparative genomics power did not saturate with the number of species compared , as the multi-species metrics tended to show continued improvement from each progressively larger group of informants studied ( Figure 5 ) . The overall improvement did become more incremental as the number of informants grew , which could be interpreted either as diminishing returns from additional genomes , or simply as the expected asymptotic increase in performance towards an achievable optimum . Importantly , the improvement from more informants was far more pronounced among short exons than long exons ( Figure 5B ) ; this suggests that , while long exons are easy to discover even with few species , still more informants may significantly improve the discovery of short coding exons , and perhaps other classes of small elements . Thus , especially for small elements , we apparently have not yet reached a saturation point with twelve metazoan species spanning a total of 4 . 13 substitutions per neutral site . We chose to express discovery power as a function of the neutral substitution rate estimated for the species compared ( Figure 3 ) . While this rate provides a compelling measure of expected discovery power [13] , it is important to note that genetic distance between species ( whether measured by neutral substitution rate or other metrics [21] , [53] ) is far from the only consideration that should guide comparative informant selection . For example , population dynamics affect the strength of selection relative to neutral drift , and thus may skew the relationship between neutral divergence and the significance of observed conservation in some lineages [54] , [55] . Additionally , the genome size and the density and type of repetitive elements in an informant genome may affect the ability to sequence , assemble , and align it to a target genome , especially if low-coverage [18] or short-read [56] , [57] sequencing strategies are used . Accurate alignment is further complicated by variation in the rates of chromosomal rearrangement and segmental duplication and loss , which are likely to affect the proportion of the genome that can be accurately recognized as orthologous , even for species that show similar nucleotide divergence . Much more fundamentally , distant species share less in common biologically; indeed , the 12 Drosophila species were selected in part to represent the diverse ecological niches they occupy [58] and the neutral distance they span ( approximately corresponding to the distance between human and reptiles ) . Thus , while our results suggest that such distant species may nonetheless be highly informative given high-quality sequences and alignments , future empirical studies should compare them to the use of many species at closer distances , such as those represented by the eutherian mammals , for gene identification . One application of the metrics we have studied will be their integration into de novo gene structure predictors based on semi-Markov conditional random fields , which can combine multiple discriminative metrics in a manner not unlike our LDA hybrid . Our results suggest that these systems should be able to use multiple informant species and multiple metrics to identify protein-coding sequences with higher accuracy , especially on short exons . Still , it is not obvious that these trends in the metrics' performance necessarily imply higher-accuracy prediction of complete gene structures , since the latter also strongly depends on the detection of splice sites and other sequence signals [12] , [59] . Additionally , like the more advanced metrics we studied , such systems tend to be highly parameterized and thus dependent on high-quality training data , which may not be available in less well-studied species . More fundamentally , the probabilistic models used in gene predictors make simplifying assumptions about gene structures that lead to many incorrect predictions , and that cannot be relaxed just by using more powerful metrics . For example , they currently cannot predict nested and interleaved genes , which are fairly common in metazoan genomes [5] , [50] , [60]–[62] , since these structures violate Markov independence assumptions . A similar challenge is presented by alternative splice isoforms with mutually exclusive exons that do not splice to each other in-frame . The methods we have studied also have other important applications , such as assessing and refining existing annotations , and searching the genome for coding regions that are systematically missed or erroneously modeled by other methods . In particular , the effectiveness of comparative methods for detecting short coding regions may prove crucial in identifying short proteins , which are known to serve important biological roles but have probably been systematically under-represented in genome annotations [63]–[66] . They also provide a promising way to search for gene structures that violate traditional assumptions entirely , such as stop codon readthrough , translational frameshifts and polycistronic transcripts , which also might be more common in animal genomes than currently appreciated [5] .
We used “Comparative Analysis Freeze 1” assemblies of the twelve Drosophila genomes [21] available from the following web site: http://rana . lbl . gov/drosophila/assemblies . html . We used two different genome alignment sets [22] . One was derived from a synteny map generated by Mercator ( C . Dewey , http://www . biostat . wisc . edu/~cdewey/mercator/ ) and sequence alignments generated by MAVID [43] . The other genome alignments were generated by MULTIZ [42] . These alignments are available from the following web site: http://rana . lbl . gov/drosophila/wiki/index . php/Alignment . We obtained FlyBase release 4 . 3 annotations from the following web site: ftp://ftp . flybase . net/genomes/Drosophila_melanogaster/dmel_r4 . 3_20060303/gff/ . We estimated branch lengths in the phylogenetic tree for the flies ( shown in Figure 3 ) based on four-fold degenerate sites in alignments of orthologous protein-coding genes . We identified one-to-one orthologs based on FlyBase annotation release 4 . 3 for D . melanogaster and community annotations for the 11 other species [21] , yielding 12 , 861 four-fold sites . Then , to estimate branch lengths , we ran PHYML v2 . 4 . 4 [67] with an HKY model of sequence evolution , a fixed tree topology ( Figure 3A ) , and remaining parameters at default values . For comparison with vertebrates , we estimated the branch lengths for 28 vertebrates using 10 , 340 four-fold sites , based on alignments of genes with one-to-one orthologs in human , dog , and mouse [68] . We obtained the MULTIZ vertebrate alignments from the UCSC Genome Browser [69] . We randomly sampled 2 , 734 of the 13 , 733 euchromatic genes in FlyBase annotation release 4 . 3 , and then selected all 10 , 722 non-overlapping exons of all transcripts of those genes . We chose this strategy of randomly sampling genes and selecting all exons of those genes , rather than directly sampling exons , to facilitate studying how the power of each metric varies across different functional categories of genes . Although not by design , the length distribution of sequences in our test set ( median = 224 nt , mean = 404 nt , sd = 570 nt ) is very similar to the length distribution of exons in the genome ( median = 220 nt , mean = 408 nt , sd = 568 nt ) . Each known exon was evaluated in its annotated reading frame . For each known exon in our dataset , we selected four non-coding regions of the same length and strand . We selected each of these regions by randomly choosing a start coordinate in the BDGP Release 4 assembly of the D . melanogaster euchromatic chromosome arms , and ensuring that the resulting region did not overlap an annotated coding exon . We also chose only regions consisting of at least 50% nucleotide characters ( as opposed to Ns ) . The codon reading frame for the non-coding regions was always set arbitrarily to 0 ( that is , they were always considered to begin with a complete codon ) . We removed in-frame stop codons in D . melanogaster from the non-coding regions ( the length of each control region matched the corresponding exon after removing stop codons ) . All the regions in the dataset were selected without regard to how well they were aligned in either genome alignment set we used . The coordinates , sequences , and alignments of our dataset are available for download ( Text S1 ) . CSF and the single-sequence metrics ( except for Fourier transform ) require training to estimate parameters . To avoid overfitting , we trained and applied them using four-fold cross validation: we randomly partitioned the dataset into four subsets , and then generated scores for each subset by training on the other three subsets . We then combined the scores for the subsets to obtain scores for the entire dataset . We applied the other metrics directly to each sequence . We computed ROC curves for each metric by choosing 250 cutoffs representing quantiles of the score distribution over the entire dataset , and at each cutoff , evaluating sensitivity and specificity when sequences scoring above the cutoff are considered positively classified , and sequences scoring less than or equal to the cutoff are negatively classified . Some metrics failed to produce a score for some sequences; for example , comparative metrics produced no score for sequences in which no alignment was present . These sequences were regarded as negatively classified at all cutoffs , reflecting a non-coding default hypothesis . Our ROC curves may therefore underestimate the sensitivity or overestimate the specificity that each comparative method would exhibit if given perfect alignments of all orthologous elements . We computed the MAE as the highest average sensitivity and specificity among the 250 points on the ROC curve , and the AAC by trapezoidal integration over these points . We created hybrid metrics by combining the pre-computed scores of the input metrics using linear discriminant analysis ( LDA ) and a support vector machine ( SVM ) . In both cases , prior to combination , the scores of each input metric were normalized to have zero mean and unit variance across the entire dataset . The normalized scores from each input metric were then used as feature vectors representing each sequence in the dataset . We trained and applied the hybrid metrics using four-fold cross-validation . We applied LDA with default settings in MATLAB . For SVM , we used SVMlight 4 . 00 [73] with a linear kernel and default cost parameters . We used the prediction confidence computed by the svm_classify program as the SVM hybrid metric score for each sequence . Multidimensional scaling ( MDS ) takes a high-dimensional matrix of pairwise similarities between items ( in our case , metrics ) , and assigns each item to a point in a low-dimensional space ( in our case , two dimensions for visualization ) , such that the distance between any two points approximately represents the dissimilarity of the corresponding items . We applied MDS to generate the visualization in Figure 2A using the R function cmdscale with default parameters . We defined the similarity between two metrics as S ( i , j ) = cor ( Ri , Rj ) , where Ri is the vector of ranks of the known exons according to the scores computed by metric i . For example , if the known exons are ordered in some way E1 , E2 , E3 , and metric i assigns them scores Mi ( [E1 , E2 , E3] ) = [0 . 2 , 1 . 0 , −0 . 5] , then Ri = [3] , [1] , [2] .
|
Comparing the genomes of related species is a powerful approach to the discovery of functional elements such as protein-coding genes . Theoretically , using more species should lead to more discovery power . Many questions remain , however , surrounding the optimal choice of species to compare and how to best use multi-species alignments . It is even possible that practical limitations in the sequencing , assembly , and alignment of genomes could effectively negate the benefit of using more species . Here , we used 12 complete fly genomes to study a variety of metrics used to identify protein-coding genes , including methods that analyze only the genome of interest and comparative methods that examine evolutionary signatures in genome alignments . We found that species over a surprisingly broad range of phylogenetic distances were effective in comparative analyses , and that discovery power continued to scale with each additional species without apparent saturation . We also examined whether comparative methods systematically miss genes considered fast-evolving , and studied how performance is influenced by genome alignment strategies . Our results can help guide species selection for future comparative studies and provide methodological guidance for a variety of gene identification tasks , including the design of future de novo gene predictors and the search for unusual gene structures .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"computational",
"biology/evolutionary",
"modeling",
"computational",
"biology/comparative",
"sequence",
"analysis",
"computational",
"biology/genomics"
] |
2008
|
Performance and Scalability of Discriminative Metrics for Comparative Gene Identification in 12 Drosophila Genomes
|
Although there have been great advances in our understanding of the bacterial cytoskeleton , major gaps remain in our knowledge of its importance to virulence . In this study we have explored the contribution of the bacterial cytoskeleton to the ability of Salmonella to express and assemble virulence factors and cause disease . The bacterial actin-like protein MreB polymerises into helical filaments and interacts with other cytoskeletal elements including MreC to control cell-shape . As mreB appears to be an essential gene , we have constructed a viable ΔmreC depletion mutant in Salmonella . Using a broad range of independent biochemical , fluorescence and phenotypic screens we provide evidence that the Salmonella pathogenicity island-1 type three secretion system ( SPI1-T3SS ) and flagella systems are down-regulated in the absence of MreC . In contrast the SPI-2 T3SS appears to remain functional . The phenotypes have been further validated using a chemical genetic approach to disrupt the functionality of MreB . Although the fitness of ΔmreC is reduced in vivo , we observed that this defect does not completely abrogate the ability of Salmonella to cause disease systemically . By forcing on expression of flagella and SPI-1 T3SS in trans with the master regulators FlhDC and HilA , it is clear that the cytoskeleton is dispensable for the assembly of these structures but essential for their expression . As two-component systems are involved in sensing and adapting to environmental and cell surface signals , we have constructed and screened a panel of such mutants and identified the sensor kinase RcsC as a key phenotypic regulator in ΔmreC . Further genetic analysis revealed the importance of the Rcs two-component system in modulating the expression of these virulence factors . Collectively , these results suggest that expression of virulence genes might be directly coordinated with cytoskeletal integrity , and this regulation is mediated by the two-component system sensor kinase RcsC .
Salmonellae remain major global pathogens causing a broad spectrum of disease ranging from gastroenteritis to typhoid fever [1] , [2] . The emergence of multidrug resistant salmonellae is complicating the management of disease [3] , [4] . Hence , there is an urgent need to identify novel bacterial targets for the development of new antimicrobial agents or vaccines to combat infection . The view that bacteria do not possess a cytoskeleton has radically changed in recent years with the discovery of intracellular filamentous protein assemblies with cell-shape defining function [5] . Although there is little primary sequence identity between eukaryotic cytoskeletal proteins and those in prokaryotes , proteins with actin- and tubulin-like structural motifs have been identified in bacteria . Bacterial cytokinesis is dependent on FtsZ which contains a structural fold mirroring tubulin . FtsZ displays similar dynamic properties to tubulin and is able to polymerise unidirectionally in a GTP-dependent manner to produce polymeric filaments [6] , [7] . Polymers of FtsZ are able to assemble into a transient helical structure and subsequently form a ring-like structure around the circumference of the mid-cell [8] . This Z-ring is required for recruiting proteins for the assembly of the cell division complex [8] . The intermediate filament-like protein crescentin determines the vibroid shape of Caulobacter crescentus cells [9] . The bacterial proteins MreB , Mbl , and ParM display the structural and dynamic properties of eukaryotic actin [10] . Amongst these proteins , MreB is the most homologous to actin in terms of primary sequence , structure , and size [11] , [12] . The most conserved region of this actin-superfamily is the ATPase domain . MreB can polymerise into helical filamentous structures important for cell morphology . Live cell microscopy in Bacillus subtilis revealed that MreB forms large cables which follow a helical path close to the cytoplasmic membrane [5] . An equivalent MreB protein has been found in Escherichia coli . When MreB is depleted , rod-shaped B . subtilis and E . coli cells become spherical [5] , [13]–[15] . In C . crescentus MreB has been implicated to play a role in the control of cell polarity [16] . In rod-shaped bacteria the MreB polymeric structures control the localisation of cell wall growth by providing a scaffold for enzymes involved in cell wall assembly [17] . The MreB operon in E . coli and B . subtilis encodes for a number of additional genes , which do not possess any similarity to actin [18] . These include the cellular membrane proteins MreC and MreD , which also have a helical disposition . MreC forms a dimer and interestingly in C . crescentus MreC is localised in spirals in the periplasm [19] . Recent studies by Rothfield and colleagues provide convincing evidence to suggest that in E . coli MreB , MreC and MreD form helical structures independently of each other [20] . Using affinity purification and bacterial two hybrid assays , MreC and MreD appear to interact together [13] . In E . coli there is evidence to suggest that MreB interacts with MreC , but this may not be the case in Rhodobacter sphaeroides or C . crescentus [21] . As well as playing a key role in cell morphogenesis , MreB also has a pivotal function in chromosome segregation [22]–[24] . Adding the MreB inhibitor A22 [S- ( 3 , 4-Dichlorobenzyl ) isothiourea] to synchronised cultures of C . crescentus inhibited segregation of GFP-tagged chromosomal origins [22] . However MreB may not function in chromosome segregation in Bacillus [15] . Recently another helically distributed cytoplasmic membrane protein which interacts with MreB named RodZ has been identified [25]–[27] . Cellular components including the RNA degradosome and lipopolysaccharide have also been identified to be localised in helical structures within the cell [28] , [29] . In spite of these major advances in our understanding of the structure and organization of the bacterial cytoskeleton , there are major gaps in our knowledge of its role in bacterial pathogenicity . In this study we wished to gain insights into understanding the function of the bacterial cytoskeleton in the pathogenicity of Salmonella .
The in vivo experiments were covered by a Project License granted by the Home Office under the Animal ( Scientific Procedures ) Act 1986 . This license was approved locally by the University of Cambridge Ethical Review Committee . S . Typhimurium SL1344 and mutant derivatives used in this study are described in Table 1 . Strains were routinely grown in Luria-Bertani ( LB ) broth with appropriate antibiotics at the following concentrations: ( kanamycin 50 µg ml−1 ) , ampicillin ( 100 µg ml−1 or 30 µg ml−1 for pNDM220 ) . A22 ( Calbiochem ) was added at 10 µg ml−1 . Bacteria were grown overnight in 5 ml LB , before 25 µl of culture was used to inoculate 25 ml of fresh LB in a 250 ml flask and grown at 37°C shaking ( 200 rpm ) unless otherwise stated . ΔmreC was maintained in media containing 100 µM IPTG , however for phenotypic testing this was removed unless otherwise mentioned . For the SPI-1 T3S studies cells were grown overnight in LB before subculturing 1/100 into 25 ml fresh LB and growing at 37°C for approximately 5 hrs with good aeration until OD600nm∼1 . 2 in 250 ml flasks [30] . For the SPI-2 T3S studies cells were grown in SPI-2 induction media ( 100 mM Tris-base , 0 . 1% w/v casamino acids , 0 . 1% w/v glycerol , 10 µM MgSO4 , 40 µg ml−1 histidine , pH 5 . 8 ) . Cells were grown overnight in LB before subculturing 1/100 in 25 ml SPI-2 inducing media before growing for 16 h at 37°C in 250 ml flasks before sampling . Cells were inoculated from a fresh LB plate onto the semi-solid motility agar ( 10 g l−1 Bacto-tryptone , 5 g l−1 NaCl , 3 g l−1 agar ) and incubated upright for a minimum of 5 h . Distinct zones of cell motility were measured and compared to WT SL1344 and non-motile SL1344 strains . Chromosomal gene deletions were constructed using the lambda Red method as described previously [31] , before transducing the mutation into a genetically clean parent strain using bacteriophage P22int . In the case of ΔmreC and ΔmreD the mutations were transduced into a parent strain containing pTK521 ( lac-mreBCD E . coli ) to complement the mutation in the presence of 100 µM isopropyl beta-D-1-thiogalactopyranoside ( IPTG ) . Gene deletion primers typically encompassed the first and final 20 bases of the coding sequence of the respective gene were synthesised . However , as the mreC and mreD gene coding sequences overlap by a single base , to ensure only a single coding sequence was disrupted the respective mreC 3′ primer and mreD 5′ primer were moved internally into their coding sequence such as to produce no overlapping mutations . Gene deletions for the two-component systems ( ΔqseF , ΔphoBR , ΔyjiGH , ΔbaeSR , ΔbasSR , ΔhydH , ΔqseBC ΔtctDE , ΔcpxAR , ΔrcaA , ΔrcsB , ΔrcsC , ΔrcsD , ΔrcsDB , and ΔrcsCBD ) , were constructed in SL1344 WT using classical lambda Red methods before transducing into the ΔmreC strain using bacteriophage P22int . Primers are listed in Table 2 . GFP was amplified from pZEP08 and cloned along with a new multiple cloning site into the EcoRI and HindIII sites of pBR322 to create pBR322GFP . mreB along with its natural promoter was amplified from genomic DNA and cloned into the EcoRI and XbaI sites of pBR322GFP , before the mreB-gfp fusion was subcloned from the pBR322mreB-gfp into pNDM220 using the EcoRI and BamHI sites . Flagella and SPI1 transcriptional reporter plasmids were transformed into SL1344 and ΔmreC mutant cells . Expression from the lux transcriptional reporters was measured during the growth cycle of 10−3 diluted overnight cultures cells grown in microtitre plates ( 200 µl total volume ) for a minimum of 15 h at 37°C with periodic shaking . Optical density ( 600nm ) and relative luminescence was measured at 15 minute intervals using a Tecan Infinity200 luminometer . Samples were tested in triplicate , and repeated at least 3 times . The hilA and rcsC open reading frames were amplified from SL1344 genomic DNA and cloned into the EcoRI and XbaI or the EcoRI and HindIII sites of pBAD24 to create pBADhilA and pBADrcsC respectively . Whole cell total protein samples were obtained by pelleting an appropriate volume of bacterial culture , followed by resuspension in SDS-loading buffer and boiling for 10 mins . Culture supernatants were filter sterilized ( 0 . 22 µm ) and proteins were ammonium sulphate precipitated ( 4 g 10 ml−1 supernatant ) overnight at 4°C . Precipitated secreted proteins were resuspended in H2O and then combined with an equal volume of sample buffer ( Biorad ) . Whole cell and culture supernatant samples were run on 12% SDS/PAGE and transferred on Protran nitrocellulose transfer membranes ( Schleicher & Schuell ) using a wet transfer apparatus ( Biorad ) . Western blot analysis was performed using polyclonal SipA , SipB , SipC or PrgH for testing SPI-1 T3S functionality , coupled with a goat anti-mouse horseradish peroxidase-labelled secondary antibody ( Dako Cytomation ) . Detection was carried out using 4-chloro-1-naphthol ( Sigma ) according to the manufacturer's instructions . Female C57BL/6 mice were purchased from Harlan Olac Ltd . , ( Blackthorn , Bicester , UK ) . Mice were used when over eight weeks of age . Bacterial suspensions for injection were grown for 16 h as a stationary culture at 37oC in LB broth . Bacteria were diluted in PBS prior to injection into a lateral tail vein . Mice were killed by cervical dislocation and the livers and spleens aseptically removed . Each organ was homogenised ( separately ) in a Seward Stomacher 80 Biomaster ( Seward ) in 10 ml of distilled water and viable bacterial counts in the homogenate were assayed on pour plates of LB agar . Representative bacterial colonies were kept and re-tested for phenotypic changes . Wild type Salmonella SJW1103 cells with chromosomal N-terminal GFP fusion to fliG ( YVM004 ) [32] were P22 transduced with the mreC::kan mutation to create YVM004 ΔmreC . This strain , along with the WT control , was subsequently transduced with a chromosomally-based inducible flhDC locus derived from TH2919 [33] . Cells were grown to the appropriate growth phase ( mid-log for SPI-1 and flagella , or stationary phase for SPI-2 ) in relevant media ( LB or SPI-2 inducing media ) . Flagella visualisation strains ( fliG-gfp ) , were mounted on 1% agarose beds for imaging . Samples for visualising the type 3 secretion apparatus were fixed in 4% paraformaldehyde diluted in PBS for 1 h before washing for 15 minutes in three changes of PBS . Samples were incubated with either αSipA , αSipB , αSipC , αSipD ( SPI-1 ) or αSseB ( SPI-2 ) antibodies diluted 1∶1000 in PBS for 3 h with gentle agitation . Samples were subsequently washed in PBS before incubating in 1∶1000 Alexa Fluor 488 conjugated goat anti-rabbit antibody ( Invitrogen-Molecular Probes , Paisley , U . K . ) , washed for 30 mins in fresh PBS before mounting onto agarose beds . Half of each organ was fixed overnight in 4% paraformaldehyde diluted in PBS , washed for 90 min in three changes of PBS and then immersed in 20% sucrose ( in PBS ) for 16 h at 4oC before being embedded in Optimal Cutting Temperature ( OCT ) ( Raymond A Lamb Ltd , Eastbourne , U . K . ) in cryomoulds ( Park Scientific , Northampton , U . K . ) . Samples were frozen and stored at -80oC . 30 µm sections were cut , blocked and permeabilised for 10 min in a permeabilising solution containing 10% normal goat serum and 0 . 02% Saponin in PBS ( Sigma , Poole , UK ) . Sections were stained with 1∶1000 dilution of rat anti-mouse CD18+ monoclonal antibody ( clone M18/2 , BD Pharmingen ) , together with a 1∶500 dilution of rabbit anti-LPS O4 agglutinating serum ( Remel Europe Ltd ) , for 16 h at 4oC . Subsequently , sections were washed in PBS then incubated with 1∶200 Alexa Fluor 568-conjugated goat anti-rat antibody ( Invitrogen-Molecular Probes , Paisley , U . K . ) and a 1∶1000 dilution of Alexa Fluor 488-conjugated goat anti-rabbit antibody ( Invitrogen-Molecular Probes , Paisley , U . K . ) . All sections were mounted onto Vectabond-treated glass slides ( Vector Laboratories Ltd . ) using Vectashield containing DAPI ( Vector Laboratories Ltd . ) . All phase contrast and fluorescence images were captured using an Andor iXonEM+ 885 EMCCD camera coupled to a Nikon Ti-E microscope using a 100x/NA 1 . 4 oil immersion objective . Images were acquired with NIS-ELEMENTS software ( Nikon ) and processed using ImageJ . Fluorescence images were deconvolved using Huygens Deconvolution software ( Scientific Volume Imaging ) . Cell measurements were taken on a Nikon Ti-E microscope with NIS-ELEMENTS software . Immunofluorescence images from tissue sections were analysed multi-colour fluorescence microscopy ( MCFM ) using a Leica DM6000B Fluorescence microscope running FW4000 acquisition software . The effect of Salmonella infection on transepithelial resistance ( TER ) was determined for differentiated Caco-2 cells as previously described [34] . Briefly , the Caco-2 cells were grown on transwell inserts ( Corning , UK ) until differentiated ( 12–14 days ) , before the transepithelial resistance was measured for each well . Salmonella strains were then added to the cells at a multiplicity of infection ( MOI ) of 20 , and the cells incubated for 4 h . TER measurements were taken every hour and the results given as a ratio of TER ( t ) / ( t0 ) to show the percentage change in TER over the course of the experiment . Data were collated and analysed for statistical differences ( Student's t-test ) in Minitab . Samples for the assay of translocated effector proteins were isolated from differentiated Caco-2 cells grown in 6 well plates after infection with an MOI of 20 for 4 h . Excess bacteria were washed off before the cells were solubilised in 0 . 01% Triton X-100 and centrifuged to remove bacteria and host cell membranes . The host cell cytoplasmic fractions were analysed by western blotting with αSipB antibody .
We wished to identify and characterise putative Salmonella cytoskeletal gene homologues . A BLAST search of the S . Typhimurium genome sequence database ( www . ncbi . nlm . nih . gov ) [35] for the known E . coli actin-homologue MreB identified a putative mre operon of high sequence identity . Comparison of the Salmonella genes to those of E . coli showed 100% ( mreB ) , 88% ( mreC ) and 94% ( mreD ) homology at amino acid level , comparisons of these same genes to those in B . subtilis revealed sequence homologies of 57% , 24% and 27% respectively . In order to determine the localisation of MreB in Salmonella , vectors expressing N and C terminal fusions of MreB to GFP were used . The N-terminal fusion plasmid has already been described [36] , and we constructed a C-terminal fusion vector . Both constructs revealed a helical distribution of MreB along the long axis of the cell . The helices were discerned by assembling a series of z-stack images taken in successive planes by using Metamorph imaging and Huygens deconvolution software ( Figure 1A ) . The mreB gene appears to be essential in bacteria including Salmonella ( data not shown ) , and ΔmreB viable cells often contain compensatory mutations [37] . Each of the components of the cytoskeletal complex , for example MreB , MreC , or MreD , are essential for its function . As an alternative strategy to study the function of the cytoskeleton we therefore generated a mreC depletion strain under conditions designed to minimise selective pressures for undefined secondary compensatory mutations [37] . Using the lambda Red one-step gene disruption method , we successfully constructed a mreC::kan mutant in the S . Typhimurium wild-type strain SL1344 [31] . This mutation leaves intact the first gene in the operon mreB . Using bacteriophage P22int the mreC::kan mutation was then transduced into a genetically “clean” SL1344 strain harbouring plac-mre operon ( pTK521 ) [14] and the resulting strain designated ΔmreC . The plac-mre operon is a low copy number plasmid expressing the mre operon from the IPTG-inducible lac promoter . The identity of the mutation was confirmed by PCR and DNA sequencing . Expression of MreC was assessed by western blotting in the mutant strains , revealing no detectable levels MreC unless complementation was induced ( Figure S1 ) . In addition to the ΔmreC mutant , the lambda Red method was used to generate ΔmreD . When the morphology of the ΔmreC mutant was examined microscopically , the cells were no longer rod-shaped but spherical ( Figure 1B ) . Upon the addition of IPTG the morphology of the ΔmreC strain was restored to the wild-type rod shape . Under microscopic examination the ΔmreD mutant displays a similar morphological phenotype to the ΔmreC . WT cells were measured to be on 1 . 61 ( +/−0 . 49 ) µm in length and 0 . 75 ( +/− 0 . 17 ) µm in width , whereas the ΔmreC cells were 2 . 03 ( +/−0 . 60 ) µm in length and 1 . 21 ( +/−0 . 41 ) µm in width . Complementation of the ΔmreC mutant with 100 µM IPTG resulted in wild type shaped cells 1 . 82 ( +/−0 . 44 ) µm in length and 0 . 78 ( +/−0 . 24 ) µm in width . Measurements were taken from a minimum of 350 cells per strain . Growth rates of the strains were determined in LB media at 37°C revealing a ∼50% increase in the lag phase of the ΔmreC mutants ( Figure S2 ) , which subsequently grow at a comparable rate to that of the wild type or complemented mutant strains during log phase . The motility phenotype of ΔmreC was examined on semi-solid agar plates . In contrast to the isogenic parent , the ΔmreC cells were no longer motile . Surprisingly , this motility defect has not been reported in either E . coli or B . subtilis . Cellular and secreted proteins of the parent SL1344 and ΔmreC were examined by SDS-PAGE and western blotting using antibodies directed against the phase-1 and phase-2 flagellin subunits FliC and FljB . Neither of these subunits were present in either the secreted or cellular proteins , explaining the inability of the cells to swim ( data not shown ) . The non-motile phenotype was fully complementable in trans upon the addition of IPTG to the mutant strain harbouring pTK521 ( Figure S3 ) . We observed that the Salmonella ΔmreC depletion strain was non-motile and failed to express flagella subunits FliC or FljB . The regulation and assembly of flagella in Salmonella is complex . Flagella genes are arranged into 14 operons and their transcription is organised into a regulatory hierarchy of early ( class I ) , middle ( class II ) , and late genes ( class III ) [38] . The class I flhDC operon is the master regulator , with FlhD and FlhC forming a heterotetramer that is required for transcriptional activation of the class II genes , which encode the hook-basal body complexes and the alternative sigma factor FliA ( sigma28 ) . FliA alone or with FlhDC , activates expression of the class III operon genes , which encode the filament protein , hook-associated proteins , motor proteins , and chemotaxis proteins [39] , [40] . The class III genes are further subdivided into fliA-independent expression class IIIa or class IIIb [41] . In order to systematically investigate the mechanistic basis for the ΔmreC motility phenotype we have taken selected class I , II , and III regulated flagella gene promoter fusions to a luciferase reporter gene , and monitored their expression by luminescence in wild type and ΔmreC strains . Constructs with flhD ( class I ) , fliA , flgA , ( class II ) , and fliC ( class III ) promoters fused to the luciferase reporter gene were used . The reporter plasmid pSB401 has a promoterless luxCDABE operon and was used as a control . The class I flhD promoter displayed a reduction in the level of expression in ΔmreC compared to the wild-type strain suggesting the class I promoter has reduced activity . Notably greater changes in the expression profiles occur in other class II and class III genes . The class II promoters for the operons encoding the transcriptional regulators fliAZY and flgAM display significant reductions in expression levels in ΔmreC ( Figure 2 ) . As predicted from the western blotting data expression of the fliC class III promoter was significantly reduced . Collectively , the promoter-reporter activity data can account for the motility defect . Type 3 secretion systems are essential for the virulence of a range of pathogens including Salmonella [42] , [43] . The secretion apparatus assembles into a supramolecular needle-complex . Secreted effector proteins in the bacterial cytoplasm traverse through the needle-complex and the bacterial multi-membrane envelope , directly into host cells [44]–[46] . The apparatus anchors to the cell envelope via a multi-ring base . Salmonella possess two T3SS's encoded by pathogenicity islands ( SPI's ) . The SPI-1 T3SS is important for invasion of intestinal epithelial cells and the SPI-2 T3SS plays a central role in survival within the hostile environment of a macrophage . The SPI-1 T3S system translocates virulence effector proteins into the cytosol of epithelial cells resulting in rearrangements of the actin cytoskeleton which enable Salmonella to invade [47] . To investigate whether the mreC mutation has an impact on SPI-1 T3S , we used western blotting to determine the presence and functionality of the system using antibodies to an apparatus protein PrgH as well as the effector proteins SipA and SipC , in both SL1344 and ΔmreC . In contrast to the wild-type SL1344 , the T3S structural and effector proteins were not expressed in the cellular or secreted fractions from the ΔmreC depletion mutant ( Figure 3A ) . This suggests that SPI-1 T3S in the ΔmreC mutant is not fully functional . The expression and secretion phenotypes were fully complementable in trans upon the addition of IPTG ( data not shown ) . The functional assembly of SPI-1 T3SS was also confirmed using transepithelial resistance ( TER ) assays in differentiated Caco-2 cells , showing a reduced ability to disrupt epithelial tight junctions in the ΔmreC mutant compared to the wild type strain ( Figure 4 ) . To further assess the disruption of the functionality of the SPI-1 T3S , a translocation assay was performed in Caco-2 cells infected with the strains . Host cell cytoplasmic proteins were probed for the bacterial effector protein SipB using western blotting ( Figure S4 ) . This revealed the inability of the ΔmreC mutants to infect host epithelia and disrupt their tight junctions . In addition , ΔmreC was fully complementable in this assay following IPTG induction . The SPI-2 T3SS is pivotal for the establishment of the Salmonella containing vacuole ( SCV ) inside macrophages and subsequent survival [43] . We next investigated the effect of the ΔmreC mutation on the functionality of the SPI-2 T3SS . The strains were grown under SPI-2 inducing conditions and the T3S of the translocon protein SseB monitored . SseB together with SseC and SseD function as a translocon for other effector proteins and SseB is normally found associated with the outer surface of Salmonella . Thus membrane fractions were purified to monitor expression and T3S by western blotting . This revealed that in contrast to the SPI-2 negative control ( ssaV ) , SseB was secreted and associated with the bacterial membrane surface in both the wild-type and ΔmreC strains ( Figure 3A ) . This provides qualitative evidence to suggest that in contrast to the SPI-1 T3SS , the SPI-2 T3SS appears to remain functional . Several environmental signals and transcriptional factors modulate expression of the SPI-1 T3SS . We wished to understand the mechanistic basis by which expression of the SPI1-T3SS is down-regulated . Within SPI-1 there are key transcriptional activators which regulate expression of SPI-1 genes: HilC , HilD , HilA , and InvF . Both HilC and HilD activate expression of SPI-1 genes by binding upstream of the master regulatory gene hilA to induce its expression[48] . HilA binds and activates promoters of SPI-1 operon genes encoding the type 3 secretory apparatus , several secreted effectors , and the transcriptional regulator InvF . InvF activates expression of effector genes inside SPI1 and also effector genes outside SPI-1 such as sopB and sopE [47] . Expression of selected SPI-1 T3SS genes was monitored using transcriptional promoter reporters in ΔmreC , using constructs harbouring the hilA , hilC , hilD , invF and sopB promoters fused to the promoterless luxCDABE operon that produces light in response to gene expression [49]–[51] . Each construct was introduced into both wild-type SL1344 and ΔmreC depletion mutant , and the level of expression of the promoters in these strains monitored by luminescence assays . WT SL1344 and ΔmreC cells harbouring pCS26 or pSB401 vectors alone were used as controls , and did not produce any luminescence as expected . The reporter assays revealed that the SPI-1 transcription factor gene promoters for hilA , hilC , hilD , and invF were completely inactive in ΔmreC in contrast to the wild-type strain . However the promoter of sopB located in SPI-5 remained active but its activity was marginally lower than in the wild-type strain ( Figure 3B ) . The regulation of many T3SS genes often require multiple signals for maximal expression and clearly other signals remain in the ΔmreC depletion mutant which drive expression of the SopB in SPI-5 . Expression of SPI-2 T3SS genes were monitored using a transcriptional reporter for the SPI-2 gene ssaG , whose promoter was cloned upstream of the luxCDABE luciferase operon in the plasmid pMK1-lux [52] . The construct was transformed into wild-type SL1344 and ΔmreC , and the luminescence and OD600 measured during growth in SPI-2 inducing conditions ( Figure 3B ) . The ssaG promoter remains active in the ΔmreC mutant although expression appears to be delayed , and is marginally less than in WT . This evidence supports the western blot data with αSseB and suggests that in contrast to the SPI-1 T3SS , the SPI-2 T3SS remains functional in the absence of the cytoskeleton . Two-component regulatory systems are vital in sensing environmental and cell surface signals , enabling bacteria to rapidly adapt to ever changing conditions [53] , [54] . These signals are detected by histidine protein sensor kinases , which subsequently transfer phosphate groups to an aspartate residue in the response regulator proteins , thus modulating their regulatory activities . The environmental signals are thus transmitted by a phosphorelay system to regulate gene expression . In order to identify putative regulators of the ΔmreC observed phenotypes , we have constructed knockout mutations in a range of two-component systems . As an initial screen , a panel of nine separate two-component system mutant strains were constructed as double mutants with ΔmreC . One two-component system sensor kinase mutation ΔrcsC resulted in recovery of SPI-1 effector expression in the ΔmreC background as judged by western blotting using αSipC sera ( Figure 5 panels A and B ) . Interestingly the amount of SipC protein expressed and secreted from the cell was less than the wild-type suggesting there are additional repressors continuing to operate ( Figure 5 panels A and B and Figure S5 ) . Furthermore , disruption of rcsC also significantly de-repressed motility ( Figure 6 and Figure S6 ) in a ΔmreC mutant similar to SPI-1 expression , again suggesting there are additional repressors involved . Expression of the RcsC protein in trans was able to restore the phenotype of ΔmreC ΔrcsC back to the equivalent of a ΔmreC strain , with respect to repressing SPI-1 type 3 secretion and motility . These complementation studies provide further evidence supporting the regulatory role of RcsC in the ΔmreC phenotypes ( Figure S7 ) . Rcs is a highly complex multi-component phosphorelay system and was originally identified in regulating genes involved in capsule synthesis in Escherichia coli [55] , [56] . The RcsC sensor kinase phosphorylates RcsD , which subsequently phopshorylates the DNA binding response regulator RcsB . The unstable RcsA protein and additional auxillary proteins can also interact and regulate RcsB . The Rcs system is involved in down-regulating the expression of flagella , SPI1-T3S and increasing biofilm formation [57] . We therefore also constructed ΔmreC ΔrcsB , ΔmreC ΔrcsD , ΔmreC ΔrcsDB and ΔmreC ΔrcsCBD mutants , which however did not restore either SPI-T3S or motility ( Figures 5 , 6 , and S6 ) . We propose that in the absence of RcsC signalling , phosphorylated levels of RcsB are depleted enabling de-repression of FlhDC and motility . The presence of RcsDB appears essential for restoring motility in the absence of RcsC [55] . The functionality of SPI-1 T3SS in the ΔmreC ΔrcsC and ΔmreC ΔrcsDB mutants were assessed in a TER assay , which revealed partial restoration of tight junction disruption in the ΔmreC ΔrcsC mutant , but not in the ΔmreC ΔrcsDB ( Figure S8 ) . It has been suggested that the outer membrane protein RcsF may perceive some of the environmental signals necessary to activate the Rcs phosphorelay system . To investigate this we constructed a ΔmreC ΔrcsF mutant which failed to restore motility or SPI-1 T3S and appeared phenotypically identical to ΔmreC ( Figure 5 , S6 ) . This would suggest that RcsF is not involved in the observed ΔmreC phenotypes . Furthermore as the auxillary protein RcsA can interact and regulate RcsB , we therefore disrupted the rcsA gene in ΔmreC and which also resulted in no impact on the observed phenotypes ( Figure 5 , S6 ) . In summary , we propose that RscC is sensing cell surface perturbations [58] in ΔmreC , resulting from a disrupted cytoskeleton , and subsequently down-regulating the expression of SPI-1 T3S and motility . This signalling appears to be independent of both RcsF and RcsA . A cell permeable compound named A22 [S- ( 3 , 4-Dichlorobenzyl ) isothiourea] has been demonstrated to perturb MreB function [59] . As an alternative approach to genetically disrupting the essential gene mreB , we exposed wild-type Salmonella cultures to A22 and observed a morphological change from rod to spherical-shaped cells . In addition we phenotypically screened and tested A22-treated cells for motility and T3S . The A22-treated cells were phenotypically identical to ΔmreC with respect to cell shape , motility , SPI-1 T3S , and also SPI-2 T3S ( data not shown ) . The effects of A22 were completely reversible following its removal ( data not shown ) . Thus the chemical genetic inactivation of MreB , independently corroborates the phenotypic observations made with ΔmreC . The ΔmreC defect clearly has an impact on the expression of important virulence determinants of Salmonella in vitro . We therefore wished to investigate the contribution of the bacterial cytoskeleton on the virulence of Salmonella in vivo using the mouse model . We observed that the SPI-1 T3SS in ΔmreC is completely down-regulated , and as this virulence system is important for infection through the oral route of inoculation the strain would be attenuated . We therefore explored the colonization of ΔmreC using the intravenous route allowing us to examine the impact of the host on the further down-stream stages of infection . Groups of 5 female C57/BL6 mice were inoculated intravenously with circa 103 colony forming units of either control SL1344 or ΔmreC . The times taken for clinical symptoms to appear were determined . Viable bacterial numbers in the spleen and liver for SL1344 were determined at days 1 and 4 , and ΔmreC at days 1 , 4 , 7 , and 10 . The in vivo bacterial net growth curves vividly demonstrate two clear phenotypic effects upon the growth of ΔmreC compared to the wild-type . Firstly , there is a greater initial kill of ΔmreC , and this is secondly followed by a slower net growth rate . However , in spite of the reduced growth rate of ΔmreC , the bacterial numbers steadily increase over 6 days . This eventually causes the onset of clinical symptoms necessitating termination of the experiment at day 10 ( Figure 7 ) . During these stages Salmonella infect and multiply within macrophages and the SPI-2 T3SS is essential for survival . Thus providing further evidence to support the presence of a functional SPI-2 T3SS in ΔmreC . Collectively , these observations imply the mreC defect reduces the virulence of the strain , but does not completely abrogate its ability to multiply and cause disease systemically in vivo . Strains recovered from in vivo passage were tested for changes in morphology , motility and T3S , and were found to be identical to the input strain . Furthermore the in vivo morphology of the strain within livers and spleens was determined by immunofluorescence microscopy . Sections of livers and spleens were taken and stained as described in the materials and methods . Figure 8 demonstrates the Salmonella ΔmreC mutant strain retains the round morphology in vivo compared to the rod shaped wild-type control . Collectively these data suggests that the mutation has remained stable during the in vivo passage for the virulence phenotypes tested . The regulation and assembly of SPI-1 T3SS and flagella are complex . When the bacterial cytoskeleton is disrupted both the SPI-1 T3SS and flagella expression are down-regulated . A hypothesis is that the integrity of the cytoskeleton is essential for the correct assembly of these complex macromolecular structures and in its absence the SPI-1 and flagella gene expression are down-regulated to conserve resources . Alternatively , in the absence of a functional cytoskeleton the bacterial cell is stressed and shuts down the expression of energetically expensive “non-essential” machinery . To test these ideas we wished to force on the expression of SPI-1 T3S and flagella genes , and examine whether these systems are correctly assembled and functional . We therefore expressed in trans from heterologous inducible promoters either the flagella master regulator FlhDC or the SPI-1 T3S regulator HilA in a panel of strains including ΔmreC . Strikingly , expression of FlhDC restored both the expression and assembly of flagella on the cell surface as determined by fluorescence microscopy ( Figure 9A ) and motility assays ( data not shown ) in ΔmreC . Furthermore , expression of HilA in trans up-regulated expression of the SPI-T3SS and its assembly on the cell surface as determined immunofluorescence microscopy ( Figure 9B ) western blotting with αSipB antibody ( Figure S9 ) or functionally by TER measurements ( Figure 4 ) . In contrast to SPI-1 T3SS and flagella , the expression of the SPI-2 T3SS was not turned off in the ΔmreC mutant as shown in ( Figure 9C ) . Interestingly , in WT cells the SPI-1 T3S apparatus and flagella appear to be present in around six to eight copies mainly along the long axis of the cell . In marked contrast the SPI-2 apparatus is typically present in one or two copies located at the poles of the bacterial cell [42] , whereas their localisation appears less clear in the ΔmreC mutant , possibly due to perturbations in the cell envelope and the indistinct cell polarity in these cells caused by disruption of the cytoskeleton . The complementation of the functional assembly of SPI-1 T3SS was also confirmed using TER assays , where the levels of decrease in resistance after infection with ΔmreC strain reverted to that of the parent strain upon induction of the transcriptional regulator hilA ( Figure 9B and S9 ) , or complementation of the ΔmreC mutation ( Figure 4 ) . Taken together the data support the notion that the cytoskeleton is not required for the correct assembly of these virulence factors but essential for their expression .
Bacterial cells possess dynamic cytoskeletons composed of diverse classes of self-assembling polymeric proteins . Some of these proteins resemble eukaryotic actin , tubulin , and intermediate filaments both structurally and functionally [5] , [7] , [11] , [12] . The bacterial tubulin FtsZ plays a key role in cell division . Bacterial actins provide vital functions in maintaining cell morphology , segregating DNA , and positioning bacterial organelles . It has recently been demonstrated in Helicobacter pylori , that MreB is essential not for cell shape but for maintenance of the full enzymatic activity of urease , an essential virulence factor [60] . Furthermore the MreB cytoskeleton is also essential for the polar localisation of pili in Pseudomonas aeruginosa [61] . Using a variety of approaches we have demonstrated the importance of the bacterial cytoskeleton in the pathogenicity of Salmonella . MreC and MreD form a complex in the cytoplasmic membrane , which subsequently interacts with MreB . The mreB gene appears to be essential in many organisms including as we discovered in Salmonella . Viable mreB mutants often contain compensatory changes in other genes e . g . ftsZ which compensate for the lethality of the mreB lesion [37] . As an alternative strategy to investigate the function of the bacterial cytoskeleton and avoid these deleterious effects , we carefully constructed depletion mutants of mreC in strains harbouring a single-copy plasmid expressing the MreB operon from the lac promoter . In addition we confirmed the phenotypic effects of the mreC genetic lesion by disrupting the functions of MreB using a chemical genetics approach and inactivating MreB with A22 . Removal of the gratuitous inducer IPTG from the growth medium of the ΔmreC depletion mutant resulted in cells changing from rod to a spherical shaped morphology . Using fluorescence microscopy MreB was observed to be no longer distributed in a helical fashion throughout the cell but rather diffusely throughout the cytoplasm ( data not shown ) . Presumably MreB polymers are no longer able to contact the cytoplasmic membrane via MreD attachment sites resulting in mis-assembly of the entire cytoskeleton . In growing cells , this disruption of the cytoskeleton leads to loss of the rod-shape . We next examined the motility of ΔmreC depletion strain to assess the functionality of flagella . The strains were non-motile and western blotting revealed absence of the flagellin filament subunit proteins FliC and FljB in both secreted and also cytoplasmic protein fractions , suggesting expression of these alternative subunits had been switched off . Flagella gene expression is complex and involves a regulatory hierarchy of Class I , Class II , and Class III genes [38] . The class I flhDC operon is the master regulator , and FlhDC complex is required for transcriptional activation of the class II genes including the specialized flagellar sigma factor FliA . FliA alone or with FlhDC complex , activates expression of the class III operon genes encoding motor proteins , hook-associated proteins , the filament protein , and chemotaxis proteins [39] , [40] . Expression of the FlhDC complex was reduced but still appeared comparable between the wild-type and the ΔmreC suggesting changes in the promoter activity of flhDC alone are not responsible for the observed phenotype . Class II gene expression was significantly reduced . Expression of the Class III gene fliC was completely down-regulated confirming the western blot observations . Hence these independent observations are in accordance with the ΔmreC motility data . Thus in the absence of the cytoskeleton expression of class II and class III flagella genes appears to be down-regulated . Expression of the SPI-1 T3S system is essential for invasion of intestinal epithelial cells and the SPI-2 T3SS plays a central role in survival within the hostile environment of a macrophage [43] . Western blotting revealed the SPI-1 T3S structural protein PrgH and the effectors SipA and SipC were no longer expressed or secreted in the ΔmreC depletion mutant . The phenoptype was fully complementable by the addition of IPTG . Several environmental signals and transcriptional factors modulate expression of the SPI-1 and SPI-2 T3SS [43] , [45] , [62] . We wished to understand the mechanistic basis by which expression of the SPI1-T3SS is down-regulated . Within SPI-1 there are key transcriptional activators which regulate expression of SPI-1 genes: HilC , HilD , HilA , and InvF . Using promoter-luciferase transcriptional reporter assays it was revealed that the SPI-1 transcription factor gene promoters for hilA , hilC , hilD , and invF were completely inactive in ΔmreC , in marked contrast to the control wild-type strain . Surprisingly , the promoter of sopB located outside of SPI-1 in SPI-5 remained active but its activity was marginally lower than in the wild-type strain . The regulation of many T3SS genes often require the input of multiple signals for maximal expression and clearly other signals remain in the ΔmreC depletion mutant which drive expression of the SopB in SPI-5 . It therefore appears that the SPI-1 T3SS is completely down-regulated in the absence of an cytoskeleton by an unidentified regulatory factor . In contrast , the SPI-2 T3SS remains functional as evidenced by western blotting with SseB antibody and promoter-reporter assays . This is further corroborated with the in vivo evidence that following systemic inoculation , ΔmreC is able to survive and multiply within the host . This takes place within the hostile environment of the macrophage where SPI-2 T3S is essential for biogenesis of the Salmonella containing vacuole and survival [43] , [63] , [64] . We wished to gain further insights into the mechanistic basis of the down-regulation of both SPI- T3SS and motility in ΔmreC . Two-component systems play an essential role in sensing and responding to environmental and cell surface signals [54] . To investigate if two-component systems contribute to the regulation of the ΔmreC phenotypes , we constructed a panel of separate two-component system mutant strains in an ΔmreC background . The double mutants were screened for recovery of motility and expression of the SPI-1 T3SS . A mutation in the rcsC sensor kinase gene resulted in significant but not complete recovery of both motility and expression of the SPI-1 T3SS . The Rcs phosphorelay system regulates a broad range of genes from capsule synthesis in E . coli to increasing biofilm formation [58] . RcsC also plays an important role in repressing expression of flagella and SPI-1 T3SS in Salmonella Typhi [57] . The RcsC sensor kinase normally phosphorylates RcsD , which subsequently phosphorylates the DNA binding response regulator RcsB . However , in ΔmreC ΔrcsDB and ΔmreC ΔrcsCBD there was no restoration of either motility or expression of the SPI-1 T3SS suggesting that RcsC signals repression and requires the presence of rcsDB to mediate this effect . We propose that in ΔmreC , the sensor kinase RscC detects cell surface perturbations and down-regulates expression of flagella and the SPI-1 T3S apparatus [58] . This signalling is independent of both the outer membrane lipoprotein RcsF sensor and the auxilliary regulatory protein RcsA . There are a number of explanations to provide a bacterial rational for this shutdown in expression . In the absence of a functional cytoskeleton the flagella and SPI-1 T3SS are either not being correctly assembled , triggering a feedback loop to repress expression , or alternatively are down-regulated to prevent the cell from wasting valuable resources under these conditions . To test the assembly idea , we forced on the expression of flagella and SPI-1 T3SS genes by expressing the regulators flhDC or hilA in trans in ΔmreC . Using independent methods we observed the correct assembly and function of these macromolecular machines suggesting the cytoskeleton is not essential for functionality . The cytoskeleton could also have a role in sensing cellular stress , as has recently been suggested by Chiu and colleagues [65] . They propose that the integrity of the cytoskeleton may be exploited by the cell to monitor oxidative stress and physiological status . If the cytoskeleton disintegrates in the absence of MreC , this may be sensed by the cell leading to a shut-down of the SPI-1 T3S apparatus and down-regulation of flagella protein expression . We have provided mechanistic insights into the regulation of motility and SPI-1 T3S in ΔmreC . We have identified the two-component system sensor RcsC as an important regulator controlling expression of these systems , presumably as a consequence of sensing membrane perturbations brought about by the disruption of the cytoskeleton [58] . With a non-functional SPI-1 T3SS , we would expect the ΔmreC would be attenuated in mice when administered by the oral route as it is unable to invade intestinal epithelial cells by the SPI-1 T3SS . We therefore explored the colonization of ΔmreC in vivo using the intravenous route of inoculation [66] . This provides an opportunity to examine the impact of ΔmreC on the down-stream stages of infection . Salmonella infect and multiply within macrophages during the systemic stages of infection . Survival within the hostile environment of the macrophage would require a functional SPI-2 T3SS in the Salmonella-containing vacuole to remodel the host cell environment and survive attack from reactive oxygen free radicals [64] , [67] , [68] . By examining the in vivo net bacterial growth curves within livers and spleens two clear phenotypic effects were revealed with ΔmreC compared to the wild-type . Greater initial killing of ΔmreC is followed by a slower net growth rate with the bacterial numbers steadily increasing over six days . Clinical symptoms begin to appear and by day ten these symptoms necessitate termination of the experiment . The phenotypic data clearly imply the ΔmreC defect reduces the colonization of Salmonella , but does not completely abrogate its ability to multiply and cause disease systemically in vivo . This would suggest that the second T3S in Salmonella encoded on SPI-2 remains sufficiently functional to permit growth in the absence of the cytoskeleton . In the absence of an intact cytoskeleton in ΔmreC the expression of the SPI-1 T3SS and flagella are clearly down-regulated . Strikingly however , the SPI-2 T3SS appears to remain functional contributing to the virulence of the ΔmreC strain observed in vivo . A possible explanation could be that the regulation of the SPI-2 T3SS is co-ordinated independently of the integrity of the cytoskeleton in contrast to flagella and SPI-1 T3SS . Collectively these data highlight the importance of the bacterial cytoskeleton in the ability of Salmonella to cause disease , and may provide opportunities for the development of new antimicrobials to target the cytoskeleton .
|
Salmonella are major global pathogens responsible for causing food-borne disease . In recent years the existence of a cytoskeleton in prokaryotes has received much attention . In this study the Salmonella cytoskeleton has been genetically disrupted , causing changes in morphology , motility and expression of key virulence factors . We provide evidence that the sensory protein RcsC detects changes at the cell surface caused by the disintegration of the bacterial cytoskeleton and modulates expression of key virulence factors . This study provides insights into the importance of the integrity of the bacterial cytoskeleton in the ability of Salmonella to cause disease , and thus may provide a novel target for antimicrobial drugs or vaccines .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"bacterial",
"diseases",
"cellular",
"structures",
"infectious",
"diseases",
"cell",
"biology",
"biology",
"microbiology",
"molecular",
"cell",
"biology"
] |
2012
|
The Bacterial Cytoskeleton Modulates Motility, Type 3 Secretion, and Colonization in Salmonella
|
Mammalian sex chromosomes stem from ancestral autosomes and have substantially differentiated . It was shown that X-linked genes have generated duplicate intronless gene copies ( retrogenes ) on autosomes due to this differentiation . However , the precise driving forces for this out-of-X gene “movement” and its evolutionary onset are not known . Based on expression analyses of male germ-cell populations , we here substantiate and extend the hypothesis that autosomal retrogenes functionally compensate for the silencing of their X-linked housekeeping parental genes during , but also after , male meiotic sex chromosome inactivation ( MSCI ) . Thus , sexually antagonistic forces have not played a major role for the selective fixation of X-derived gene copies in mammals . Our dating analyses reveal that although retrogenes were produced ever since the common mammalian ancestor , selectively driven retrogene export from the X only started later , on the placental mammal ( eutherian ) and marsupial ( metatherian ) lineages , respectively . Together , these observations suggest that chromosome-wide MSCI emerged close to the eutherian–marsupial split approximately 180 million years ago . Given that MSCI probably reflects the spread of the recombination barrier between the X and Y , crucial for their differentiation , our data imply that these chromosomes became more widely differentiated only late in the therian ancestor , well after the divergence of the monotreme lineage . Thus , our study also provides strong independent support for the recent notion that our sex chromosomes emerged , not in the common ancestor of all mammals , but rather in the therian ancestor , and therefore are much younger than previously thought .
Several recent studies [1–3] of mammalian retroduplicate genes ( i . e . , intronless duplicate genes generated by the reverse transcription of mRNAs from “parental” source genes [4 , 5] ) have revealed a peculiar pattern with respect to their chromosomal origin: an excess of functional retrogenes stem from the X chromosome . It was suggested that these autosomal retroduplicate counterparts of X-linked genes carry out functions of the silenced parental genes that are necessary or advantageous during the transcriptional silencing of the X chromosome in the meiotic phase of spermatogenesis ( termed male meiotic sex chromosome inactivation [MSCI] [6] ) , and were therefore selectively fixed during evolution [1 , 7] . In support of this notion , a number of X-derived retrogenes were found to be expressed in testis [1–3 , 7] , and for some retrogenes , it was shown that they are expressed during meiosis while their parental genes are shut off ( e . g . , [8] ) . In addition , loss of function of two X-derived retrogenes was shown to lead to severe defects of male meiotic functions in humans and mice [9–11] , suggesting that such genes are needed to replace their parental genes during male meiosis . When did the selectively driven , out-of-X movement of genes begin ? If MSCI is responsible for export of gene copies from the X , answering this question would provide a unique means to date the evolutionary onset of MSCI .
To trace the evolution of gene movements in mammals , we first screened for intronless retroposed gene copies ( retrocopies ) and their parental genes in three eutherian ( “placental” mammal ) genomes and one metatherian ( “marsupial” ) genome ( opossum ) , using a refinement of our previously described procedure [2] ( Materials and Methods ) . This analysis identified several thousand retrocopies in each of the therian genomes analyzed ( Table 1 ) . Thus , the process of retroposition has significantly shaped , not only the genomic landscape of eutherians [1 , 2] , but also that of its sister lineage , the marsupials . We then extracted two subsets from these retrocopy data for each species ( see Materials and Methods for details ) : One was enriched for functional retrocopies ( retrogenes; Tables 1 and S1–S4 ) , whereas the other contained retropseudogenes with open reading frame disruptions ( premature stop codons and frameshifts ) that likely preclude gene function ( Table 1 ) . The analysis of chromosomal locations of parental genes revealed that X-linked genes of all genomes analyzed have spawned a large excess of functional retrogenes compared to autosomal genes , whereas no such bias is observed for parental genes that gave rise to retropseudogenes ( Table 1 ) . Thus , preferential fixation of functional X-derived genes by natural selection occurred , not only in eutherians [1 , 2] , but also in metatherians . The latter is consistent with a recent study that showed that MSCI occurs in marsupials [12] . Before examining the evolutionary history of X movement patterns in more detail , we sought to obtain further evidence for the hypothesis that MSCI is the driving force for the preferential copying of genes from the X to autosomes , which was so far based on the analysis of individual genes ( see Introduction ) . To this end , we analyzed expression patterns of retrogenes and their parental genes using genome-wide murine expression data [13] from testicular germ-cell populations , total testis , ovary , and 14 somatic tissues ( Figure 1A; Materials and Methods ) . We find that all parental genes are broadly expressed ( median: 16 , mean: ∼14 . 2 tissues ) , in significantly more tissues than other genes in the genome ( median: 15 , mean: ∼11 tissues , p < 10−11 , Mann-Whitney U test; Figure 1A ) , which substantiates previous notions that retrogenes stem from housekeeping genes with important functions in all or most tissues [2 , 7] . In contrast , the majority of X-derived retrogenes ( 12 of 17 , ∼71% ) are specifically expressed in testes ( Figure 1A and Table S5 ) . X-derived retrogenes show a striking excess of testis-specific cases compared to their parental genes ( 0 of 21 specifically expressed in testes ) or other genes in the genome ( 790 of 14 , 991 , 5 . 3%; p < 10−17 , Fisher exact test ) . We note that similar patterns have been described in Drosophila [14] , a genus in which the out-of-X movement of genes was originally observed [15] . X-derived retrogenes in our data are also significantly more frequently expressed ( specifically or nonspecifically ) in testis ( 17 of 17 , or 100% , with testis expression ) compared to other , autosome-derived retrogenes ( 41 of 53 , or ∼77% , with testis expression; two-tailed p < 0 . 05 , Fisher exact test; 26 of 53 , or 49% , are testis-specific ) . This points to a selective enrichment of testis functions among X-derived retrogenes during evolution , although retrogenes generally seem to be frequently expressed in testis , consistent with previous studies [2 , 3] . In order to functionally compensate for their parental genes in testes ( Figure 1A ) , expression of X-derived retrogenes would be specifically required in testicular meiotic germ cells ( spermatocytes ) , where their parental genes are silenced , but not in premeiotic spermatogonia ( Figure 1B ) . Our expression analysis of premeiotic , meiotic , and postmeiotic cells revealed a striking pattern ( Figure 1B and Table S5 ) , consistent with a compensation function of retrogenes during but—surprisingly—also after meiosis ( see [6] for recent evidence of active postmeiotic silencing of the X ) . In spermatogonia , X-linked parental genes show high and their retrogene copies low expression activity . Conversely , X-derived retrogenes are highly expressed in spermatocytes and postmeiotic spermatids , while their parental genes are silenced . The overall propensity of retrogenes—including retrogenes with autosomal progenitors ( Figure 1B ) —to be expressed in spermatocytes/spermatids is probably due to the “hypertranscription” state of autosomal chromatin in these cell types ( [3] and references therein ) . This likely facilitated the initial transcription of retrocopies after their emergence , allowing them to obtain functions in the late stages of spermatogenesis . However , X-derived retrogenes are more frequently expressed in spermatocytes than retrogenes with autosomal parental genes ( 16 of 17 , or 94% , X-derived vs . 39 of 53 , or 74% , autosome-derived; one-tailed p < 0 . 1 , Fisher exact test ) , and at higher levels ( median log2-transformed expression signal: ∼10 . 9 vs . 8 . 8 , one-tailed p < 0 . 05 , Mann-Whitney U test ) . A similar pattern is observed in postmeiotic spermatids ( 100% vs . 75% expressed , two-tailed p < 0 . 05; median signal: ∼11 . 0 vs . ∼10 . 5 , one-tailed p = 0 . 15 ) . In addition , based on expression cluster analyses [13] ( Materials and Methods ) , we find that a significant excess of X-derived retrogenes show transcriptional induction in meiosis when compared to retrogenes that stem from autosomes ( 10 of 17 , or ∼59% , vs . 16 of 53 , or ∼30%; two-tailed p < 0 . 05 , Fisher exact test ) . Our expression analyses substantiate the hypothesis that retrogenes that stem from the X have been fixed during evolution and shaped by natural selection to compensate for parental ( housekeeping ) gene silencing during ( and after ) MSCI . Thus , sexual antagonism ( i . e . , evolutionary conflict between males and females ) , which was previously considered as an alternative driving force for the fixation of X-derived retrogenes [1 , 16] , likely played less significant roles for the selectively driven export of X-linked genes in mammals ( at least for those that are specifically expressed during/after meiosis ) . In contrast to the mammalian pattern , X chromosome inactivation during spermatogenesis does not seem to be a major contributor to the out-of-X movement of genes in Drosophila [17] . Rather , it appears that the increased residency time of the X chromosomes in females accounts for the observed pattern in this genus [17] . Thus , interestingly , the predominant selective forces associated with the export of X-linked genes appear to differ between fruitflies ( sexually antagonistic selection ) and mammals ( MSCI ) . To date the evolutionary onset of the out-of-X movement of genes in mammals , we screened for the presence/absence of human retrogenes in genomes representing the three major mammalian lineages ( see Materials and Methods for details ) . In addition to three eutherian and one marsupial genome ( opossum ) , this analysis included a genome ( platypus ) of the most basal mammalian lineage , the egg-laying monotremes ( Figure 2 ) . For the purpose of this dating , it is necessary to focus the X-related part of the analysis on the ancestral part of the human X , termed X conserved region [18] ( XCR ) , which is shared across mammals . The dating of human XCR-derived retrogenes uncovered a striking pattern ( Figure 2 ) . Although a number of autosomal retrogenes were produced in the common mammalian ancestor more than approximately 210 million years ago ( Mya ) as well as in the common therian ancestor between approximately 180 and 210 Mya , X-derived genes only started to appear after the eutherian–metatherian split ( <180 Mya ) on both of the descendent lineages . The approximately 1 , 300% excess of X-derived retrogenes in the common human–dog ancestor ( branch C ) is highly significant ( p < 0 . 01 , resampling test ) , which suggests strong selection driving the fixation of X-derived retrogenes between 90 and 180 Mya on the eutherian lineage . Similarly , there is an approximately 860% excess of old ( pairwise dS > 0 . 5 between parental gene and retrogene ) marsupial-specific X-derived retrogenes , which suggests selective export of genes from the X early in the metatherian lineage ( i . e . , early on branch H; p < 0 . 01 , resampling test ) . Importantly , the X-to-autosome parental gene ratio is significantly higher on branch C ( human–dog ancestor ) than on branch B ( common therian ancestor ) , where the zero observed out-of-X cases correspond to the random expectation ( two-tailed p < 0 . 01 , Fisher exact test ) . These findings demonstrate a significant shift in the selective forces—likely due to the emergence of MSCI—driving genes out of the X around the time of divergence of the two therian lineages . Thus , selective gene export driven by chromosome-wide MSCI originated either just before ( not leaving enough time for an X-skew in the retrogene generation pattern on branch B ) or—less parsimoniously—soon after the eutherian–marsupial split around 180 Mya , which would imply two independent origins of MSCI in eutherians and metatherians , respectively . We find that the first described X-derived human retrogene with parental replacement function , PGK2 ( [8] ) , originated in the common human–dog ancestor approximately 90–180 Mya ( Table 2 ) , contrary to a previous study that suggested an origin in the therian ancestor [19] . The PGK1 parental gene has independently spawned three PGK retrogenes on the marsupial lineage ( Figure S1A; Table S4 , identifiers MD5 , MD6 , and MD12 ) . One of these marsupial PGK genes ( Table S4 , MD6 ) shows a high divergence from its parental gene at silent sites ( dS ∼ 0 . 77 , corresponding to an age of roughly 140 million years ) , indicating an origin shortly after the eutherian–metatherian split . The Cetn-2 ( Centrin ) parental gene ( a gene required for centromere structure and function [20] ) similarly gave rise to retrogenes independently in the human–dog ancestor ( Figure S1B and Table 2 ) and in metatherian evolution ( Figure S1B; Table S4 , MD9 ) . Both the PGK and Centrin retrogenes evolved highly specific testis expression patterns in eutherians [8 , 21] ( Figure 1 and Table S5 , identifiers MM6 and MM17 ) . These data suggest a strong selective pressure to generate autosomal copies of these important housekeeping genes soon after the evolutionary onset of MSCI in both placental and marsupial mammals . Several other parental genes with fundamental cellular functions also spawned functional retrogene copies early in eutherian or metatherian evolution . For example , retrogenes encoding proteins involved in protein synthesis ( Table 2 , HS7 and HS11 ) , the core transcription machinery ( HS6 and HS13 ) , nucleotide synthesis ( HS9 ) , and energy metabolism ( HS8 ) originated in the common eutherian ancestor . Our study on chromosomal gene movements in mammals has general implications for the origin and evolution of mammalian sex chromosomes . The X and Y chromosomes started to evolve from an ancestral autosomal pair when the SRY gene—the primary sex determinant—emerged on the proto-Y chromosome in a mammalian ancestor [22 , 23] . Suppression of recombination between the proto-X and -Y chromosomes initially encompassed the long arm of the X chromosome ( containing the SRY gene ) and then spread to include the entire XCR . This barrier to recombination between the X and Y was crucial for their differentiation and therefore marks the origin of these sex chromosomes [24] . It likely also triggered silencing of genes in the unpaired ( nonrecombining ) regions of the X during the meiotic phase of spermatogenesis—the process of MSCI—through a more general molecular mechanism ( meiotic silencing of unsynapsed chromatin , MSUC ) that silences unpaired DNA during meiosis [6 , 25] . Our study of chromosomal gene movements suggests that MSCI emerged late in the common therian ancestor , around 180 Mya . Intriguingly , given that MSCI likely reflects the spread of the recombination barrier between the X and Y ( see above ) , this observation also suggests that these chromosomes originated after the separation of the therian and monotreme lineages , which is later than the previously suggested origin [22 , 23] , in the common ancestor of all mammals approximately 240–310 Mya . Our findings are consistent with a recent study of monotreme sex chromosomes [26] . Contrary to previous studies , which suggested that the platypus X chromosomes are related to both the therian X and bird Z chromosomes [27 , 28] , this work only finds homologous relationships between the sex chromosomes of monotremes and birds [26] . A recent origin of X and Y sex chromosomes in therians also implies that all other properties and evolutionary forces associated with the differentiated X and Y chromosomes—such as somatic X chromosome inactivation ( XCI ) seen in females and sexual antagonism—emerged recently in therians . A recent origin of XCI—which may be derived from MSCI ( [25 , 27] ) —in the common therian ancestor is consistent with the presence of XCI in eutherians and marsupials , as well as the recent origin of the XIST gene , crucial for XCI in eutherians , in the common eutherian ancestor [29 , 30] . In conclusion , our analyses of gene movement patterns have shed new light on the origin and properties of mammalian sex chromosomes . They suggest that in addition to the well-known phenotypes that distinguish therian mammals from monotremes , such as placentation , which evolved together with viviparity [31] , therian mammals have evolved a unique sex chromosome system that includes dosage compensation and MSCI .
We identified retrocopies in the human , mouse , dog , and opossum genomes using a previously described procedure [2] . The analysis was based on Ensembl [32] ( http://www . ensembl . org ) genome annotations ( versions: human 29 , mouse 32 , dog 34 , and opossum 41 ) . dN and dS statistics for retrocopy/parental gene pairs were estimated using the tool codeml as implemented in the PAML package [33] ( http://abacus . gene . ucl . ac . uk/software/paml . html ) . For each species , we established a dataset enriched for functional retrogenes , i . e . , retrocopies with intact ORFs and dN/dS less than 0 . 5 ( p < 0 . 05 ) in the comparison between the parental genes and retrogenes ( suggesting purifying selection on both the parental and retrocopy sequence [1] ) . The test is based on a likelihood ratio test [34] that compares a codeml model in which dN/dS is fixed to 0 . 5 ( null model ) to a model where dN/dS is estimated from the data . Retrocopies were mapped to Ensembl annotations by overlapping the retrocopy coordinates with those from Ensembl exons . We used normalized mouse microarray data generated in a previous study [13] for the parent–retrogene expression analyses . Parental genes were linked to probe sets using the Ensembl annotation provided by Affymetrix . Given that a number of retrocopies that integrated into introns of “host” genes are annotated in Ensembl as alternative splice variants of their host genes , we used a distinct procedure to link Affymetrix probe sets to retrocopies: First , we used BLAT ( [35] ) to map all probe sets onto the mouse genome sequence . A probe set was then assigned to a retrocopy if its best hit overlapped with the retrocopy . When multiple probe sets represented the expression of a retrocopy or a parental gene , we selected the probe set with the highest expression value in all testis measurements and considered it as representative . We excluded highly similar parent–retrogene pairs that may potentially cross-hybridize by requiring a minimum divergence at silent sites ( dS ) of 0 . 1 . The overall procedure yielded expression data for 70 retrogenes and 116 parental genes ( 62 parent–retrogene pairs for which expression data are available for both members of the pair ) . Expression data preprocessing , statistical filtering , gene clustering , and the testis-specificity determination procedure were established using the procedures described in [13] . In this study , probe sets were clustered and classified into four broad somatic ( SO ) , mitotic ( MI ) , meiotic ( ME ) , and postmeiotic ( PM ) expression clusters , showing transcriptional peaks in Sertoli cells , spermatogonia , pachytene spermatocytes , and round spermatids , respectively . We empirically considered a gene to be significantly expressed , when its probe set had a signal greater than log2 ( 100 ) . We used two criteria to establish testis specificity for a gene . First , we chose all probe sets that are expressed in at least one male germline sample ( Sertoli cells , spermatogonia , spermatocytes , spermatids , tubules , or total testis ) , but not in any somatic tissue analyzed ( expression signals < log2 ( 100 ) ) . Among these , we selected probe sets with expression signals that are at least 2-fold higher in the male germline sample ( s ) than in the somatic control samples . We dated human retrocopies by establishing the presence/absence of orthologous copies in the mouse , dog , opossum , and platypus genomes . For the therian genomes , we used a previously established procedure based on pairwise chained alignments of genomes ( retrieved from the UCSC genome database , http://genome . ucsc . edu/ ) for this phylogenetic dating [2] . Briefly , we first extracted the best alignments that overlap with the genomic location of retrocopies and that are greater than 15 kb ( this length ensures that the alignment also covers surrounding , non–retrocopy-derived sequences in the two species ) . We then scanned the alignments for aligned blocks that overlapped with the retrocopy . If the total length of the overlap corresponded to at least 60% of the length of the human retrocopy , the retrocopy was considered to be present in the other species . Conversely , when no such overlap was found , the retrocopy was assumed to be absent . Presence/absence of retrogenes shared between human and opossum in the platypus genome was established using a manual procedure , due to the incomplete assembly of this genome . Chained alignment data were visually inspected for the presence of significant blocks that overlap human retrocopies . Synteny of chains was validated by checking for the presence of genes in the flanks of the chains in platypus that are orthologous to the genes flanking the retrocopy in the human genome . Finally , the phylogenetic age of retrocopies was determined based on the pairwise presence/absence data obtained for all genomes; we assumed that a human retrocopy emerged on the branch before the divergence of the most-distant species in which its presence could be confirmed . We used our automatic dating procedure ( see above ) to determine the presence/absence of mouse , dog , and opossum retrocopies in other therian genomes . Retrocopies with no orthologs detected were considered to be specific to the mouse , dog , or opossum lineage , respectively . In addition , we split the set of opossum-specific retrocopies into two subsets of retrocopies that are estimated to be generally older ( parent–retrogene pairwise dS > 0 . 5 ) or younger than 90 million years ( parent–retrogene pairwise dS < 0 . 5 ) , which approximately corresponds to the human–dog lineage split time . The treshold dS = 0 . 5 is assumed to roughly correspond to a divergence of 90 million years , as human–opossum orthologs have a median dS ∼ 1 , and the two species are estimated to have diverged about 180 Mya [36] . We used standard Fisher exact and Mann-Whitney U tests . In addition , we used the resampling test described in [2] to assess the significance of the excess of parental genes on the X chromosome ( the proportion of X-linked genes was set as the null expectation ) .
|
Our sex chromosomes have profoundly differentiated since evolving from an ancestral pair of non-sex chromosomes ( autosomes ) . In this study , we first show that X chromosome–derived retrogenes ( genes that arose as duplicates of “parental” X-linked genes ) are specifically expressed during the meiotic and postmeiotic stages of spermatogenesis , thus functionally replacing their parents during , but also after , the process of male meiotic sex chromosome inactivation ( MSCI ) . We then show that the “export” of retroposed gene copies from the X chromosome started rather recently during mammalian evolution , on the eutherian ( “placental” mammal ) and marsupial lineages , respectively . This suggests that MSCI—the main driving force for this out of the X gene “movement”—originated around the separation of these two major ( therian ) mammalian lineages , approximately 180 million years ago . Given that MSCI was likely triggered as soon as the proto-X and -Y chromosomes ceased to recombine ( an event that marks the origin of these sex chromosomes ) , our data also support the recent notion that our sex chromosomes and those of other therians emerged , not in the common ancestor of all mammals , but—probably rather late—in the therian ancestor .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics",
"evolutionary",
"biology"
] |
2008
|
Chromosomal Gene Movements Reflect the Recent Origin and Biology of Therian Sex Chromosomes
|
Spliceosomal SNRNP200 is a Ski2-like RNA helicase that is associated with retinitis pigmentosa 33 ( RP33 ) . Here we found that SNRNP200 promotes viral RNA sensing and IRF3 activation through the ability of its amino-terminal Sec63 domain ( Sec63-1 ) to bind RNA and to interact with TBK1 . We show that SNRNP200 relocalizes into TBK1-containing cytoplasmic structures upon infection , in contrast to the RP33-associated S1087L mutant , which is also unable to rescue antiviral response of SNRNP200 knockdown cells . This functional rescue correlates with the Sec63-1-mediated binding of viral RNA . The hindered IFN-β production of knockdown cells was further confirmed in peripheral blood cells of RP33 patients bearing missense mutation in SNRNP200 upon infection with Sendai virus ( SeV ) . This work identifies a novel immunoregulatory role of the spliceosomal SNRNP200 helicase as an RNA sensor and TBK1 adaptor for the activation of IRF3-mediated antiviral innate response .
The innate immune system is the first line of defense against pathogens , and it relies on the recognition of pathogen-associated molecular patterns ( PAMPs ) by specific pattern recognition receptors ( PRRs ) . Upon viral infection , intracellular foreign nucleic acids are detected by specific DExD-box RNA helicases of the RIG-I-like receptor ( RLRs ) family: RIG-I ( also known as DDX58 ) , MDA5 ( also known as IFIH1 ) , and LGP2 ( also known as DHX58 ) [1] . In response to sensing viral RNA , these RLRs associate with the MAVS adaptor ( also called IPS-1 , Cardif , and VISA ) [2–5] to induce its multimerization [6 , 7] and to activate multiple kinases , including IKK , TBK1 , and IKBKE . Upon signal transduction , the activation of transcription factors such as AP-1 ( also known as ATF-2/c-jun ) , NF-κB , IRF3 , and IRF7 induces the expression of pro-inflammatory and antiviral cytokines and chemokines . Type I Interferons ( IFNs ) then trigger the activation of STAT1 , STAT2 , and IRF9 , forming a transcription factor complex known as IFN-stimulated gene factor 3 ( ISGF3 ) to ultimately induce a large number of IFN-stimulated genes ( ISGs ) . In a recent genome-wide RNAi screening that assessed virus-induced IFNB1 transcription [8] , spliceosomal factors , including the SNRNP200 RNA helicase , that positively modulate the RLR-mediated antiviral pathway were identified . Few studies have described a contribution of spliceosomal factors in pathogen-mediated immune responses , though studies have examined the effects of alternative mRNA splicing of innate immunity genes , such as DDX58 , MyD88 , and IRF3 [9–11] . Interestingly , many DExD/H-box RNA helicases were recently identified as viral nucleic acids sensor and/or mediator components of antiviral innate immunity [12 , 13] . DHX15 and DHX9 helicases were shown to interact with MAVS , following dsRNA recognition , and to activate NF-κB , IRF3 , and MAPK pathways in myeloid dendritic cells ( mDC ) [12 , 14] . An RNA helicase complex composed of DDX1 , DDX21 , and DHX36 was reported to induce type I IFN through TRIF-dependent signaling in mDC [15] . Two other helicases , DDX60 and DDX3 , were shown to bind DDX58/MDA5 and to enhance its recognition of dsRNA while also enhancing downstream type I IFN production [16 , 17] . DDX3 acts as an adaptor protein of TBK1 and IKBKE , thereby synergistically enhancing IFNB1 promoter induction [18 , 19] . Finally , DDX41 helicase is a DNA sensor that activates type I IFN via a STING-TBK1 complex [20] . In the present study , it was found that silencing SNRNP200 , a core spliceosome RNA helicase and unique member of the Ski2-like subfamily , leads to a strong decrease in the antiviral innate response by positively regulating IRF3 signaling upon Sendai virus ( SeV ) infection . In SNRNP200 knockdown ( KD ) cells , unlike the expression of wild-type ( WT ) protein , expression of the S1087L variant associated with retinitis pigmentosa 33 ( RP33 ) is unable to rescue IFNB1 transcription . The functional rescue phenotype correlates with the ability of the amino-terminal Sec63 domain ( Sec63-1 ) of SNRNP200 to bind surrogate polyinosinic-polycytidylic acid ( poly I:C ) and viral RNA . For instance , upon infection by SeV , viral RNA allows SNRNP200 to relocalize into TBK1-containing cytosolic structures . A physical interaction between SNRNP200 and TBK1 was also observed , and this interaction was mapped to the Sec63-1 domain . Finally , a significantly hindered antiviral response was demonstrated in human monocyte-derived macrophages ( MDM ) silenced for SNRNP200 and in peripheral blood cells ( PMBCs ) from RP33 patients with pathogenic missense S1087L mutation in SNRNP200 . Ultimately , this study revealed a novel immunoregulatory role of spliceosome SNRNP200 helicase in viral RNA sensing and in promoting IRF3-dependent antiviral innate immune responses .
A genome-wide gene silencing screen that assessed the transcriptional activity of the IFNB1 promoter following SeV infection was previously performed to identify novel regulators of innate immunity [8] . Six genes that encode spliceosome components that reduced IFNB1 transcription upon gene silencing were identified ( Fig 1A ) . Among these genes , one encodes an RNA helicase ( SNRNP200 ) and two ( SF3A1 and SRSF1 ) were shown to regulate innate immune responses by the alternative splicing of either Myd88 or IRF3 [10 , 11] . To further explore RNA helicases that play central roles in splicing and that often function in proofreading events in pre-mRNA splicing [21] , an RNAi mini-screen using five independent lentiviruses that express short hairpin RNA ( shRNA ) targeting most spliceosomal RNA helicases was performed ( S1A Fig ) . SNRNP200 was the only RNA helicase assigned to the Ski2-like helicase subfamily that showed a significant reduction in IFNB1 promoter-driven reporter activity . The KD of SNRNP200 was validated through the marked decrease in the mRNA and the protein levels while its specificity was validated by the absence of any off-target effects on other spliceosome gene hits ( S1B and S1C Fig ) . The depletion of SNRNP200 reduced IFN-β production at 8 hours post-infection reaching levels comparable to those obtained in DDX58 KD cells at 48 hours post-infection ( Fig 1B ) . Interestingly , while the depletion of SNRNP200 completely inhibited IFIT1 ( also known as ISG56 ) induction ( S1C Fig ) , its overexpression could not increase IFIT1 levels in neither the non-infected nor the SeV-infected cells . To further investigate SNRNP200’s contribution to the antiviral response , the viral susceptibility of SNRNP200 KD cells was monitored in a time-course experiment by following IFIT1 induction along with the production of infectious particles and the viral protein levels ( Fig 1C and 1D ) . In control HEK 293T cells transduced with a non-target sequence shRNA ( shNT ) -expressing lentivirus , SeV protein was only detectable at 24 hours post-infection which coincided with IFIT1 induction ( Fig 1C ) . In contrast , in SNRNP200 KD cells , SeV protein was readily detectable at 8 hours post-infection becoming more significant at 24 hours post-infection . However , IFIT1 induction was only detected at 48 hours post-infection . Of notable importance , SNRNP200 KD cells were observed to yield up to a 2-log increase in viral titers when compared to the control ( Fig 1D ) . Correlating with a reduced early IFNB1 induction , it was confirmed that silencing SNRNP200 also increased the replicative potential of influenza A virus ( FLUA ) and hepatitis C virus ( HCV ) in HEK 293T and Huh7 cells , respectively ( S2 Fig ) . Based on the epistasis analysis , the transcriptional activity of the IFNB1 promoter is slightly affected by ectopic expression of constitutively active IRF3 ( 5D ) [22] , while it was completely blocked in SeV infected or MAVS overexpressing cells transduced at a high multiplicity of infection ( MOI of 20 ) with lentiviral-expressing shRNA ( Fig 1E ) . Similar results were obtained in A549 cells ( S3A and S3B Fig ) . Interestingly , upon SeV infection and in contrast to IRF3 overexpression , the ectopic expression of IRF3 ( 5D ) could rescue the induction of IFIT1 ( Fig 1F ) , suggesting a role for SNRNP200 in IRF3 activation . The effect of SNRNP200 KD in NF-κB-dependent transcription was then investigated using a reporter assay ( p2xNF-κB_fLUC ) in HEK 293T cells . It was found that SNRNP200 KD cells display no attenuation of poly ( I:C ) - , MAVS- , TBK1- , or p65-mediated activation of the NF-κB promoter ( S4A Fig ) . In contrast , in these cells , there was a significant inhibition of SeV- , poly ( I:C ) - , TBK1- , and IFN-α-mediated activation of the ISG56 promoter ( S4B Fig ) . It was confirmed that SNRNP200 silencing does not affect NF-κB-dependent transcription in SeV-infected A549 cells through the quantification of TNF , NFKBIA , and TNFAIP3 mRNA levels using qRT-PCR ( S3C Fig ) . Interestingly , neither the TRIF nor the cGAS/STING pathways are affected in SNRNP200 KD cells whereas the RLR pathway , which converge to the TBK1-mediated phosphorylation of their respective adaptors ( TRIF , STING , and MAVS ) to recruit IRF3 and to license IRF3 for phosphorylation [23] . These data suggest that , upon RNA virus infection , SNRNP200 may function between MAVS signal transduction and TBK1-mediated IRF3 licensing . These observations led to exploring a specific regulatory role of SNRNP200 , a core component of U4/U6-U5 small nuclear RNA ( snRNA ) [24] , in the downstream activation of IRF3 , of production of IFNB1 and ultimately of an optimal antiviral response . To understand the manner in which SNRNP200 , upon viral infection , contributes to IRF3-mediated IFNB1 production , the effect of SNRNP200 silencing on the expression of established members of the RLR pathway was evaluated using a western blot analysis ( S5A Fig ) . First , a decreased protein expression of IRF3 in SNRNP200 KD cells was observed . This correlated with the blockage of the SeV-mediated induction of IFIT1 , DDX58 , and IFIH1 proteins . The decreased IRF3 protein levels were further confirmed at the mRNA level by qRT-PCR , paralleling the reduced mRNA levels of SNRNP200 and its effector genes ( IFNB1 , IFIT1 , DDX58 , and IFIH1 ) ( S5B Fig ) . While the residual IRF3 protein levels of KD cells are sufficient for the activation of the cGAS/STING pathway ( S4B Fig ) , a complete inhibition of IRF3 phosphorylation at serine 386 following SeV infection ( Fig 2A , see IRF3-p386 ) was observed . This suggested a specific contribution of SNRNP200 during IRF3-activation-mediated IFNB1 production . A weak decrease of the basal protein expression levels of DDX58 in SNRNP200 KD cells was also observed ( S5A Fig ) . The mRNA levels , however , were comparable to the control shNT cells ( S5B Fig ) , suggesting that SNRNP200 enhances the RLR-mediated antiviral signaling potential of DDX58 at the basal level . In contrast , the protein expression of MAVS , TBK1 , IKBKE , RELA ( p65 ) , and TRAF3 , which all contribute to the signal propagation for IFNB1 induction , remained unchanged in all conditions ( S5A Fig ) . Similar observations were made for the protein expression of the housekeeping genes ACTIN , TUBULIN , and GAPDH ( S5A Fig ) . To better evaluate the outcome of the reduced basal protein levels of DDX58 and IRF3 on IFNB1 production , ectopic expression of DDX58 and IRF3 in SNRNP200 KD cells was used in an attempt to restore antiviral response . Surprisingly , overexpression of DDX58 and IRF3 neither alone nor in combination could restore SeV-mediated IFIT1 induction ( Fig 2B and S6A and S6B Fig ) . Furthermore , ectopic expression of neither DDX58 nor IRF3 restored IFNB1 promoter reporter activity or IFN-β cytokine production upon SeV infection ( Fig 2D and 2E ) . This was in stark contrast to the almost complete rescue achieved by the ectopic expression of IRF3 ( 5D ) ( Fig 2D and 2E ) . The phosphorylation of IRF3 at serine 386 ( IRF3-p386 ) was also noted as a key event for IRF3 activation and , as such , the proportion of IRF3-p386 in relation to total IRF3 was investigated . The quantitative ratios of IRF3-p386 to total IRF3 in context of endogenous or overexpressed DDX58 , IRF3 , IRF3 ( 5D ) , and cGAS/STING were determined ( Fig 2B ) . When comparing control shNT-treated with SNRNP200 KD cells in the context of SeV-mediated infection , regardless of whether or not DDX58 or IRF3 was overexpressed , a significant reduction in the IRF3-p386/IRF3 ratios ( from 0 . 6–0 . 9 to 0 . 1–0 . 2 ) was observed ( Fig 2B ) . This establishes the requirement of SNRNP200 for downstream IRF3 activation independent of its effects on protein expression . Furthermore , ectopic expression of IRF3 ( 5D ) in SNRNP200 KD cells yielded IRF3-p386/IRF3 ratios comparable to those of the control shNT-treated cells ( 0 . 8 vs 0 . 6–1 . 3 ) correlating with the almost complete restoration of SeV-mediated IFNB1 production . The effect of SNRNP200 KD on the DNA sensing arm of the antiviral response downstream of cGAS/STING was also investigated . It was shown that SNRNP200 is dispensable for cGAS/STING-mediated IFIT1 induction , IFN-β production , and IFNB1 promoter activity ( Fig 2B , 2D and 2E ) , further implying a specific role for SNRNP200 in the RLR-mediated IRF3 signaling pathway upon RNA virus infection . Interestingly , the higher IRF3-p386/IRF3 ratios ( 4 . 5 ) in SNRNP200 KD cGAS/STING expressing cells versus control cells ( 2 . 3 ) largely reflects a significant increase in IRF3 activation and hence in IFIT1 induction ( Fig 2B ) . This suggests that SNRNP200 potentially competes with the STING adaptor during TBK1-mediated IRF3 phosphorylation . Although IRF3 expression slightly increased IFN-β secretion and the IFNB1 promoter activity when activation of the cGAS/STING pathway and SeV infection were combined in SNRNP200 KD cells ( Fig 2D and 2E ) , similar IFNB1 induction was observed in uninfected cells ( Fig 2C ) , demonstrating that IRF3 protein levels in SNRNP200 KD cells have little functional consequence on the cytosolic DNA sensing pathway . Finally , the IRF3 mRNA splice junctions and the presence of exons were investigated to explain the reduced mRNA and protein levels of IRF3 . No splicing variants were observed excluding an alternative splicing regulation of IRF3 and supporting a reduction in the efficiency of pre-mRNAs splicing to explain the phenotype of SNRNP200 KD cells ( S7 Fig ) . Interestingly , the induction of IFIT1 , DDX58 , and IFIH1 was also inhibited upon treatment of SNRNP200 KD cells with IFN-α ( Fig 2A ) . Furthermore , similar levels of IFNα/β receptor alpha chain ( IFNAR1 ) , of STAT1 , and of phosphorylation at tyrosine 701 ( STAT1pY701 ) were observed between control and SNRNP200 KD cells following stimulation with IFN-α . This suggests the involvement of SNRNP200 at a later stage of type I IFN signaling downstream of STAT1 phosphorylation . The negative effect of SNRNP200 KD on IFN-α signaling was also demonstrated by the reduced mRNA levels of IFN inducible genes IFIT1 , DDX58 , and IFIH1 in A549 cells ( S3C Fig ) . To comprehensively understand the effect of SNRNP200 KD on type I IFN production ( early ) and signaling ( late ) , expression profiling studies of non-stimulated ( NS ) , SeV-infected and IFN-α-treated SNRNP200 KD cells versus control shNT HEK 293T cells were performed to assess differential gene expression ( Fig 3 and S8 Fig ) . The effect of SNRNP200 KD on basal gene expression of NS cells was characterized; 2 , 880 altered transcripts ( cutoff of 1 , 5 log2 fold induction ) that are primarily associated with immune system functions and cell cycle regulation were found based on a Reactome pathway enrichment analysis ( S8A Fig ) . A list of transcriptionally altered genes by SeV infection or by IFN-α stimulation in control shNT cells was then established . Using a stringent significance cutoff ( p ≤0 , 001 ) , 52 genes altered by SeV infection and 55 genes altered by IFN-α stimulation were found to be transcriptionally affected by SNRNP200 ( Fig 3A ) . Within these subsets of genes , 13 were in common reflecting the expected overlap of the early ( SeV-mediated IFNB1 production ) and the late ( type I IFN signaling ) arms of the antiviral response . Within the subset of commonly affected genes , all showed altered expression upon SNRNP200 silencing with a mean difference of 3 , 8 log2 fold change when compared to the shNT control cells ( Fig 3B ) . On the other hand , the two subsets of 39 SeV-specific and 42 IFN-α specific genes were hindered by SNRNP200 silencing by a mean difference of 2 , 6 and 2 , 5 log2 fold change , respectively , when compared to the shNT control cells ( Fig 3C and 3D ) . This demonstrated that SNRNP200 plays a distinct role in the early and the late antiviral response pathways . The relationship between the affected gene subsets was assessed by resolving their interaction and functional alignment networks . The assessment showed that the top GO term for genes exclusively affected by SeV is “response to virus” , and the top GO term for IFN-α is “response to type I interferon” . Furthermore , it showed that SeV-specific genes affected by silencing SNRNP200 are IRF3-dependent ( IFNB1 , IL29 , and BIRC6 ) and that several IFN-α-specific genes are JAK-STAT1-dependent ( ACVR1C , CIQA , IFIT5 , and OAS1 ) . The latter might be explained by the reduction of IFN-induced IRF9 and STAT1 mRNA levels encoding key transcription factors of ISGF3 that mediates signaling of type I and III IFNs ( S2 Table ) . The molecular signature of differential gene expression strengthens the observation that SNRNP200 silencing hinders IRF3-dependent gene induction , leading to a general weakening of the RLR signaling pathway , and further suggests that SNRNP200 plays a distinct regulatory role in type I IFN signaling . The results suggest that SNRNP200 specifically regulates IRF3 activation upon RNA virus infection to promote IFNB1 induction and IFN effector responses , and thus demonstrates its importance in controlling viral infections . To examine the manner in which SNRNP200 directly contributes to IRF3-mediated IFNB1 activation upon SeV infection , a series of recombinant SNRNP200-truncated mutants were tested for their ability to rescue IFNB1 reporter activity and ISG expression in SNRNP200 KD cells ( Fig 4 and S9 Fig ) . None of the truncated mutants could induce an antiviral response in SNRNP200 KD cells ( with the exception of a weak IFNB1 activation by expression of a D1-D3 construct ) . Indeed , the deletion of the C-terminal Sec63 domain ( Sec63-2 ) alone , which was reported in yeast ( Sec63-2 deleted Brr2 protein ) to reduce ATPase/helicase activity and splicing [25] , completely abolished the activation of IFNB1 promoter-driven reporter activity and the induction of IFIT1 and DDX58 upon SeV infection . To further explore a dual regulatory role in splicing and RNA-mediated antiviral responses , the described SNRNP200 heterozygous mutations associated with the autosomal dominant RP33 disease were considered [26–28] . In particular , the SNRNP200 S1087L and R681C variants located within the Sec63-1 homology domain and the N-terminal RecA-like ATPase/helicase domains , respectively , were investigated ( Fig 4A ) . It was first demonstrated that the ectopic expression of RNAi-resistant WT SNRNP200 rescues the SeV-mediated IFN-β secretion and the IFNB1-driven reporter activity in KD cells , further validating the specificity and minimal off-target effects of shSNRNP200 and its associated immunoregulatory phenotype ( Fig 4 ) . Surprisingly , expression of the SNRNP200 S1087L mutant completely eliminated the ability to rescue IFNB1 activation ( Fig 4B and 4C ) . Similar results were obtained using qRT-PCR as the rescue of endogenous IFNB1 mRNA levels was achieved only by the expression of WT SNRNP200 ( S10 Fig ) . Concordantly , WT SNRNP200 , but not the S1087L mutant , restores IRF3 protein levels , and more importantly , restores the phosphorylation of IRF3 at serine 386 as well as the inducible levels of DDX58 and IFIT1 upon SeV infection ( Fig 4D ) . WT SNRNP200 , but not the S1087L mutant , also restores IFN-α-dependent DDX58 and IFIT1 induction . It was also determined that expression of R681C variant only slightly rescues IFNB1 promoter-driven reporter activity and IFN-β secretion ( Fig 4B and 4C ) . Interestingly , while investigating a mutation within the ATP binding motif , it was found that the ectopic expression of a SNRNP200 C502A variant elicited an IFNB1 response independent of viral infection ( S11 Fig ) in line with the recently reported natural gain-of-function of DDX58 and IFIH1 ATPase-deficient variants [29] . The constitutive induction of IFNB1 with expression of SNRNP200 C502A is further enhanced upon SeV infection to levels similar to the WT enzyme ( Fig 4B and 4C ) , thereby suggesting the requirement of a functional SNRNP200 ATPase in conferring specificity to viral RNA and in preventing signaling through the recognition of self-RNA . Thus , the data firmly establishes a critical role of the Sec63-1 domain to promote virus-mediated IFNB1 production and further suggests a contribution of the N-terminal ATPase/helicase domain in sensing viral RNA . DExD/H-box helicases , such as RIG-I , are engaged in antiviral innate immunity because they detect viral nucleic acids and prevent the recognition of self-RNA through ATP hydrolysis ( 29 ) . As the Sec63-1 containing S1087L mutation was reported to diminish binding to RNA duplex and to reduce RNA-stimulated ATPase/helicase activity without any discernible effect on the folding of SNRNP200 ( 27 ) , it was hypothesized that this natural loss-of-function mutation abolishes the recognition of viral RNA for IFNB1 induction . To determine whether or not the S1087L variant impaired the binding of the immunostimulatory RNA in SeV-infected cells , the in vitro ability of exogenously expressed SNRNP200 to bind biotinylated polyinosinic-polycytidylic acid ( poly ( I:C ) ) was measured using an RNA pull-down and western blot analysis of bead-bound protein fractions ( Fig 5 ) . It was shown that FLAG-WT SNRNP200 binds poly ( I:C ) , which is used as a viral double-stranded RNA ( dsRNA ) surrogate , only in SeV-infected cell extracts ( Fig 5A ) . Furthermore , a complete loss of poly ( I:C ) binding by the FLAG-SNRNP200 S1087L variant was observed . Interestingly , the FLAG-Sec63-1 domain , but not the FLAG-Sec63-2 , is sufficient to bind poly ( I:C ) ( Fig 5B ) . These observations were confirmed with biotinylated HCV genomic RNA . As expected , WT SNRNP200 and the Sec63-1 domain , but neither the S1087L variant nor the Sec63-2 domain , were able to successfully pull-down HCV RNA ( Fig 5C ) . To provide insight into the interaction of SNRNP200 and immunostimulatory RNA molecules , SNRNP200’s ability to bind a synthetic 5’-triphosphate ( 5’ppp ) and a double-stranded stretch of RNA using the full-length HCV genome , produced in vitro by transcription with a T7 polymerase , was investigated ( Fig 5D ) . A comparable binding of FLAG-WT SNRNP200 to the untreated and to the calf-intestine alkaline phosphatase ( CIAP ) -treated blunt-ended HCV RNA was observed . This comparable binding suggests that the 5’ppp moiety is not essential for the recognition of viral dsRNA . FLAG-SNRNP200 does not bind dsDNA molecules unlike the FLAG-cGAS control which does [30] . This is reflected by the lack of pull-down by FLAG-SNRNP200 with biotinylated polydeoxyadenylic acid-polythymidylic acid ( poly ( dA:dT ) ) and polydeoxyguanylic acid-polydeoxycytidylic acid ( poly ( dG:dC ) ) homopolymer molecules ( Fig 5D ) . To assess the requirement of a protein complex with DDX58 , MAVS , or TBK1 for binding HCV RNA , expressions of these proteins both individually and together were silenced , and RNA pull-down assays were performed to detect SNRNP200 ( Fig 5E ) . It was determined that SNRNP200 binds HCV RNA regardless , ruling out a contribution of these proteins in its ability to recognize viral RNA . Finally , FLAG-tagged WT SNRNP200 and S1087L variant were immunoprecipitated upon SeV infection , and the co-purified RNA molecules were analyzed using qRT-PCR in SNRNP200 KD and in control shNT cells ( Fig 5F ) . Increased amounts of actin mRNA for both immunoprecipitated proteins were found compared to the eYFP control ( normalized to RNA levels of cell lysates ) . A significant enrichment of SeV RNA , which is more important in SNRNP200 KD cells than in shNT control cells that express the endogenous untagged protein , was observed upon immunoprecipitation of FLAG-WT SNRNP200 , demonstrating a direct binding to viral genomes . The amount of SeV RNA recovered with the WT was almost 10- to 20-fold higher than with the S1087L variant in KD cells ( and 3-fold in shNT cells ) , reflecting an altered RNA binding ability of the mutant . Despite the weak binding of SeV RNA by S1087L which is possibly due to its N-terminal RecA domains , the loss-of-function in IFNB1 induction reveals that this interaction is not biologically active . The data demonstrate that the Sec63-1 domain of SNRNP200 is a major determinant of viral RNA recognition and consequently of SNRNP200’s ability to activate antiviral innate immune responses . To better define a specific immunoregulatory role of SNRNP200 , binding partners were identified by screening proteins of the antiviral signaling pathways upon immunoprecipitation of FLAG-tagged SNRNP200 . This method successfully allowed the detection of a constitutive interaction between SNRNP200 and the ubiquitously expressed kinase TBK1 . This interaction was also detected when using the SNRNP200 S1087L mutant ( Fig 6A ) . The ability of various SNRNP200-truncated mutants ( Fig 4A ) to bind TBK1 ( Fig 6B ) was then assessed . A mutagenesis analysis showed that the Sec63-1 domain of SNRNP200 is required and sufficient for TBK1 interaction ( Fig 6C ) , which is similar to the observation for RNA binding ( Fig 5B and 5C ) . Both Sec63 homology domains of SNRNP200 contain a helical bundle ( HB ) and immunoglobulin-like ( IG ) sub-domains separated by a helix loop helix ( HLH ) motif . To more accurately map the TBK1 binding domain , the sub-domains of Sec63-1 were expressed separately . A weak interaction with the HLH-IG sub-domain was observed , suggesting its contribution to the binding of TBK1 ( Fig 6C ) . It was also demonstrated that the C-terminal Sec63-2 domain could not bind TBK1 , which corroborates the detected interaction of the N-terminal truncated D1-3 , D1-4 , and D1-5 mutants with TBK1 ( Fig 6B and 6C ) . In reciprocal experiments , immunoprecipitation of FLAG-tagged TBK1 confirmed the interaction with ectopically expressed SNRNP200 in uninfected and in SeV-infected cells ( Fig 6D ) . In addition , the kinase-dead TBK1 mutant ( K38A ) was still shown to interact with SNRNP200 , demonstrating that this interaction is not dependent on TBK1 activity ( Fig 6D ) . Finally , the interaction was further confirmed in A549 cells by the co-immunoprecipitation of endogenous SNRNP200 and TBK1 proteins ( S12 Fig ) . To further assess the interaction of SNRNP200 and TBK1 , their intracellular localization was investigated by examining confocal fluorescence microscope images of HEK 293T cells and of HeLa cells in response to SeV infection ( Fig 7 and S13 Fig ) . It was observed that FLAG-SNRNP200 ( HEK 293T cells ) and endogenous SNRNP200 ( Hela cells ) are localized to the nucleus and cytoplasm with a diffuse staining prior to stimulation . Upon viral infection , a subcellular fraction of SNRNP200 relocalizes with TBK1 into perinuclear cytoplasmic speckles ( Fig 7A and S13B Fig ) . SNRNP200 and TBK1 colocalization can be easily observed in the 3D-stack and lateral view of infected cells ( Fig 7B and 7C ) . Unlike WT SNRNP200 , the staining of the FLAG-SNRNP200 S1087L mutant shows neither relocalization of the protein nor colocalization with TBK1 into these cytoplasmic speckles upon infection ( Fig 7A and S13A Fig ) , correlating with its lack of RNA binding ( Fig 5 ) . Thus , the data suggest that viral RNA recognition by the Sec63-1 domain is responsible for the relocalization of SNRNP200 to perinuclear cytoplasmic speckles , and that SNRNP200 possibly functions as a novel adaptor via its interaction with TBK1 to promote IRF3 phosphorylation and antiviral innate responses . The regulation of antiviral responses by SNRNP200 was further investigated in immune cells using primary cultures of purified human monocyte-derived macrophages ( MDM ) . It was found that SeV infection leads to an increase in the immunodetection of SNRNP200 without affecting mRNA levels ( Fig 8A ) , as observed in SeV-infected and in IFN-α-treated HEK 293T cells ( S14 Fig ) . More importantly , the silencing of SNRNP200 in MDM decreases the induction of IFIH1 and IFIT1 , and completely blocks IRF3 Ser386 phosphorylation within 3 hours post-infection ( Fig 8B ) . Kinetic studies on IFN-β production have further demonstrated a complete blockage of its secretion at 3 hours post-infection ( Fig 8C ) . In contrast to the unchanged TNFα mRNA levels , a decrease in IFNB1 mRNA was observed , which correlates with the reduced SNRNP200 mRNA at 1 hour post-infection in MDM ( Fig 8D , 8E and 8F ) . Interestingly , the duration of SNRNP200 gene silencing is not sufficient to affect the steady-state levels of IRF3 protein , though it completely inhibits its phosphorylation . In addition , SNRNP200 KD increased SeV protein levels , as observed in HEK 293T cells ( Figs 1C and 8B ) . These results confirm a regulatory role of SNRNP200 in the IRF3-mediated antiviral response in human macrophages . RP is an inherited degenerative eye disease that causes severe vision impairment and blindness due to mutations in several core spliceosomal proteins . The antiviral responses of peripheral blood mononuclear cells ( PBMCs ) of RP33 patients that are genotyped for a particular monoallelic mutation in SNRNP200 were characterized: p . S1087L- c . 3260C>T in the Sec63-1 domain and p . R681C c . 2122G>A in the N-terminal helicase domain ( see S1 Table for donor information ) . Interestingly , all RP33 patients showed a complete blockage of IFN-β cytokine production at 3 hours post-infection with a significant two-fold reduction in IFN-β secretion at 7 hours ( Fig 9A and 9B ) . The decreased IFN-β production was corroborated by the reduction of virus-induced IRF3-dependent IFNB1 and IFIT1 mRNA . On the other hand , NF-κB-dependent TNF mRNA levels were not significantly affected ( Fig 9C , 9D and 9E ) . The IRF3 mRNA levels determined by qRT-PCR showed no difference between healthy donors ( HD ) and RP33 patients ( Fig 9F ) . Finally , a cytokine 41-plex assay performed on supernatants of infected PBMCs from HD and RP33 patients showed a significant decrease in IFN-α2 , but showed similar cytokine/chemokine levels of RANTES , IL6 , CXCL10 , and IL1B ( S15 Fig ) . The defective antiviral response of PBMCs from RP33 patients demonstrates that SNRNP200 plays a crucial role in regulating the IRF3-dependent pathway of IFNB1 production and does so without altering the NF-κB-dependent inflammatory pathway .
SNRNP200 RNA helicase is ubiquitously expressed in cells and is a core component of the spliceosome . Its plays a key role in unwinding U4/U6 snRNA to form a highly structured RNA interaction network among the U2 , U6 , and U5 snRNA and the pre-mRNA required for activation of the spliceosome [31 , 32] . Despite this critical function for pre-RNA splicing , to the authors’ knowledge , no data has been found that suggests a role of SNRNP200 in host defense . Furthermore , few studies have described a contribution of spliceosomal proteins for innate immunity . Two spliceosomal proteins ( SRSF1 and SF3A1 ) have been identified from the genome-wide gene silencing screening ( Fig 1A ) that have previously been reported to be involved in the generation of alternative splice variants of important innate immune regulators . Depletion of SRSF1 in human A549 lung cancer cells reduces IFN-β through the expression of alternative IRF3 spliced variants [10] , while SF3A1 silencing leads to a decreased induction of pro-inflammatory cytokines by promoting an alternative splice form of MyD88 [11] . Based on this study , evidence of a novel role of the spliceosomal SNRNP200 RNA helicase in the regulation of IRF3-mediated antiviral response upon the RNA virus infection in human cells is presented: 1 . SNRNP200 KD cells infected with SeV and FLUA show higher virus titers and viral proteins ( Fig 1 and S2 Fig ) , suggesting that SNRNP200 is involved in host defense mechanisms; 2 . SNRNP200 KD cells reduce virus-mediated IFN-β production ( Figs 1B , 2E , 4C and 8C ) ; 3 . Epistasis studies suggest a role for SNRNP200 within the antiviral response during IRF3 activation ( Figs 1E , 1F , 2B , 2D and 2E ) ; 4 . SNRNP200 solely regulates the RLR pathway and does not affect the TRIF or the cGAS/STING pathways when activating IFNs production ( Fig 2B , 2D , 2E and S4 Fig ) ; 5 . SNRNP200 requires a competent Sec63-1 domain and functional ATPase/helicase activity to promote IRF3-dependent IFNB1 activation ( Fig 4 ) ; 6 . The SNRNP200 Sec63-1 domain binds immunostimulatory RNA molecules ( Fig 5 ) ; 7 . SNRNP200 interacts with endogenous TBK1 through its Sec63-1 domain ( Fig 6 ) ; and 8 . PBMCs of RP33 patients ( who have one allele carrying the dominant S1087L or R681C mutation ) showed a reduction of IFN-β secretion when challenged with SeV ( Fig 9 ) . Thus , SNRNP200 , the only Ski2-like RNA helicase involved in pre-mRNA splicing , regulates IRF3-dependent IFNB1 production upon RNA virus infection through the recognition of viral RNA promoting the phosphorylation of IRF3 and possibly functions as an adaptor protein through its constitutive interaction with TBK1 . The results reveal a unique molecular mechanism regarding the way SNRNP200 regulates the antiviral response . Mechanistically , it was found that , upon viral infection , SNRNP200 relocates to some undefined cytoplasmic structures were it is able to directly sense viral RNA . This activation results in a striking virus-induced association of SNRNP200 and TBK1 into larger order punctate perinuclear structures . This mobilization of SNRNP200 and TBK1 promotes IRF3 phosphorylation that , in turn , translocate to the nucleus to transactivate the IFN-β promoter and induce the production of ISGs to fully engage antiviral immunity ( for a model , see Fig 10 ) . The major mechanism by which SNRNP200 functions is as a spliceosomal helicase that unwinds the U4/U6 snRNA , providing key remodeling activity for spliceosome catalytic activation , and thus it regulates the expression of a large and disparate group of genes associated with the cell cycle [33] . Indeed , the transcriptional profiles of HEK 293T cells indicate a large group of differentially expressed genes upon SNRNP200 silencing that are associated with the immune system and the cell cycle , as shown by the Reactome Pathway Enrichment Analysis ( S8A Fig ) . Nevertheless , among the total SeV- and IFN-induced genes of control shNT cells , silencing SNRNP200 KD affects specific SeV- and IFN-inducible genes as well as common genes by more than 1 , 5 log2 fold induction ( see Venn diagrams of S8 Fig ) . The analysis of the gene network confirmed an enrichment for innate immunity gene function ( Fig 3E ) and revealed that altered genes ( 52 for SeV , 55 for IFN , and of which 13 are common to both ) are highly connected to IRF3 and IFNB1 by a molecular signature , which indicates that SNRNP200 silencing hinders IRF3-dependent gene induction . One possible mechanism is that SNRNP200 affects the pathway at a transcriptional level , as first revealed by the observation that SNRNP200 alters expression of the key transcription factor IRF3 , which is essential for IFNB1 transcription . The decrease of IRF3 mRNA and protein levels correlates with the reduced SNRNP200 mRNA and protein levels as well as with the reduced expression of effector genes upon infection in KD cells ( S5 Fig ) ; however , the experiments did not identify splicing variants to explain the reduced IRF3 protein levels ( S7 Fig ) , ruling out an alternative splicing regulation of IRF3 mRNA for SNRNP200 depleted cells . While the reduced expression of IRF3 and DDX58 proteins ( S5A Fig ) may contribute to the phenotype , considerable evidence suggests that it is not the primary mechanism responsible for the decreased IFNB1 activation of SNRNP200 KD cells . First , ectopic expression of IRF3 and/or DDX58 fails to restore the virus-mediated IFNB1 production or IFIT1 expression in SNRNP200 KD cells , while expression of the constitutively active IRF3 ( 5D ) fully rescues the antiviral response ( Figs 1F , 2B , 2D , 2E and S6 Fig ) . Second , activation of the cGAS/STING pathway involved in the recognition of cytosolic DNA is not affected by SNRNP200 KD , despite the reduced IRF3 protein levels ( Fig 2C ) which slightly restricts IFNB1 production when cGAS/STING activation and SeV infection are combined ( Fig 2D and 2E , see cGAS+STING versus cGAS+STING+IRF3 ) . Indeed , the full activation of the cGAS/STING/TBK1/IRF3 pathway in SNRNP200 KD cells further supports a specific role for SNRNP200 upon activation of the RLR/MAVS/TBK1/IRF3 pathway by RNA virus infection . Finally , the silencing of SNRNP200 completely blocks IRF3 Ser386 phosphorylation in MDM , even when IRF3 protein levels are similar to control cells , resulting in the blockage of IFN-β secretion at 3 hours post-infection ( Fig 8C ) . On the other hand , gene profiling data clearly illustrated that SNRNP200 affects the expression of a large group of genes associated with the immune system and the cell cycle of unstimulated cells ( S8A Fig ) . Furthermore , it was observed that silencing SNRNP200 reduces type I IFN signaling downstream of STAT1 phosphorylation through a molecular mechanism that requires further investigation . Thus , it cannot be ruled out that the perturbation of pre-mRNA processing leading to impaired expression of immune-related genes possibly contributes to the reduced antiviral response of SNRNP200 KD cells via a global transcriptional regulatory role . Nonetheless , the abrogated phosphorylation of IRF3 ( Fig 2A and 2B for IRF3p386/IRF3 ratios and 8B ) provides the first mechanistic insight to explain the phenotype of SNRNP200 depleted cells . Based on these results , a direct role of SNRNP200 was considered for the activation of IRF3 that leads to IFNB1 production . To determine the regulatory function of SNRNP200 in IRF3 phosphorylation , its ability to interact with known members of the RLR signaling pathway was evaluated and its interaction with TBK1 was carefully characterized . The TBK1 binding site to the Sec63-1 domain ( Fig 6C ) was mapped , and a colocalization of SNRNP200 and TBK1 in cytoplasmic speckles triggered by SeV infection was observed ( Fig 7 and S12 Fig ) . This indicates the involvement of a cytoplasmic SNRNP200-TBK1 protein complex in the modulation of IRF3 phosphorylation required for IRF3’s dimerization and nuclear translocation for IFNB1 transcription and for ISGs production [34] in a mechanism similar to that described for DDX3 helicase [19] . The hypothesis that SNRNP200 directly functions as a sensor of viral RNA was tested using RNA pull-down experiments . It was demonstrated that SNRNP200 , and more specifically its Sec63-1 domain , binds viral surrogate poly ( I:C ) and HCV genomic RNA ( Fig 5B and 5C ) . The RNA-binding ability of SNRNP200 mainly involves the recognition of dsRNA , as seen with poly ( I:C ) , while the presence of the 5’triphosphate moiety as well as the expression of DDX58 , MAVS , and TBK1 proteins are not required for binding dsRNA molecules ( Fig 5D and 5E ) . As expected , SNRNP200 RNA helicase does not bind dsDNA ( Fig 5D ) , which supports the observation that SNRNP200 KD does not affect ISG56 promoter-driven activity upon activation by the cGAS/STING pathway ( S4B Fig ) . Surprisingly , in contrast to DDX58 , which binds poly ( I:C ) in the absence of viral infection ( S16 Fig ) , the binding of SNRNP200 to dsRNA was only observed following SeV infection . This observation cannot be explained , especially because SNRNP200 binds SeV RNA with high affinity , as reflected by the significant enrichment of the protein pull-down over cell lysates of SNRNP200 KD cells ( with low SeV RNA levels detected due to the rescue of antiviral response by WT expression ) ( Fig 5F ) . It was further demonstrated that expression of the naturally occurring SNRNP200 S1087L variant located in the Sec63-1 domain , which is associated with RP33 [OMIM:610359] , is unable to bind dsRNA and SeV RNA ( Fig 5A and 5F ) , does not relocalize with TBK1 upon SeV infection ( Fig 7A ) , and cannot restore the antiviral response in SNRNP200 KD cells ( Fig 4D ) . Thus , the results suggest that a pre-activation of SNRNP200 upon recognition of viral RNA allows its relocalization to perinuclear cytoplasmic speckles with TBK1 . Finally , the study of ATPase/helicase-deficient SNRNP200 variants that affect the antiviral response and the discovery of a mutant ( C502A ) that elicits an IFNB1 response independent of viral infection and fully rescues IFN-β in SNRNP200 KD cells ( Fig 4C and S11 Fig ) further support the SNRNP200 ATPase/helicase function in conferring specificity to viral RNA and preventing signaling through the recognition of self-RNA , as recently reported for the natural gain-of-function DDX58 and IFIH1 ATPase-deficient variants [29] . These data demonstrate a direct regulatory role of SNRNP200 via its Sec63-1 domain and its ATPase/helicase function in the recognition of viral RNA and in the relocalization into cytoplasmic speckles with TBK1 to promote IRF3 activation and the antiviral response . To ascertain the regulatory role of SNRNP200 in immune cells , depletion experiments were carried out with purified MDM isolated from PBMCs of healthy donors . As observed with the cell lines , SNRNP200 KD led to a hindered IFN-β production by blocking IRF3 phosphorylation without altering IRF3 expression in MDM ( Fig 8B and 8C ) . There was also an increased viral susceptibility , illustrating a relevant role of SNRNP200 in the antiviral response of human macrophages . Furthermore , the loss-of-function mutations of the human SNRNP200 gene that cause autosomal-dominant RP33 were exploited . RP is a rare inherited disease of retinal dystrophies with an incidence of one in 3 , 000–4 , 000 , of which 1 . 6% bear mutations in the SNRNP200 gene [35] . The investigated S1087L is a disease-associated mutation with complete penetrance in the RP33-linked family [36–38] . With access to PBMCs of three RP33 patients who had one allele carrying the dominant S1087L or R681C mutation , a decreased IRF3-dependent antiviral response was confirmed , when challenged with SeV , by the specific reduction of IFN-β and IFN-α2 cytokine secretion , without any effect on other tested cytokines ( Fig 9 and S15 Fig ) . Thus , further evidence with human cells of patients with RP33 showed that SNRNP200 positively regulates antiviral responses independently from its primary core function in pre-mRNA splicing . The recent resolution of the SNRNP200 structure ( amino acids 395 to 2129 ) provides the spatial relation between duplicated N-terminal and C-terminal cassettes that both contain a RecA1-RecA2 DEAD-box helicase domain and a Sec63 homology domain [27] . Both cassettes are required for optimal helicase activity and splicing function , but only the N-terminal cassette is reported to be catalytically active [27] . The 3D structure of the Sec63-1 homology region resembles the structure of isolated C-terminal Sec63 units in yeast and in human enzymes [25 , 39] . The 3D structure showed that the serine 1087 is located on a long scaffolding helix ( referred as a ratchet-helix ) , within the HB domain . The HB domain forms the top , while the RecA domains form the bottom , of a central tunnel that act as a strand separation device for RNA . The testing of a leucine at position 1087 exhibited decreased RNA binding and reduced ATPase and helicase activities compared to the WT [27] . This mutation is believed to decrease spliceosome activation and to explain its linkage to RP33 [25] . In this study , the S1087L mutant completely abolished the recognition of poly ( I:C ) and of viral RNA molecules and was completely unable to relocalize into TBK1-containing cytoplasmic speckles , which could explain the hindered antiviral response . Further proteomic studies should provide a more comprehensive explanation for this mechanism along with the identification of interaction partners that mediate the cytoplasmic relocalization of a SNRNP200-TBK1 complex upon infection . Structurally , SNRNP200 also decisively differs from other spliceosomal helicases , as it belongs to the Ski2-like subfamily , which is a small family within the superfamily 2 helicases ( founder member: yeast Ski2 ) involved in a variety of RNA processing and degradation events [40] . In addition to SNRNP200 , which exhibits the ATP-dependent unwinding activity of U4/U6 RNA duplexes during pre-mRNA splicing [22 , 41 , 42] , other helicases , such as SKIV2L and DDX60 , promote exosome-mediated RNA decay [16 , 43] . SKIV2L is involved in the elimination of incompletely spliced RNA transcripts upon stress responses , which triggers a sterile RNA-activated antiviral innate response due to its inhibition [43] . Indeed , SKIV2L-deficient patients exhibit a constitutive type I IFN signature in their peripheral blood that results in a human auto-immune disorder . Moreover , the yeast Ski2 and Ski2-like helicase 1 ( Slh1 ) have been implicated in antiviral defense by blocking the translation of RNA lacking a 3' poly ( A ) structure [44 , 45] . Further studies are required to assess the potential role of SNRNP200 for the recognition of host RNA molecules upon stress responses . Finally , to the authors’ knowledge , there has been no association between RP33 pathology and immune disorders . The findings in the PBMC of patients with a monoallelic point mutation in SNRNP200 established a deregulation of innate immunity , which may affect cell viability in different retinal disorders , as these cells and neural cells are usually non-proliferative and long-lived . Although RP33 is a rare event , it may be clinically relevant in identifying the mechanism of disease onset at a molecular level in relation to a deregulation of the innate response and the control of cell viability . Indeed , mutation E50K of optineurin , a critical regulator of antiviral signaling [46] , promotes interaction with TBK1 and is associated with familial primary open-angle glaucoma [47] . The dysfunction of optineurin and TBK1 in retinal cells is assumed to play a significant role in glaucomatous and other retinal diseases by affecting the autophagy process and survival [48 , 49] . In summary , it has been demonstrated that upon virus infection , the SNRNP200 RNA helicase in combination with TBK1 via its Sec63-1 domain recognizes viral RNA , relocalizes into TBK1-containing cytoplasmic structures , and positively regulates IRF3 phosphorylation to promote the antiviral response . The regulatory role of SNRNP200 is confirmed in the MDM and PBMCs of RP33 patients due to the impaired production of IFN-β upon viral infection . The data revealed a crucial immunoregulatory role for the SNRNP200 helicase as well as for the Sec63-1 domain within SNRNP200 , which acts as an RNA sensor and as an adaptor for TBK1 to promote the IRF3-mediated antiviral innate immune response . Taken together , the data illustrate a novel function for SNRNP200 that is clearly distinguishable from its function in spliceosome activation and in pre-mRNA splicing . Through the development of immunomodulatory molecules , exploiting the function of human encoded regulators of the antiviral response presents an alternative strategy in treating a broad range of viral infections .
This study was approved by Institutional Review Board ( IRB ) of the participating institution ( McGill Children’s Hospital , McGill University Health Centre and Centre Hospitalier Universitaire de l’Université de Montréal ( CHUM ) ) and written informed consent was obtained from all participants before participation . SF3A1 , NHP2L1 and PHF5A cDNAs were purchased from GE Dharmacon/Open Biosystems . Following PCR-amplification , PCR products were cloned into pcDNA3 . 1-Hygro-MCS using EcoRV/HindIII[6] . SNRNP200 was subcloned from the pBluescriptSK-hBrr2 obtained from R . Lührmann [50] into pcDNA3 . 1-Hygro ( + ) ( Life Technologies ) using NotI and XhoI restriction sites . SNRNP200 deletion mutants and S1087L point mutation were generated by PCR . All constructs were verified by Sanger sequencing and subsequent western blot analysis . If necessary , validated constructs where subcloned into pcDNA3 . 1-MCS-FLAG . pIFNB1-LUC and p2xNF-κB-LUC luciferase reporter constructs were previously described [51–53] . Generation of stable HEK 293T cells harboring the pIFNB1-LUC and pEF1α-LUC promoters was previously described [8] . Human embryonic kidney HEK 293T ( ATCC ) , human epithelial adenocarcinoma HELA ( ATCC ) and human hepatoma cell lines Huh7 / Huh7 . 5 ( ATCC ) were cultured in Dulbecco's modified Eagle's medium ( DMEM , Wisent ) . Human lung adenocarcinoma epithelial A549 ( ATCC ) were cultured in Ham’s F-12 medium ( Life Technologies ) . Both media were supplemented with 10% fetal bovine serum , 100 units/ml penicillin , 100 μg/ml streptomycin and 2 mM glutamine ( all from Wisent ) at 37°C in an atmosphere of 5% CO2 . Transient transfections were performed with lipofectamine 2000 ( Life Technologies ) according to manufacturer’s protocol . Peripheral blood mononuclear cells ( PBMCs ) were isolated from fresh heparinized peripheral blood samples by Ficoll-Histopaque gradient centrifugation ( Sigma-Aldrich ) . Unfrozen PBMCs were washed twice in 10 ml of sterile RPMI 1640 and re-suspended in RPMI 1640 supplemented with 10% FBS . PBMCs were counted using a haemocytometer and counts were adjusted using trypan blue exclusion to plate 1x106 PBMCs in 100 μl RPMI 1640 supplemented with 10% FBS in 96-well plate . For monocyte-derived macrophage ( MDM ) , PBMCs were harvested has described above and monocyte were isolated using MACS Monocyte Isolation Kit II human ( Miltenyi Biotec ) as per manufacturer’s protocol before differentiation into MDM for five days in the presence of 10 ng/mL granulocyte-monocyte colony stimulating factor ( M-CSF , R&D ) . shRNAs from MISSION TRC shRNA lentiviral library ( Sigma-Aldrich ) were used as followed: shRNA targeting SNRNP200 ( TRCN0000051831 ) , SF3A1 ( TRCN0000006597 ) , PHF5A ( TRCN0000074878 ) , NHP2L1 ( TRCN0000074799 ) , or shRNA non-target ( NT ) . shRNA were transfected in combination with a standard packaging mix ( 1 . 5 μg pMDLg/pRRE , 1 . 5 μg pRSV-REV and 3 μg pVSVg ) as previously described[54] . siRNA ON-TARGETplus SMARTpool , Human SNRNP200 and siRNA non-targeting #1 Human , ON-TARGETplus ( GE Healthcare , Dharmacon ) , Santa Cruz HELIC2 siRNA ( h ) ( sc-75243 ) were transfected with lipofectamine RNAi Max ( Life Technologies ) for 48 hours according to manufacturer’s protocol . For assays in 96-well plates , cells were seeded in white 96-well plates at a density of 5 , 000 HEK 293T pIFNB1_LUC and 1 , 250 293T pEF1α-LUC in 100 μl of complete phenol-red free DMEM containing 4 μg/ml polybrene . Infection with lentivirus encoding shRNA were carried out immediately after cell seeding at a MOI of 10 ( except when specified otherwise ) and incubated for three days at 37°C in an atmosphere of 5% CO2 . Cells were infected with 100 HAU/ml of SeV ( Cantell Strain , Charles River Labs ) for 16 hours before cell lysis and firefly luminescence reading in a 100 mM Tris acetate , 20 mM Mg acetate , 2 mM EGTA , 3 . 6 mM ATP , 1% Brij 58 , 0 . 7% β-mercaptoethanol and 45 μg/ml luciferine pH 7 . 9 buffer . All infections were performed in an enclosed in a class II cabinet . Assays used in western blot or qRT-PCR experiments where scaled up accordingly and carried out with the maternal HEK 293T or appropriate cell line . For influenza infection , 3x105 HEK 293T pIFNB1_LUC cells were seeded in 6-well plates . The next day , cells were transfected with an influenza vRNA reporter plasmid [55] and infected with 0 . 1 ul of purified influenza virus ( A/PR/8/34 , from Charles River ) . Five days later , 20 ul of cell supernatant was used to quantify the Gaussia luciferase using a Gaussia Luciferase Assay HTS ( Nanolight Technology ) . Cell lysates were used to quantify the IFNB1 induction according to the Firefly luminescence assay described above . J6/JFH ( p7-Rluc2a ) virus production was conducted as previously described [56] . Briefly , HCV DNA template used for in vitro transcription was linearized using XbaI and subsequently transcribed using TranscriptAid T7 High Yield Transcription Kit according to manufacturer protocol ( Life Technologies ) . The resulting HCV RNA was then electroporated into Huh7 . 5 and virus-containing culture medium was collected , filtered ( 0 . 45 μm ) and kept at -80°C . For infection , 100μl of virus was added to 5 , 000 Huh7 cells that had been plated in 96-well white opaque plates one day before . Culture medium was replaced six hours later and Huh7 cells were transfected with pIFNB1-LUC ( 50 ng/well ) the next day . Three days later Huh7 cells were washed twice with PBS , before Rluc and Fluc quantification using the Dual-Luciferase Reporter Assay System ( Promega ) according to the manufacturer protocol . Cells were washed twice with ice-cold phosphate-buffered saline ( PBS; Wisent ) , harvested and lysed in 10mM Tris-HCl , 100mM NaCl , 0 . 5% Triton X-100 , pH7 . 6 with EDTA-free Protease Inhibitor Cocktail ( Roche ) . Cell lysates were clarified by centrifugation at 13 , 000 g for 20 min at 4°C and subjected to sodium dodecyl sulfate-polyacrylamide gel ( SDS-PAGE ) . Western blot analysis was performed using mouse anti-PHF5A ( Abnova ) , anti-IRF3 ( Santa Cruz ) , anti-TRAF3 ( Santa Cruz ) , anti-RIG-I ( Alexis Biochemicals ) , anti-ACTIN ( Chemicon International ) , anti-TBK1 ( Imgenex and Santa Cruz ) , anti-IKBKE ( Santa Cruz ) , anti-TUBULIN ( ICN ) , anti-GAPDH ( RDI ) and rabbit anti-SNRNP200 ( Sigma-Aldrich ) , anti-SF3A1 ( Santa Cruz ) , anti-RELA ( Santa Cruz ) , anti-NHP2L1 ( Abcam ) , anti-DDX3X ( Bethyl ) , anti-DDX60 ( Abcam ) , TRIF ( Cell Signaling ) , anti-ISG56 ( Novus Biologicals ) , anti-MDA5 ( Alexis Biochemicals ) , anti-MAVS ( Alexis Biochemicals ) , anti-IKBKE ( eBioscience ) , STAT1 ( ABCAM ) , STAT1 tyr701 ( ABCAM ) , IFNAR1 ( Santa Cruz ) and anti-IRF3-P-ser386 ( Abcam ) . HRP-conjugated secondary antibodies were from Bio-Rad . The chemiluminescence reaction was performed using the Western Lighting Chemiluminescence Reagent Plus ( PerkinElmer ) . For co-immunoprecipitation , FLAG-tagged protein expressing cells were harvested and lysed as described above . Resulting cell extracts were adjusted to 1 mg/ml and subjected to IP as follows: preclearing of the lysates was done by incubating lysates with 40 μl of a 50:50 slurry of immunoglobulin G-Sepharose ( GE Healthcare ) prepared in the lysis buffer with IgG beads for 1 hour . Pre-cleared lysate were immunoprecipitated by adding 20 μl of M2 anti-FLAG affinity gel ( Sigma-Aldrich ) prepared in TBS buffer ( 50 mM Tris-HCl , 150 mM NaCl , pH 7 . 4 ) overnight as described by the manufacturer . Immunoprecipitates were washed five times in lysis buffer . For interaction analysis , elution was performed using 250 ng/μl purified FLAG peptide for 45 min at 4°C ( Sigma-Aldrich ) . Eluates were analyzed by western immunoblotting . The microarray studies were performed with HEK 293T cells transduced with lentiviral-expressing shNT ( control ) or shSNRNP200 RNA targeting SNRNP200 gene for three days following 16 hours infection with SeV ( 100 U/ml ) or 16 hours of treatment with a mixture of IFN-α from human leukocytes ( 400 U/ml; Sigma-Aldrich ) . A total of 10 μg of RNA was reverse transcribed using oligo ( dT ) 16–18 primers and SuperScript II Reverse Transcriptase ( Life Technologies ) according to the manufacturer's instructions . Following purification using QIAquick PCR Purification kit ( Qiagen ) , up to 1 μg of purified cDNA was mixed with 5'-Cy3 labeled random nonamers ( Trilink Biotechnology ) and heated at 95°C for 10 minutes and transferred on ice for 10 minutes . Samples were mixed with 1 mM dNTP and 2 μl of 3’-5’ exo-Klenow fragment ( New England Biolabs ) and incubated at 37°C for 2 hours . The labeling reaction was stopped using 50 μM EDTA and the DNA precipitated using 0 . 5 M NaCl and 1 volume isopropanol , washed with 80% ethanol and resuspended in water . Hybridizations were carried out using the Human GE 4x44K v2 Microarrays ( Agilent Technologies ) containing probes targeting 27 , 958 Entrez Gene RNAs . Arrays were scanned at 5 μm resolution using a GenePix4000B scanner ( Molecular Devices ) . Data from scanned images were extracted using GenePix 6 . 1 ( Axon ) and processed and normalized using ArrayPipe ( v2 . 0 ) . Processed data was used as input for linear modeling using Bioconductor's limma package , which estimates the fold-change between predefined groups by fitting a linear model and using an empirical Bayes method to moderate standard errors of the estimated log-fold changes in expression values from each probe set . P values from the resulting comparison are adjusted for multiple testing according to the method of Benjamini and Hochberg . Subsequently , gene enrichment analysis are conducted using DAVID [57 , 58] , STRING [59 , 60] and Gene networks were constructed using GENEMANIA [60] . RNA pull-down assay was performed using Dynabeads M270 Streptavidin ( Life Technologies ) . Dynabeads were incubated with biotin-labeled RNA ( poly I:C ( InvivoGen ) and full-length Jc1 HCV ) for 1 hours according to manufacturer’s protocol . Biotin-HCV RNA was obtained by subjecting linearized HCV DNA to T7 reverse transcription ( TranscriptAid T7 High Yield , Life Technologies ) and biotin-dUTP ( Enzo Life Sciences ) . Saturated beads were added to whole 100 μg cell lysate and incubated , in a cold room , on a rotating wheel . Beads were washed three times and RNA-bound proteins were eluted after boiling in 0 . 1% SDS and analyzed by western blot . Poly ( dA:dT ) and Poly ( dG:dC ) were purchased from Sigma and labeled using Label IT Nucleic Acid Labeling Kit ( Mirus Bio ) and biotin-DNA pull-down assays were performed as described above . Total cellular RNA was extracted with the RNeasy Mini kit ( Qiagen ) . Reverse transcription was performed on 1 μg total cellular RNA using the High Capacity cDNA Reverse Transcription kit ( Applied Biosystems ) . In order to amplify only the cDNA , primers were located in the splicing junction between two exons . PCR reactions were performed using 1 . 5 μl of cDNA samples ( 15 ng ) , 5 μl of the Fast TaqMan PCR Master Mix ( Applied Biosystems ) , 10 pmol of each primer ( IDT ) and 5 pmol of the UPL probe ( Roche ) in a total volume of 10 μl . The ABI PRISM 7900HT Sequence Detection System ( Applied Biosystems ) was used to detect the amplification level and was programmed to an initial step of 3 minutes at 95°C , followed by 40 cycles of 5 seconds at 95°C , 30 seconds at 60°C and 1 second at 72°C . All reactions were run in duplicate on biological duplicate and the average values were used for quantification . ACTIN ( β-actin ) or GAPDH ( Glyceraldehyde 3-phosphate dehydrogenase ) and HPRT1 ( hypoxanthine phosphoribosyltransferase 1 ) were used as endogenous controls . The relative quantification ( RQ ) of target genes was determined by using the ΔΔCt method with the Sequence Detection System ( SDS ) 2 . 2 . 2 software ( Applied Biosystems ) . Plaque assays were conducted in VERO cells and MDCK . 2 cells ( ATCC ) using a method described elsewhere [61] . Briefly , supernatants were harvested from infected cells and used to inoculate in serial dilutions VERO ( SeV ) and MDCK . 2 cells ( FLUA ) for 45 minutes and 1 hour , respectively . After infection , cells were wash with PBS and an overlay of 0 , 6% agarose was superimposed to 2X DMEM medium . At 72 hours post-infection , cells were colored with crystal violet , washed with PBS and colonies ( lysed cells ) were counted to compute viral titers . ELISA assays were carried out with 50 μl of cell culture supernatants using the VeriKine Human Interferon Beta Elisa Kit ( PBL Assay Science ) according to the manufacturer’s protocol . Samples were run as technical duplicates on biological triplicates . HEK 293T were seeded in cover slip-containing 24-well plates and co-transfected with FLAG-SNRNP200 WT or S1087L mutant and MYC-TBK1 24 hours later . The following day , cells were infected or not with SeV for 16 hours before being washed twice with PBS , fixed with 4% paraformaldehyde-containing PBS during 20 minutes at room temperature and then permeabilized in 0 . 2% Triton X-100/PBS during 15 minutes . Blocking was made in PBS with 10% normal goat serum , 5% bovine serum albumin ( BSA ) and 0 . 02% sodium azide during 45 minutes at room temperature . Following three rapid washes , cells were labelled with rabbit anti-FLAG ( Sigma-Aldrich ) and mouse anti-MYC ( Santa Cruz ) primary antibodies diluted in 5% BSA/0 . 02% sodium azide/PBS for 2 hours . Slides were washed three times in PBS and then labeled with anti-rabbit or anti-mouse AlexaFluor 488 , 594 or 647 secondary antibodies ( Life Technologies ) diluted in 5% BSA/0 . 02% sodium azide/PBS for 1 hour . Cells were extensively washed and incubated with Prolong Gold with DAPI ( Life Technologies ) . Alternatively , nuclei were labeled with Syox Green ( Life Technologies ) . Labelled cells were then examined by laser scanning microscopy using a TCS SP5 ( Leica ) .
|
The innate immune system is the first line of defense against pathogens and relies on the recognition of molecular structures specific to pathogens by sensor receptors . These receptors activate a signaling cascade and induce a protective cellular innate immune response . In this study , we provide evidence for a role of the spliceosomal SNRNP200 RNA helicase in promoting antiviral response that is clearly distinguishable of the one in pre-mRNA splicing . The depletion of SNRNP200 in human cells resulted in a reduced interferon-β ( IFNB1 ) production and increased susceptibility to viral infection . We showed that SNRNP200 positively regulates activation of the key transcription factor IRF3 via interaction with TANK kinase 1 ( TBK1 ) . Upon infection , SNRNP200 binds viral RNA and relocalizes into TBK1-containing cytoplasmic structures to promote IRF3 activation and IFNB1 production . Of clinical relevance , we observed a significantly hindered antiviral response of PBMCs from patients carrying a dominant SNRNP200 mutation associated with retina pigmentosa type 33 ( RP33 ) , an inherited degenerative eye disease . We showed that the RP33-associated S1087L SNRNP200 mutant has lost the ability to bind RNA and that its expression fails to rescue antiviral response in SNRNP200 silenced cells . Our study provides new insights into a role within the antiviral response for spliceosome SNRNP200 helicase as an RNA sensor and TBK1 adaptor in IRF3 signaling .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"phosphorylation",
"medicine",
"and",
"health",
"sciences",
"293t",
"cells",
"enzymes",
"biological",
"cultures",
"immunology",
"enzymology",
"plasmid",
"construction",
"rna",
"helicases",
"dna",
"construction",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"proteins",
"gene",
"expression",
"cell",
"lines",
"molecular",
"biology",
"immune",
"response",
"biochemistry",
"rna",
"spliceosomes",
"helicases",
"post-translational",
"modification",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences"
] |
2016
|
Spliceosome SNRNP200 Promotes Viral RNA Sensing and IRF3 Activation of Antiviral Response
|
More than 80% of schistosomiasis patients in China live in the lake and marshland regions . The purpose of our study is to assess the effect of a comprehensive strategy to control transmission of Schistosoma japonicum in marshland regions . In a cluster randomized controlled trial , we implemented an integrated control strategy in twelve villages from 2009 through 2011 in Gong'an County , Hubei Province . The routine interventions included praziquantel chemotherapy and controlling snails , and were implemented in all villages . New interventions , mainly consisting of building fences to limit the grazing area for bovines , building safe pastures for grazing , improving the residents' health conditions and facilities , were only implemented in six intervention villages . Results showed that the rate of S . japonicum infection in humans , bovines , snails , cow dung and mice in the intervention group decreased from 3 . 41% in 2008 to 0 . 81% in 2011 , 3 . 3% to none , 11 of 6 , 219 to none , 3 . 9% to none and 31 . 7% to 1 . 7% , respectively ( P<0 . 001 for all comparisons ) . In contrast , there were no statistically significant reductions of S . japonicum infection in humans , bovines and snails from 2008 to 2011 in the control group ( P>0 . 05 for all comparisons ) . Moreover , a generalized linear model showed that there was a higher infection risk in humans in the control group than in the intervention group ( OR = 1 . 250 , P = 0 . 001 ) and an overall significant downward trend in infection risk during the study period . The integrated control strategy , designed to reduce the role of bovines and humans as sources of S . japonicum infection , was highly effective in controlling the transmission of S . japonicum in marshland regions in China . Chinese Clinical Trial Registry ChiCTR-PRC-12002405 .
Schistosomiasis is an important public health issue in China where it continues to pose a serious threat to human well-being [1]–[3] . Since the People's Republic of China was established in 1949 , the Chinese government has given high priority to the control of schistosomiasis , establishing a number of special bodies to manage control activities from the national to the township level [4]–[6] . These policies resulted in remarkable success in that the number of schistosomiasis patients were reduced by over 90% , from about 11 . 6 million cases in the 1950s to 726 , 000 cases in 2004 [7] . A further report in 2004 indicated that disease transmission had been interrupted or controlled in 42% of provinces , 40% of counties , and 53% of towns , previously endemic for schistosomiasis [8] . However , prospects for the control of schistosomiasis have been less optimistic in recent years , particularly since the termination of the World Bank Loan Project ( WBLP ) for schistosomiasis control at the end of 2001 [7] . Compared with the second national survey of S . japonicum in China in 1995 , the third national survey conducted in 2004 showed that the prevalence of S . japonicum infection in humans had not substantially changed in the lake and marshlands and other areas of Southern China [7] . By the end of 2009 , a total of 365 , 770 cases of S . japonicum were estimated in China , and 89 counties had not yet reached the criterion for transmission control which stipulated that the human prevalence should be less than 1% over a length of time [9] , [10] , i . e . the prevalence in these counties was >1% . Over the past 2–3 decades , the strategies for S . japonicum control in southern China , including the vast lake and marshland regions , has involved chemical mollusciciding , alteration of the oncomelanid intermediate snails habitats , and synchronous chemotherapy with praziquantel for all villagers and their bovines [6] , [11] . Historically , these strategies achieved some effect , but recent studies demonstrated there were some problems as the control options resulted in environmental pollution leading to ecological damage [11] , [12] . Moreover , owing to the high rates of reinfection in both humans and bovines , frequent flooding , and the complex environment , more persons have been infected and the habitat of the Oncomelania snails has increased in the lake and marshlands [11] , [13] . Consequently , a more effective strategy was needed urgently in these areas . From 2005 through 2007 , an important study of schistosomiasis japonica control was undertaken by Wang et al . in villages along the Poyang Lake in Jiangxi Province involving a comprehensive integrated approach aimed at reducing S . japonicum transmission to snails from cattle and humans , which play key roles as sources of S . japonicum [14] . The integrated strategy , which included removing bovines from snail-infested grasslands , providing farmers with mechanized farming equipment , building safe water systems , providing adequate sanitation , and implementing health education and synchronous chemotherapy with praziquantel for both villagers and bovines , was highly effective [14] . On the basis of identifying and controlling the main S . japonicum infection sources , we implemented a similar strategy ( March 2009 through November 2011 ) to that employed by Wang et al . in the marshlands of Gong'an County [14] , Hubei Province , another major endemic area for schistosomiasis japonica . The objective of the study was assessing the strategy's effect in marshland regions .
Written ethical approval for this study was obtained from the Ethics Review Committee of Hubei Provincial Center for Disease Control and Prevention ( no . 200803 ) . Written informed consent was obtained from all adults and from parents or guardians of minors before participation in the study . The participants had the opportunity to withdraw from the study at any time . Before beginning work on the study , the bovine owners provided consent to have their animals involved in the study . Moreover , permits for the bovines were obtained from Gong'an County Animal Husbandry Bureau . All animal work was carried out under the guidance of the Institute for Laboratory Animal Research ( ILAR ) , and approved by Ethics Review Committee of Animal Experiments , Hubei Provincial Center for Disease Control and Prevention ( no . 2008a05 ) . Both doses of praziquantel ( i . e . , single 40 mg/kg dose for humans identified as stool egg-positive or 25 mg/kg for infected bovines ) were within Chinese Ministry of Health published guidelines [15] . Gong'an County is located in typical marshland with a water area of 364 hectares ( km2 ) along the mid-to-lower reaches of the Yangtze River in southwest Hubei Province , China . In 2009 , there were 294 schistosomiasis-endemic villages ( out of a total of 326 ) , 539 cases of advanced schistosomiasis , and 36 , 612 cases of chronic schistosomiasis in Gong'an County; the prevalence of schistosomiasis in humans and bovines was 2 . 75% and 2 . 41% , respectively [16] . We carried out the study in 12 villages from 12 towns in Gong'an County which were selected by a two-stage random sampling procedure . The 12 towns were first randomly selected from 16 towns; then 12 schistosomiasis-endemic villages were selected from the 12 towns randomly ( each town selected a village ) . Finally , these villages were randomly divided into intervention and control groups ( Figure 1 ) . E Jinghu , Gu Shengsi , Jianhong , Lianmeng , Tongqiao , Tuanjie were assigned to the intervention group and Guoqing , Nanyang , Qingyun , Qingji , Zhu Jiahu , Tongsheng were assigned to the control group ( Figure 2 ) . The villages had similar agriculture resources ( rice , cotton and brassica napus ) and the residents mainly were farmers ( about 75% ) . The residents were exposed to contaminated water when they performed their agricultural or daily activities ( i . e . cultivating crops , catching fish and washing clothes ) . The participants met the following inclusion criteria: a ) must have been a resident of the village for more than 12 months; b ) should be aged 6–65 years; c ) should continuously reside in the village for the study period; d ) have no serious diseases , such as malignant cancer . Prior to implementing the new control strategy , routine interventions had been used in Gong'an County to control S . japonicum infection , which comprised synchronous praziquantel chemotherapy of both villagers and bovines and mollusciciding of snails [17] , [18] . These procedures were continued in all 12 study villages for the duration of the study but additional interventions were incorporated into the intervention group . These comprised of the building of fences to limit the grazing area for bovines , building safe pastures for grazing , improving the residents' health conditions and facilities , and strengthening specialized schistosomiasis clinics at the village level . Five outcome measures , including the primary one ( the prevalence of S . japonicum in humans ) and four secondary ones ( the rate of S . japonicum infection in bovines , cow dung , snails and mice ) , were determined annually and used to assess the strategy's effect . Participants that were positive for both IHA and Kato-Katz were defined as infected and the prevalence of S . japonicum of participants in all 12 study villages was in October/November , after the second transmission season . During the same period , the infection of S . japonicum of all bovines was checked in all 12 study villages by the miracidium hatching test in dung [15] . In addition , we gathered fresh cow dung from the grasslands which located in the 6 intervention villages ( the sampled locations were the safe grazing pastures from 2009 to 2011 ) , and used the miracidium hatching test to detect the infection of S . japonicum in cow dung . In April and May , we systematically sampled Oncomelania hupensis snails along the river banks and in marshlands and ditches around all 12 study villages . In every village , we investigated at least 2000 sample units ( 0 . 11 m2 frame ) and gathered all the live snails in the frame , crushed them , and examined them for S . japonicum infection using microscopy . During the peak of the transmission season , July and August , we used exposure tests with mice to assess the infectivity of water in the 6 intervention villages [14] . Sentinel mice were exposed for 2 hours every day for 3 consecutive days . After 30 to 35 days , we sacrificed the mice and checked for adult worms of S . japonicum in their mesenteric veins . All statistical analyses were performed with the use of Statistics Analysis System , version 9 . 1 ( SAS Institute Inc . , Cary , NC , USA ) . Confidence intervals ( CIs ) were calculated using standard formulae based on the binomial distribution ( annual infection rate of humans ) . Chi-square test or Fisher exact probability test were used to examine the differences of proportions . A generalized linear model ( GLM ) with a logit link and a binomial error distribution was used to analyze the risk of S . japonicum infection in humans . Generalized estimating equations of parameters with an unstructured variance-covariance matrix were used to account for repeated measures on individuals during the study period . The SAS proc GENMOD was used to estimate the parameters . Two-sided P-values were calculated for all tests and P-values <0 . 05 were considered statistically significant .
The prevalence of S . japonicum in humans was 3 . 41% , 2 . 83% , 2 . 01% , and 0 . 81% from 2008 to 2011 in the intervention villages ( comparing with 2008 , P = 0 . 101 , P<0 . 001 and P<0 . 001 , respectively , Table 2 ) . In the control villages , there was no significant statistically decline in the rate of infection from 2008 to 2011 ( P>0 . 05 for all comparisons ) . The generalized linear model ( Table 3 ) , yielding odds ratios ( OR ) adjusted for participants' age and gender , showed a higher infection risk in humans in the control villages than the intervention villages ( OR = 1 . 250 , P = 0 . 001 ) ; and an overall significant downward trend in infection risk during the study period . The S . japonicum infection rates in bovines were 3 . 3% , 2 . 9% , 1 . 3% , and 0 . 0% from 2008 to 2011 in the intervention villages ( comparing with 2008 , P = 0 . 820 , P = 0 . 111 and P<0 . 001 , respectively , Table 4 ) . In contrast , there was no significant decline in the rate of infection during the 4-year period in the control villages ( P>0 . 05 for all comparisons ) . From 2008 to 2011 , 11 of 6 , 219 , 9 of 5 , 975 , none of 25 , 010 , and none of 15 , 490 snails were infected with S . japonicum in the intervention villages ( comparing with 2008 , P = 0 . 824 , P<0 . 001 and P<0 . 001 , respectively , Table 5 ) . In the control villages , there was no significant statistically decline in the rate of infection from 2008 to 2011 ( P>0 . 05 for all comparisons ) . In the intervention villages in 2008 , 21 of 533 ( 3 . 9% ) of the cow dung samples contained S . japonicum eggs . After the implementation of the new control strategy , 15 of 483 ( 3 . 1% ) , 5 of 476 ( 1 . 1% ) and none of 356 cow dung were infected in 2009 , 2010 and 2011 , respectively ( comparing with 2008 , P = 0 . 501 , P = 0 . 005 and P<0 . 001 , respectively ) . In 2008 , before implementation of the new control strategy , 19 of 60 ( 31 . 7% ) mice were infected with S . japonicum . In 2009 , 2010 and 2011 , 13 of 60 mice ( 21 . 7% ) , 7 of 120 ( 5 . 8% ) and 2 of 120 ( 1 . 7% ) mice were infected , respectively ( comparing with 2008 , P = 0 . 216 , P<0 . 001 and P<0 . 001 , respectively ) . No serious adverse events were reported in the study .
An integrated control strategy , aiming to weaken the roles of bovines and humans as sources of S . japonicum infection , was effective in controlling the transmission of S . japonicum in marshland regions of China .
|
More than 80% of schistosomiasis patients in China live in the lake and marshland regions . Hence , how to control transmission of Schistosoma japonicum in these regions is especially important . From 2009 through 2011 , we implemented an integrated control strategy , designed to reduce the role of bovines and humans as sources of S . japonicum infection , in twelve villages Gong'an County of Hubei Province , which is located in typical marshland . After three years , the rate of S . japonicum infection in humans , bovines and snails significantly declined in the six intervention villages . In contrast , there was no significant decline in these indexes in the six control villages . Moreover , there was a higher infection risk in humans in the control group than the intervention group . Our study showed that the integrated control strategy was highly effective in controlling the transmission of S . japonicum in marshland regions of China .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"preventive",
"medicine",
"infectious",
"diseases",
"schistosomiasis",
"public",
"health",
"and",
"epidemiology",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"public",
"health"
] |
2013
|
Assessing the Effect of an Integrated Control Strategy for Schistosomiasis Japonica Emphasizing Bovines in a Marshland Area of Hubei Province, China: A Cluster Randomized Trial
|
Clostridium botulinum produces botulinum neurotoxins ( BoNTs ) , highly potent substances responsible for botulism . Currently , mathematical models of C . botulinum growth and toxigenesis are largely aimed at risk assessment and do not include explicit genetic information beyond group level but integrate many component processes , such as signalling , membrane permeability and metabolic activity . In this paper we present a scheme for modelling neurotoxin production in C . botulinum Group I type A1 , based on the integration of diverse information coming from experimental results available in the literature . Experiments show that production of BoNTs depends on the growth-phase and is under the control of positive and negative regulatory elements at the intracellular level . Toxins are released as large protein complexes and are associated with non-toxic components . Here , we systematically review and integrate those regulatory elements previously described in the literature for C . botulinum Group I type A1 into a population dynamics model , to build the very first computational model of toxin production at the molecular level . We conduct a validation of our model against several items of published experimental data for different wild type and mutant strains of C . botulinum Group I type A1 . The result of this process underscores the potential of mathematical modelling at the cellular level , as a means of creating opportunities in developing new strategies that could be used to prevent botulism; and potentially contribute to improved methods for the production of toxin that is used for therapeutics .
Commonly found in any soil or water environment , the spore forming Gram-positive rod-shaped bacterium Clostridium botulinum and two other clostridia ( C . baratii and C . butyricum ) can , under suitable anaerobic conditions , release botulinum neurotoxins ( BoNTs ) [1 , 2] . BoNTs are highly potent substances with an estimated human lethal dose of ~30-100ng [1 , 2] , and are the most powerful toxins affecting human and animal health . BoNTs cause botulism , a severe neuro-paralytic disease that can lead to death in humans as well as in a range of other mammals and birds [3 , 4] . BoNTs enter into the blood stream in one of three ways: ( 1 ) toxin production by bacteria that have colonized the digestive tract of either children less than 12 months of age ( infant botulism ) or individuals with a suppressed normal intestinal flora ( e . g . , following antibiotic treatment ) including those that have anatomical or functional bowel abnormalities ( Adult intestinal toxemia botulism ) [2 , 5 , 6]; ( 2 ) infection and toxin formation in a wound ( wound botulism ) [5 , 7 , 8]; and ( 3 ) following oral ingestion of pre-formed toxin in contaminated foods ( foodborne botulism ) [2 , 5] . BoNTs target the peripheral motor nerve terminals , blocking neurotransmission by selectively hydrolysing proteins that are involved in the fusion of synaptic vesicles with the presynaptic plasma membrane , thereby preventing acetylcholine release [5 , 6 , 9–11] . Foodborne botulism is a severe and sometimes fatal disease [2] . Although there are fewer cases of foodborne illness caused by C . botulinum than by bacteria of the Salmonella genus , the death rate from botulism is relatively high , 17 . 3 percent , compared with 0 . 5 percent for Salmonella [12] . Thus , the severity of the disease and the widespread presence and persistence of C . botulinum spores make botulism a global health concern and a cause for vigilance [2] . Seven serotypically distinct botulinum neurotoxins ( serotypes A-G ) and more than 40 different subtypes [6 , 9 , 13] are produced by six phylogenetically distinct clostridia ( C . botulinum Groups I-IV and some strains of C . baratii and C . butyricum ) . Considering the highly potent nature of the neurotoxin , methods that limit the proliferation of C . botulinum and the associated neurotoxin production in food are a major priority for the food-processing industry; these processes are complicated by the physiological differences among clostridia . The structures and the mechanisms of action of BoNTs are reasonably well established [11 , 14–20] , however , regulation of botulinum neurotoxin gene ( bont ) expression and BoNT production are not fully understood . Likewise , the environmental signals which trigger the synthesis of the BoNTs and the regulatory network and actors which control the production of the toxin ( and many subsequently regulated genes ) remain to be elucidated . What is known , firstly , is that in vitro experimental methods developed and applied to monitor bont gene expression in C . botulinum show a peak in neurotoxin gene expression during late exponential and early stationary phase of population growth; expression decreases drastically during stationary phase for C . botulinum Group I type A [21–26] and for C . botulinum Group II type E [22 , 27] . However some of these studies examined a relatively small number of time points during population growth so that the full bont gene expression profile is not always observed . Moreover , these studies show that the quantity of BoNT produced can be influenced by the strain , by culture conditions and by the nutritional status of the medium—although the precise mechanisms are unknown . Secondly , bont gene expression is reported to be tightly regulated through positive and negative regulatory systems . Positively , through the participation of BotR [17] , Agr quorum sensing system [28] , CodY [29] and CLC_1094/CLC_1093 ( equivalent to CBO_1042/CBO_1041 ) , CLC_1913/ CLC_1914 ( equivalent to CBO_1967/ CBO_1968 ) and CLC_0663/CLC_0661 ( equivalent to CBO_0608/CBO_0607 ) two component signal transduction systems [30] . Negatively , through the participation of CBO0787/CBO0786 ( equivalent to CLC_0843/CLC_0842 ) [31] which is also a two component signal transduction system [32] . Thirdly , in most C . botulinum Group I type A1 strains , the genes encoding the neurotoxin ( bont ) and its associated non-toxic neurotoxin proteins ( ANTPs ) ( ntnh , has ) are located in a gene cluster and are organized in two transcriptional units ( or operons ) , namely , the ntnh-bont and ha operons [9 , 33] . The first operon ( ntnh-bont ) , which is located at the 3′ end of the botulinum locus , encompasses the bont gene immediately preceded by the ntnh gene . Both genes are co-transcribed in the same orientation , and the organization of this operon is highly conserved in all botulinum toxin forming clostridia . The second operon contains the ha genes and differs slightly between the various subtypes ( BoNT/A1 , A5 , B , C , D and G ) . The ha operon contains successive genes for the 33 kDa ( ha33 ) , 17 kDa ( ha17 ) , and 70 kDa ( ha70 ) hemagglutinins [30 , 34] . These hemagglutinin genes are localised upstream of the ntnh-bont genes and are transcribed in the opposite orientation [5 , 35 , 36] . Thus , the nontoxic proteins for subtype A1 include NTNHA ( which together with BoNT forms the minimally functional progenitor toxin complex ( M-PTC ) ) and three hemagglutinin ( HA ) proteins ( HA70 , 17 and 33 ) , which assemble ( with the M-PTC ) to form the large size toxin complex ( L-PTC ) [10 , 37 , 38] . Lastly , BoNT is released from the bacterium and exists in nature in the form of a complex [36 , 39 , 40] , i . e . not as a pure toxin [41] . The distinct neurotoxins form complexes of different sizes ( from 288 to 900 kDa ) by association with ANTPs , i . e . hemagglutinins ( HAs ) and nontoxic non-hemagglutinins ( NTNHs ) . These ANTPs spontaneously associate with BoNTs at low pH and dissociate at pH 7 . 5 and above . The associated proteins protect the neurotoxin and facilitate its absorption into the body [37 , 42] . C . botulinum Group I type A1 ( BoNT/A1 ) neurotoxins are so far the best characterized neurotoxins , a consequence of both their frequent involvement in human botulism worldwide as well as their greater potency and , therefore , suitability for therapeutics [1] . With all the aforementioned findings , it is reasonable to conclude that bont gene transcription and neurotoxin production may be influenced by the bacterial strain . In particular gene transcription may be influenced by the availability of particular nutrients ( although the precise mechanisms are unknown ) that are present during the transition from late-exponential to early-stationary phase cultures ( i . e . , growth phase dependent ) . In turn this transcription is dependent on both positive and negative regulatory elements . This evidence supports the construction of a signal transduction and sensory transcription regulatory network to describe the kinetics of neurotoxin production [32] . Current mathematical models of C . botulinum are based on statistical data aggregation and describe beliefs concerning the unknown concentrations of C . botulinum spores in the environment , the uncertain inactivation kinetics for populations of spores at high temperatures and the germination and growth of C . botulinum populations for a variety of physico-chemical conditions [26 , 43–51] . These models do not attempt to identify elements of regulatory control which are the key to transferability and to an appreciation of cell to cell variations ( in many situations foodborne botulism may be driven by very few cells so that cell variability is a crucial unknown ) . Furthermore current models sum-up many component processes , such as signalling , permeability and enzymatic activity , obscuring opportunities for improved understanding . The use of computational models amenable to simulation and to the analysis of what-if type scenarios would permit further formulation of hypotheses concerning the gene expression profiles and interactions; additionally a process of iterative computer simulation would guide future experimentation . In this report we tackle the challenge of integrating the various sources of multi-scale biological evidence into a mechanistic model using ordinary differential equations . In turn this model is used to explain the regulated toxigenesis process of C . botulinum Group I type A1 . Strains of C . botulinum Group I type A1 fall into three classes , ( i ) those that carry the neurotoxin gene in an ha cluster , ( ii ) those that carry the neurotoxin gene in an orfx cluster , and ( iii ) those that also carry a type B neurotoxin gene and form a small amount or no type B neurotoxin . We focus on the first of these , strains that carry the neurotoxin gene in an ha cluster [9 , 52] . We use biological data from the literature that relates to five strains of C . botulinum Group I type A1 ( ATCC19397 , 62A , Hall A , Hall A-hyper and ATCC3502 ) ; the close relationship between these strains having been established by whole genome sequencing , microarray analysis and MLST [53–56] . Additionally , these five strains all possess identical or very similar bont and botR genes [52 , 57 , 58] . This modelling task , to the best of our knowledge , has not been approached so far . We first review the experimental knowledge that has been published in the literature on the patterns of toxin production and toxin gene expression and then identify the main aspects of the regulation , highlighting the key molecular players . The main contribution of this report is shown in the results section where the diverse available information used in constructing the mathematical model of toxin formation by C . botulinum Group I type A1 is integrated into a complete model in an incremental way . The proposed model is implemented and simulated , to confirm its ability to reproduce the observed patterns of behaviour , in various wild type [WT] strains and in various mutant strains that have previously been experimentally characterized . We further discuss the results and focus on a review of the hypotheses that were made throughout the modelling process , identifying the opportunities they offer for the definition of specific experimental settings that would help in shedding light on several of the poorly understood steps in the process of toxin formation by C . botulinum Group I type A1 .
The proposed mathematical model is based on a continuous-deterministic approach , where the components of the model , such as concentrations , number of bacterial cells etc . , are represented as continuous variables and their variation over time is expressed through their first derivative . The dynamics of the multi-component system corresponds with a set of coupled ordinary differential equations for which numerical solutions are obtained by computer simulation . To avoid dealing with the mathematical details of differential equations , we adopted a reaction-based specification language to describe the interactions among the model variables . The whole modelling process is supported by the COPASI modelling and simulation software package [62] , which takes as its input the reaction-based specification of the model , and provides the simulated time courses of the variable dynamics . To simplify the process of model definition , we used an incremental procedure which allowed us to build increasingly complex versions of the model , each one incorporating additional pieces of biological evidence and some additional modelling assumptions . The modelling approach considers two separate levels of representation: ( 1 ) at the cell level , the dynamics of the population in a culture and ( 2 ) at a sub-cellular level , the network that regulates toxigenesis and gene expression . At a cell level the model describes the dynamics of the consumption of nutrients and of quorum-sensing signals in the culture . This is then coupled with the dynamics of the regulation and gene-expression which is described by the sub-cellular level .
Details of a computational model for the growth of a population of C . botulinum Group I type A1 cells in a culture have been reported by [44] . As thoroughly explained in [44] , the rationale underlying the need of modelling population dynamics is rooted in the experimentally observed correlation between the bacterial growth phase and the toxin production process . Further evidence supporting this correlation at genetic regulation level is described in the section on the molecular model of BoNT synthesis regulation . The mathematical modelling of C . botulinum cultures is based on a compartmentalization of the growing population of cells into three distinct groups: The initial population of C . botulinum cells is fully composed of AC cells , which later evolve to RCs and may commit to sporulation and become SCs . These processes are influenced by some biochemical species generically termed “Signal” , as shown in Fig 1 . A future development , not currently included , is to extend the present analysis to start with bacterial spores and to therefore incorporate steps for spore germination and outgrowth [64] . Different hypotheses relating to the nature of the “signal ( s ) ” ( previously described in [44] ) led to the discrimination of plausible modelling scenarios and were used to generate corresponding models that were then evaluated on their ability to reproduce the observed pattern of growth observed for C . botulinum type A1 strain ATCC 19397 . We found that a model where two distinct signal sources were considered—the first one determined by the abundance of nutrients essential to C . botulinum cell growth , which we denoted by the abstract species N , and a second one endogenously produced by the bacterial cells and used as a quorum-sensing signal , denoted by S–was most successful at explaining the pattern of growth observed for C . botulinum type A1 strain ATCC 19397 . A diagrammatic representation of this modelling option is included in Fig 1 . In this model the rate of cell reproduction increases with the nutrient concentration , N , whilst the rate of sporulation increases with the concentration of the chemical signal S . The model proposed in [44] was encoded using eight reactions . We consider here an updated version , still based on the same rationale , which is encoded by the six reactions listed in Table A of Supporting Information File 1 ( Table A in S1 Text ) . As previously reported in Figure 6 of [44] , this model produces a good fit for the experimental growth data generated for strain ATCC 19397 . Several environmental stimuli have been identified with positive and negative regulation of toxin production in C . botulinum Group I type A1 . Neurotoxin production has been reported to be associated with the transition from late-exponential to early-stationary phase cultures . This is indicated by a peak in the level of neurotoxin gene cluster expression that is clearly observable in the late-exponential to early-stationary phase of cultures and which drastically decreases during the later stationary phase ( as shown in Fig 2 ) . Moreover , the expression patterns for all the genes , in both the ntnh/bont and the ha operon , show an equivalent correlation with population dynamics ( data available in [26] , [21] , [22] and [23] ) . This points to regulatory elements that link population growth to toxigenesis in C . botulinum type A1 . In this section we expound the procedure used for calibrating model parameters , and then proceed to validate the model by checking against characteristics of toxigenesis which have been reported previously in sections on nutrient and quorum-sensing regulated population growth model and the molecular model of BoNT synthesis regulation . To find suitable values for model parameters , we used the experimental data from [59] for type A1 strain ATCC 19397 , which we considered as the 'wild type' organism for the purpose of our modelling ( WT , hereafter ) . The experimental time course for the population size ( measured in CFU/ml over time in [59] ) provides the parameters of the population sub-model , i . e . the kinetic parameters of reactions ( 1 ) to ( 6 ) provided in Table A of Supporting Information File 1 ( Table A in S1 Text ) . The amount of toxin in the supernatant measured in the same experiment in [59] ( measured in MLD50/ml over time ) provides the data for fitting the gene expression sub-model , i . e . the kinetics of reactions ( 1 ) to ( 49 ) listed in Table B of Supporting Information File 1 ( Table B in S1 Text ) . The model parameters are reported in Supporting Information File 2 ( S2 Text ) . The fitted model is compared with the WT experimental data in Fig 5A and 5B . The experimental data points are shown as empty circles , whereas the computational model is reported as continuous lines . There is an inevitable match between model outcomes and the experimental 'model training' data , which is confirmed by analysis of correlation . For the population an R2 measure is 0 . 975 while for the toxin production it is 0 . 95 . After tuning model parameters to fit WT observed behaviour , we proceeded to validate the model , by assessing its ability to reproduce the behaviours experimentally observed in the different C . botulinum mutant strains we had considered in the study . We examined four different mutations , which are implemented in the WT model exclusively by changing the initial state of the model , i . e . without any change to the kinetics of the reactions . The mutants we considered for the purposes of our validation are as follows: For each mutant , we obtain and report the toxigenesis predictions ( pattern and amount of BoNT ) from both the WT model and the mutant model . Then we examine the relationship between model predictions and the experimental results to determine the ability of the models to reproduce wet-lab evidence .
For the first time we have defined and implemented a computational model , at the molecular level , for the highly regulated process of BoNT production in C . botulinum Group 1 type A1 . In contrast to existing modelling approaches , largely aimed at risk assessment for C . botulinum , this development does not integrate out component processes such as signalling , membrane permeability and metabolic activity , and it does include elements of genetic in-formation . The model captures causal relations among the known regulators of toxigenesis , at the molecular level . This leads to a computational model which is able to embrace both the population dynamics of the cells ( so that we were able to include growth phase-dependent patterns of bacterial behaviour ) as well as behaviour of the genetic regulatory network and the molecular interactions that link toxin expression with the environmental and population generated signals . The model construction has integrated the available experimental knowledge on the factors that , at a molecular level , regulate toxigenesis in C . botulinum Group 1 type A1; previously reviewed in [32] . In addition it portrays the effects of nutrient availability and quorum-sensing molecules and their coupling with distinct sensing and response TCSs that are regulators that mediate the activation of the toxin coding genes . The model satisfies a validation based on its ability to predict the effects of various mutations that have been experimentally studied in vitro . The validation results suggest that the model is able to provide a plausible explanation for the interplay of the multiple regulation mechanisms that impact toxin production in C . botulinum Group 1 type A1 . Models that encode causality have significant advantages over purely statistical descriptions , because they lend themselves to exploration of what-if-scenarios and generating test-able hypotheses . For instance , the model proposed here can be used to predict the pheno-types of mutants that have not yet been studied in vitro . As an example , we can explore the predicted toxin production of a mutant cell where the positive regulator CodY is removed and also the negative regulator TCS CBO0787/CBO0786 is silenced ( the CO-DY_M+C786_M double mutant , based on the abbreviation used in the section on the computational model ability to reproduce additional experimental results . We can then compare the predicted concentration of toxin in the supernatant for this double mutant with that obtained by the C . botulinum WT model ( Fig 9 ) . Our model predicts that silencing the CBO0787/CBO0786 TCS rescues the CODY_M mutant ability to produce toxin , to levels similar to those of the wild type . This is a prediction that can be tested in vitro , to provide either further support for the model structure , or to generate new evidence that can be integrated into parameterization and hence improve the predictive capability . Furthermore , models can be tested for conditions not yet considered in the laboratory set-ting; thus obtaining additional predictions that could be conductive to the definition of experimental settings . This novel model is an initial attempt to elucidate toxigenesis in C . botulinum Group 1 type A1 . We expect it will require further tuning , improvements and changes . We made a substantial number of assumptions about the dynamics of the activation of promoters , which require experimental confirmation . The process of toxigenesis has been simplified not to include too many unknown details of the hemagglutinins and NTNH synthesis , together with the complexation process that generates the functional forms of the toxin . Clearly it is essential that the amount and quality of experimental results is increased . In the absence of large datasets on a specific genotype , we had to construct the model from experimental data obtained from varying , though closely related , strains . Each study used a different granularity for data collection and a distinct measurement technique , which gave us the challenging task of validating a quantitative model with qualitative data . Continuing with improving the reliability of model predictions and refining the model with the inclusion of additional experimental evidence is the subject of our on-going research work .
|
Clostridium botulinum produces botulinum neurotoxins ( BoNTs ) , highly potent substances responsible for botulism . Currently , mathematical models of C . botulinum growth and toxigenesis are largely aimed at risk assessment and do not include explicit genetic information . In this paper we present modelling based on the integration of diverse information from experimental results available in the literature . Experiments show that production of BoNTs depends on the growth-phase and is under the control of positive and negative regulatory elements at the intracellular level . Here , we integrate these regulatory elements into a combined model of population dynamics and gene regulation to build the first computational model of toxin production at the molecular level . We conduct a validation of our model against several items of published experimental data for different wild type and mutant strains of C . botulinum Group I type A1 . The result of this process underscores the potential of mathematical modelling at the cellular level , as a means of creating opportunities that could be used to prevent botulism , and potentially contribute to improved methods for the production of toxin used for therapeutics .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion",
"and",
"Conclusions"
] |
[
"clostridium",
"botulinum",
"medicine",
"and",
"health",
"sciences",
"toxins",
"gut",
"bacteria",
"pathology",
"and",
"laboratory",
"medicine",
"gene",
"regulation",
"operons",
"neurotoxins",
"toxicology",
"toxic",
"agents",
"neurotoxicology",
"dna",
"bacteria",
"bacterial",
"toxins",
"gene",
"expression",
"biochemistry",
"predictive",
"toxicology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"botulinum",
"toxin",
"organisms"
] |
2016
|
An Integrative Approach to Computational Modelling of the Gene Regulatory Network Controlling Clostridium botulinum Type A1 Toxin Production
|
Neglected Tropical Diseases ( NTDs ) are important causes of morbidity , disability , and mortality among poor and vulnerable populations in several countries worldwide , including Brazil . We present the burden of NTDs in Brazil from 1990 to 2016 based on findings from the Global Burden of Diseases , Injuries , and Risk Factors Study 2016 ( GBD 2016 ) . We extracted data from GBD 2016 to assess years of life lost ( YLLs ) , years lived with disability ( YLDs ) , and disability-adjusted life-years ( DALYs ) for NTDs by sex , age group , causes , and Brazilian states , from 1990 to 2016 . We included all NTDs that were part of the priority list of the World Health Organization ( WHO ) in 2016 and that are endemic/autochthonous in Brazil . YLDs were calculated by multiplying the prevalence of sequelae multiplied by its disability weight . YLLs were estimated by multiplying each death by the reference life expectancy at each age . DALYs were computed as the sum of YLDs and YLLs . In 2016 , there were 475 , 410 DALYs ( 95% uncertainty interval [UI]: 337 , 334–679 , 482; age-standardized rate of 232 . 0 DALYs/100 , 000 population ) from the 12 selected NTDs , accounting for 0 . 8% of national all-cause DALYs . Chagas disease was the leading cause of DALYs among all NTDs , followed by schistosomiasis and dengue . The sex-age-specific NTD burden was higher among males and in the youngest and eldest ( children <1 year and those aged ≥70 years ) . The highest age-standardized DALY rates due to all NTDs combined at the state level were observed in Goiás ( 614 . 4 DALYs/100 , 000 ) , Minas Gerais ( 433 . 7 DALYs/100 , 000 ) , and Distrito Federal ( 430 . 0 DALYs/100 , 000 ) . Between 1990 and 2016 , the national age-standardized DALY rates from all NTDs decreased by 45 . 7% , with different patterns among NTD causes and Brazilian states . Most NTDs decreased in the period , with more pronounced reduction in DALY rates for onchocerciasis , lymphatic filariasis , and rabies . By contrast , age-standardized DALY rates due to dengue , visceral leishmaniasis , and trichuriasis increased substantially . Age-standardized DALY rates decreased for most Brazilian states , increasing only in the states of Amapá , Ceará , Rio Grande do Norte , and Sergipe . GBD 2016 findings show that , despite the reduction in disease burden , NTDs are still important and preventable causes of disability and premature death in Brazil . The data call for renewed and comprehensive efforts to control and prevent the NTD burden in Brazil through evidence-informed and efficient and affordable interventions . Multi-sectoral and integrated control and surveillance measures should be prioritized , considering the population groups and geographic areas with the greatest morbidity , disability , and most premature deaths due to NTDs in the country .
Neglected Tropical Diseases ( NTDs ) are a group of communicable diseases that affect predominantly the poor and vulnerable populations in about 150 countries mainly in Africa , Asia , and Latin America and the Caribbean [1] . Most NTDs are stigmatizing , disabling , debilitating , and cause poverty-promoting chronic conditions and preventable causes of death [1 , 2] . NTDs affect more than 1 . 5 billion people , and about 3 billion people are at risk to acquire one or more NTDs worldwide [2–4] . About 150 , 000–500 , 000 deaths are attributed to NTDs annually [5 , 6] . Despite being endemic mainly in low and middle income countries , their occurrence has been on the rise in high income countries , due to the increasing population mobility and migratory movements worldwide in the past decades [2 , 7] . In Brazil , most of the world’s important NTDs are present , being responsible for the majority of the burden in Latin America [8–10] . The country recorded the largest number of cases of leprosy , dengue , schistosomiasis , Chagas disease , and leishmaniases in the region [9] . NTD burden varies by Brazilian regions , with most of these diseases occurring in areas of low socioeconomic status , mainly in the North and Northeast regions [8] . About 8 , 000–10 , 000 NTD-related deaths are recorded in Brazil annually , mostly for Chagas disease [10 , 11] . However , the true NTD burden is considered to be underestimated in Brazil [8 , 10] . The Global Burden of Diseases , Injuries , and Risk Factors Study ( GBD ) is a comprehensive and updated worldwide epidemiological study that aims at quantifying the mortality , morbidity , and disability of major diseases , injuries , and risk factors by location , sex , age group , and time period [12] . GBD study uses as the main population health metric the disability-adjusted life years ( DALYs ) , a measure of health loss due to both fatal and non-fatal disease burden [12] . DALYs are estimated by summing years lived with disability ( YLDs ) and years of life lost ( YLLs ) due to premature mortality for a given cause [12–14] . GBD 2016 estimated a total of 15 million DALYs due to NTDs worldwide in 2016 ( sum of DALYs from the 15 of 18 NTDs appearing in the priority list of the World Health Organization [WHO] ) [12] . Soil-transmitted helminthiases ( STHs ) ( 3 . 3 million DALYs ) , dengue ( 3 . 0 million DALYs ) , and schistosomiasis ( 1 . 9 million DALYs ) were the main causes of burden among all NTDs in 2016 [12] . Sub-Saharan Africa ( 5 . 3 million DALYs ) and South Asia ( 4 . 1 million DALYs ) were the regions with the highest NTD burden [12] . Despite their health , economic and social impact , few systematic and comprehensive studies to quantify and compare the disease burden due to NTDs have been conducted in Brazil to date . The quantitative assessment and timely information of NTD burden in endemic areas are important to guide health policy , allocate resources appropriately , measure progress , and monitor the effectiveness and impact of health interventions and for surveillance , prevention , and disease control programs [10 , 11 , 15 , 16] . Using GBD 2016 data , we assessed the burden of NTDs in Brazil by sex , age group , and Brazilian states , from 1990 to 2016 .
Brazil , officially called the Federal Republic of Brazil , is South America’s largest country and has a total territory of 8 . 5 million km2 with an estimated population of 207 . 7 million inhabitants in 2017 [17] . The country is divided politically and administratively into 27 federative units ( 26 states and the Federal District ) and 5 , 570 municipalities , grouped into five geographic macro-regions ( South , Southeast , Central-West , North , and Northeast ) . Brazil has the highest gross domestic product ( GDP ) among Latin America countries and the ninth of the world in 2016 [18] , but remains a country with a high income inequality ( Gini index of 0 , 549 in 2017 ) [19] . Despite the considerable reduction of poverty until 2014 , about 12 . 1% of the Brazilian population were living in extreme poverty in 2016 ( proportion of people with monthly household income per capita of up to ¼ of the minimum wage ) , and with remarkable variations regions: 23 . 1% in Northeast , 22 . 7% in North , 6 . 3% in Southeast , 6 . 0% in Central-West , and 4 . 7% in South region [20] . This research has been conducted as part of the GBD study , coordinated by the Institute for Health Metrics and Evaluation ( IHME ) at the University of Washington . Data from the GBD 2016 study were used to explore the burden of NTDs in Brazil from 1990 to 2016 . Detailed description of methods and approach used in the GBD 2016 and for estimation of specific NTDs has been published elsewhere [12–14 , 21 , 22] . Briefly , GBD 2016 provides a comprehensive annual assessment of mortality and morbidity estimates for 333 diseases and injuries and 84 risk factors for 195 countries and territories from 1990 to 2016 [12–14 , 22] . The GBD 2016 cause list hierarchy is organized into four levels of causes that are mutually exclusive and collectively exhaustive [12 , 14] . Level 1 has three broad categories: communicable , maternal , neonatal , and nutritional ( CMNN ) disorders; non-communicable diseases ( NCDs ) ; and injuries . Level 2 has 21 cause groups , such as neoplasms and cardiovascular diseases . Levels 3 and 4 are disaggregated in 168 and 276 causes , respectively [12] . NTD-related causes are included in the level 2 group “Neglected tropical diseases and malaria” , which consists of 20 infectious and parasitic diseases including malaria , NTDs prioritized by the WHO ( 15 of the 18 NTDs in 2016 ) , and other neglected diseases such as yellow fever , Ebola virus disease , and Zika virus disease [12 , 14] . Most of the WHO’s priority NTDs are part of the GBD 2016 cause list , but Buruli ulcer , chikungunya virus disease ( included as “other NTDs” ) , mycetoma , and yaws are not currently estimated and available by GBD study [12 , 14 , 16] . In this study , we included GBD 2016 estimates for 12 NTD causes that are part of the official WHO priority list in 2016 and that are endemic/autochthonous in Brazil [1 , 2 , 10 , 11]: Chagas disease , leishmaniases ( visceral and cutaneous/mucocutaneous leishmaniasis ) , schistosomiasis , cysticercosis , cystic echinococcosis , lymphatic filariasis , onchocerciasis , trachoma , dengue , rabies , soil-transmitted helminthiases ( STHs ) ( or intestinal nematode infections as designated in the GBD studies: ascariasis , trichuriasis , and hookworm disease ) , and leprosy . Although fascioliasis ( one of the trematode worm infections included within the GBD category of food-borne trematodiases ) is endemic in animals and humans in some areas of Brazil , especially in southern states [23 , 24] , its burden was not estimated for the country in the GBD 2016 . NTD causes were defined and identified according to the International Classification of Diseases , 9th Revision ( ICD-9 ) and 10th Revision ( ICD-10 ) [13 , 14] . The ICD definitions and modeling strategy for each NTD cause used in this study are described in detail elsewhere [13 , 14] . In this study , we present NTD burden estimates at national ( entire country ) and subnational level ( 27 federative units [26 states and the Federal District] , herein simply named as states ) . GBD 2016 estimated cause-specific burden for the years 1990–2016 [12] . Here we focused on NTD burden estimates for 2016 , with reference to changes in the burden since 1990 . All GBD 2016 results and metrics by location and year are available at http://vizhub . healthdata . org/gbd-compare . The GBD 2016 approach to estimate all-cause and cause-specific mortality has been described previously [13 , 21] . To assemble a comprehensive cause of death database , the GBD study uses all accessible data sources that meet quality criteria and rigorous analysis , and corrects for known bias in each data source [13] . Briefly , data sources included vital registration systems , verbal autopsy data , cancer registries , surveillance data for maternal mortality , injuries , and child death , census and survey data for maternal mortality and injuries , and police records for interpersonal violence and transport injuries [13 , 21] . For Brazil , the main source of mortality data used in GBD study was the Mortality Information System ( Sistema de Informações sobre Mortalidade–SIM ) database of the Brazilian Ministry of Health , adjusted by other national and international sources [25–28] . Data corrections were made for mortality sub-registration and redistribution of garbage codes for defined causes based on the GBD 2016 redistribution algorithms [13 , 29] . Garbage codes are the assignment of causes of death that could not or should not be classified as the underlying cause of death [13] . For GBD , each death is attributed to a single underlying cause–the cause that initiated the series of events leading to death–in accordance with ICD principles [12 , 13] . GBD 2016 used the Cause of Death Ensemble model ( CODEm ) , negative binomial regression , and natural history models to estimate the number of deaths for NTD causes by location , age , sex , and year [13] . These modeling strategies for individual NTD cause of death data were described in detail elsewhere [13] . Detailed descriptions of the GBD 2016 modeling strategy for morbidity estimation and validation have been published elsewhere [14] . GBD study uses all available data that met a minimum standard of acceptable quality for each disease [14] . To measure non-fatal disease burden , GBD 2016 used epidemiological surveillance data , published and unpublished disease registries , and published scientific reports [14] . GBD 2016 used DisMod-MR 2 . 1 , a Bayesian-regression analytic tool , to synthesize consistent estimates of prevalence and incidence of non-fatal outcomes by age , sex , year , and location using a wide range of updated and standardized analytical procedures [14] . Detailed nonfatal modeling methods are described in detail elsewhere [14] . In Brazil , the main sources of morbidity data used in GBD 2016 are the morbidity national databases of the Brazilian Ministry of Health such as the Notifiable Disease Information System ( Sistema de Informação de Agravos de Notificação–SINAN ) , Hospital Information System of the Brazilian Unified Health System ( Sistema de Informações Hospitalares do Sistema Único de Saúde–SIH/SUS ) , Outpatient Information System of the Unified Health System ( Sistema de Informações Ambulatoriais do SUS–SIA/SUS ) ; specific disease databases such as the Schistosomiasis Control Program Information System ( Sistema de Informação do Programa de Controle da Esquistossomose–SISPCE ) ; national demographic and health surveys; and published scientific literature of Brazilian population-based disease prevalence studies [25 , 26] . GBD data sources for Brazil have been described in detail previously [25–28] . The input data sources and publications for each NTD in Brazil used in GBD 2016 can be accessed at http://ghdx . healthdata . org/gbd-2016/data-input-sources . DALYs were used as the measure of total burden . DALYs are estimated by the sum of the years of life lost ( YLLs ) due to premature mortality and years lived with disability ( YLDs ) for a given disease or injury [12] . One DALY represents one year of healthy life lost [12 , 30] . Detailed methods of DALY estimation have been described in previous GBD publications [12 , 30] . YLLs are calculated multiplying the number of deaths from NTDs in each age group by the standard life expectancy at each age group [12 , 13] . For GBD 2016 , the standard life expectancy at birth is 86 . 6 years , based on the lowest observed death rates for each 5-year age group in populations greater than 5 million people in 2016 [13] . YLDs are estimated by multiplying the prevalence of each sequelae or combination of sequelae from NTDs , in each age group , sex , location , and year , by their disability weights [12 , 14] . Disability weights were derived from population-based surveys of the general public [12 , 31] . Disability weight reflects the severity of health loss associated with the respective NTD on a scale varying from 0 ( perfect health ) to 1 ( equivalent to death ) [14] . For some NTDs in which death is considered a rare event , mortality was assumed to be zero . Thus , for these causes , YLLs were not calculated and DALYs were equal to the YLDs [12 , 32 , 33] . We present results in absolute numbers and age-standardized rates ( per 100 , 000 population ) of DALYs from NTDs ( individually and as a group ) by sex , age group , year , and location , with their respective 95% uncertainly intervals ( UIs ) . Age-standardized DALY rates ( per 100 , 000 population ) were computed using the GBD world population standard [13 , 14 , 21] . We report positive and negative percentage changes to show increasing and decreasing trends from 1990 to 2016 , respectively . We also present the expected estimates of NTD burden produced by GBD 2016 based on Socio-Demographic Index ( SDI ) , a composite indicator based on the geometric mean of three measures: lag-distributed income per capita , average educational attainment over aged 15 years and older , and total fertility rate [12 , 13] . Additional detail on SDI calculation and location-specific SDI values are available elsewhere [13] . This study was based on secondary databases which are publicly available , without identification of individual data . The GBD Brazil study was approved by the Ethical Review Board of the Federal University of Minas Gerais , Belo Horizonte , Brazil ( Project CAAE n° 62803316 . 7 . 0000 . 5149 ) .
In 2016 , the selected NTDs caused an estimate of 475 , 410 DALYs ( 95% UI: 337 , 334–679 , 482; age-standardized DALY rate of 231 . 98/100 , 000 population ) in Brazil , accounting for 0 . 8% of all-cause DALYs ( about 57 . 1 million DALYs ) . Table 1 presents the national estimates for the number of DALYs and age-standardized DALY rates for the selected NTDs and the percentage change between 1990 and 2016 , as well as the expected age-standardized DALY rates based on SDI in 2016 . Chagas disease was the leading cause of DALYs among all NTDs in 2016 ( 141 , 640 DALYs [95 UI: 129 , 065–155 , 941] , age-standardized DALY rate of 70 . 69/100 , 000 population [95% UI: 64 . 49–77 . 81] ) , followed by schistosomiasis ( 102 , 259 DALYs [95 UI: 59 , 767–176 , 124] , age-standardized DALY rate of 46 . 92/100 , 000 population [95%UI: 27 . 54–80 . 71] ) , and dengue ( 92 , 538 DALYs [95% UI: 63 , 477–130 , 370] , age-standardized DALY rate of 44 . 87/100 , 000 population [95% UI: 30 . 85–63 . 10] ) ( Table 1 ) . Based exclusively on SDI ( 0 . 71 for Brazil ) , observed age-standardized DALY rates for most NTDs were higher than expected , with the largest observed-to-expected ratios verified for schistosomiasis ( 428 . 4 ) , visceral leishmaniasis ( 284 . 5 ) , and trachoma ( 81 . 3 ) ( Table 1 ) . Fig 1 shows the contribution of YLDs and YLLs to total DALYs for each NTD in 2016 , and Fig 2 illustrates the proportion of DALYs , YLDs e YLLs for each disease in relation to all NTDs . In 2016 , most DALYs due to all NTDs combined were the result of the YLD component ( 52 . 5%; 249 , 636 YLDs vs . 225 , 774 YLLs ) ( Fig 1 ) , while YLLs were the main component of DALYs for all NTDs in 1990 ( 57 . 4%; 274 , 605 YLLs vs . 204 , 081 YLDs ) . The proportion of YLLs and YLDs within DALYs varied by NTD in 2016 . YLDs were the main component of DALYs for most NTDs , accounting for all DALYs for leprosy , lymphatic filariasis , onchocerciasis , trachoma , cutaneous/mucocutaneous leishmaniasis , hookworm , and trichuriasis , and contributed to more than 80% of DALYs for schistosomiasis and ascariasis ( Fig 1 ) . YLLs were the main component of DALYs for rabies ( 100 . 0% ) , visceral leishmaniasis ( 99 . 8% ) , Chagas disease ( 82 . 3% ) , and dengue ( 51 . 5% ) ( Fig 1 ) . Among all NTDs analyzed , Chagas disease accounted for the highest proportion of total DALYs ( 29 . 8% of DALYs for all NTDs combined ) , followed by schistosomiasis ( 21 . 5% ) , dengue ( 19 . 5% ) , and STHs ( 15 . 2% ) ( Fig 2 ) . Schistosomiasis contributed for most of YLDs ( 33 . 8% ) among all NTDs , followed by STHs ( 28 . 3% ) , and dengue ( 18 . 0 ) , while Chagas disease accounted for 51 . 6% of YLLs , followed by dengue ( 21 . 1% ) and visceral leishmaniasis ( 16 . 7% ) ( Fig 2 ) . The number of DALYs due to all NTDs combined remained largely unchanged as compared to 1990 ( 478 , 686 DALYs; 95% UI: 370 , 480–645 , 029 ) , with a small decline of 0 . 7% over the 27 years ( Table 1 ) . By contrast , the age-standardized DALY rates due to all NTDs decreased 45 . 7% between 1990 and 2016 ( Table 1 ) . Fig 3 shows the time trends of age-standardized DALY rates from NTD causes in Brazil from 1990 to 2016 . Absolute number of DALYs and age-standardized DALY rates for most NTD causes decreased at national level between 1990 and 2016 ( Table 1; Fig 3 ) . The most pronounced decrease in numbers and rates were observed for onchocerciasis , lymphatic filariasis , and rabies ( a reduction of 95% or more over the last 27 years ) . The age-standardized DALY rates due to Chagas disease and ascariasis decreased by 69 . 7% and 61 . 7% , respectively . In contrast , both total DALYs and age-standardized DALY rates due to dengue , visceral leishmaniasis , and trichuriasis increased substantially between 1990 and 2016 ( Table 1 ) . For some NTD causes such as leprosy , hookworm disease , cutaneous/mucocutaneous leishmaniasis , and schistosomiasis , the absolute number of DALYs increased between 1990 and 2016 . However , the age-standardized DALY rates for these causes decreased in the period ( Table 1 ) . In 2016 , the national burden of NTDs was slightly higher in males ( 51 . 9%; 246 , 843 DALYs; age-standardized DALY rate of 253 . 14/100 , 000 population ) than in females ( 48 . 1%; 228 , 567 DALYs; age-standardized DALY rate of 214 . 35/100 , 000 population ) , with a male-female DALY ratio of 1 . 2 . With the exception of cysticercosis and STHs , age-standardized DALY rates for most NTD causes were higher in males . In 2016 , the burden of NTDs spanned all age groups , with the highest proportion of DALYs in middle-aged adults ( aged 35–64 years ) , with a peak in the aged 55–59 years ( S1 Fig ) . Fig 4 shows the distribution of DALY rates due to NTDs by sex , age group , and cause in Brazil in 2016 . DALY rates due to all NTDs were higher for males across all age groups , with exception for age groups 1–9 years . The highest difference of DALY rates between men and women was observed in the age groups older than 50 years ( Fig 4 ) . The highest DALY rates ( >500 DALYs/100 , 000 population ) for both sexes combined were observed at both extremes of age spectrum ( children under 1 year and those aged 70 years and older ) . DALY rates decreased progressively from the peak of age <1 year to age 10–14 years , and then increased progressively with age and with the peak at age 95+ years ( Fig 4; S2 Fig ) . The age pattern in males and females was somewhat similar to the pattern of both sexes aggregated . However , for males DALY rates also increased progressively from children up to age 70–74 years , decreased at age 75–79 years , and then increased progressively with a peak at age 90+ years ( Fig 4 ) . The age-specific patterns varied between the NTD causes . In childhood , mainly for age 0–4 years , visceral leishmaniasis , dengue , and STHs were the main causes of DALY rates among all NTDs ( Fig 4; S2 Fig ) . After childhood , dengue and STHs were the main causes of DALYs . Schistosomiasis , dengue , Chagas disease , and STHs were the largest causes of DALY rates in young and middle-aged adults . Chagas disease , schistosomiasis , dengue , trachoma , and STHs were the most important causes of DALYs at older age groups ( Fig 4; S2 Fig ) . For the main NTD causes , DALY rates due to leishmaniasis peaked at age <1 year ( 420 . 3 DALYs/100 , 000 ) . Dengue burden peaked at age <1 year ( 120 . 9 DALYs/100 , 000 ) , and age 95 years and older ( 144 . 4 DALYs/100 , 000 ) . STH burden peaked between 5–9 years ( 37 . 7 DALYs/100 , 000 ) . Chagas disease and schistosomiasis peaked between 70–74 years ( 323 . 0 DALYs/100 , 000 and 89 . 5 DALYs/100 , 000 , respectively ) . Cysticercosis burden peaked between 65–69 years ( 14 . 0 DALYs/100 , 000 ) . Leprosy and trachoma burden peaked at age 95 years and older ( 11 . 4 DALYs/100 , 000 and 100 . 4 DALYs/100 , 000 , respectively ) ( S2 Fig ) . There was substantial geographic variation in the burden of NTDs among the Brazilian states . In 2016 , the highest age-standardized DALY rates due to all NTDs combined at the state level were observed in Goiás ( 614 . 4 DALYs/100 , 000 ) , Minas Gerais ( 433 . 7 DALYs/100 , 000 ) , and Distrito Federal ( 430 . 0 DALYs/100 , 000 ) ( Fig 5; Table 2 ) . Between 1990 and 2016 , the absolute number of DALYs for all NTDs combined presented increase for most Brazilian states , with highest increases mainly in the states of the North and Northeast regions ( Table 2 ) . By contrast , age-standardized DALY rates decreased in most Brazilian states , with the greatest declines observed in states with the highest age-standardized DALY rates in 1990 , such as Goiás and Distrito Federal ( Table 2 ) . The only states with increase in age-standardized DALY rates were Amapá , Ceará , Rio Grande do Norte , and Sergipe ( Table 2 ) . Fig 6 shows the main causes of total DALYs among NTDs by Brazilian state in 1990 and 2016 . There was a temporal variation among the main causes of NTDs by Brazilian states . In 1990 , STHs was the leading cause of total DALYs among NTDs in 16 states and Chagas disease in 11 states . In 2016 , dengue was the leading cause in 14 Brazilian states and Chagas disease in six states , while STHs ranked first only in two states ( Fig 6 ) . Fig 7 shows the ranking of age-standardized DALY rates of the specific NTD causes by Brazilian state in 2016 . Among the leading causes of NTD burden at the national level , dengue , Chagas disease , and STHs ranked among the top five NTD causes in all 27 Brazilian states . Schistosomiasis was among the top leading five NTD causes in 17 of 27 Brazilian states and leishmaniasis , in 22 states ( Fig 7 ) .
Our findings clearly show that NTDs are important causes of health loss for both men and women , but there was considerable sex difference in the burden of many NTDs . Age-standardized DALY rates for all NTDs combined and most causes were higher in males as compared to females , reflecting the patterns for most of these diseases observed in other Brazilian large-scale epidemiological studies [10 , 34–39] . Despite this pattern , the relationship between gender and infection risk is controversial , and the causes of higher male susceptibility for some NTDs is still a matter of debate [37] . The observed findings indicate gender-specific patterns of infectious disease exposure , as the relationship between gender and risk of infection is often conditioned by different socioeconomic , environmental , occupational , and behavioral factors , as well as access to healthcare services [10 , 36 , 37 , 40] . In fact , healthcare seeking behavior in Brazilian males is more often retarded , with increased morbidity and severity for some diseases [39 , 41] . The high NTD burden in children under 1 year was mainly due to the high impact of premature mortality caused by visceral leishmaniasis , confirming the well-known pattern of disease occurrence in the child population [42 , 43] . The high burden from dengue in childhood is mainly due to the increase in severe and fatal cases of the disease among children in recent years , related to the simultaneous circulation of different serotypes in several areas of the country [26 , 44 , 45] . In addition , the high burden of NTDs in more advanced age groups can be explained by the chronicity nature of major NTDs with high mortality and morbidity impact , such as Chagas disease , schistosomiasis , leprosy , trachoma , and cysticercosis [10 , 35 , 37 , 41 , 46] . The higher burden of dengue in the elderly may reflect the simultaneous occurrence of common chronic comorbidities , such as cardiovascular diseases and cancers , increasing complications , severity , and case fatality of NTDs in this age group [41 , 47 , 48] . The elderly population should receive special attention from the moment of clinical suspicion to diagnosis and treatment of these diseases [34 , 40 , 47] . Furthermore , the considerable DALY rates among younger and economically productive age groups call for further improvements in disease-specific control actions in areas with high transmission [35 , 36 , 48] . Chagas disease is responsible for most NTD deaths recorded in the country [10] . The high burden due to Chagas disease corroborates findings of previous large-scale studies in Brazil , using mortality data [10 , 11 , 39] . Other major NTDs with predominantly chronic evolution were important causes of disability and/or premature death , such as schistosomiasis , leishmaniasis , cysticercosis , STHs , and trachoma [36 , 37 , 41 , 42 , 49–51] . For some chronic NTDs , such as Chagas disease , the highest disease burden is a result of infection in previous years . These data reinforce the need to improve epidemiological surveillance , and clinical management and to ensure adequate access to the healthcare system ( diagnosis , treatment , management , and follow-up of cases ) and social support for individuals affected by these diseases [11 , 36 , 37 , 39 , 52] . STHs are important causes of disease burden throughout the national territory , mainly among the most underprivileged population groups [49 , 53] . Dengue fever is ranked as the third leading NTD cause in 2016 . Dengue is currently the NTD with the highest absolute number of new cases in Brazil , with a relatively low case fatality rate [26 , 45] . The disease has a wide geographical distribution in the country and , despite the intensification of control measures , there has been an increase in the number of severe cases , hospitalizations , and deaths in recent years [26 , 40 , 44 , 45] . This pattern is reflected by the highest proportion of YLLs in relation to total DALYs for dengue in 2016 . The steady decline at national level between 1990 and 2016 of age-standardized DALY rates for all NTDs combined may be caused mainly by the decline of the burden for the main NTDs , such as Chagas disease and schistosomiasis . These findings corroborate the observed patterns and trends for the main NTDs in recent years , and can be attributed mainly to the impact of the specific-disease surveillance and control programs implemented in the last decades [11 , 36 , 37 , 52 , 54–56] . For Chagas disease , the implementation of control measures for vector and blood-borne transmission–such as systematic entomological surveillance and screening of blood donors–reduced considerably the number of new cases and deaths in the last decades [37 , 39 , 52 , 54 , 57] . A more pronounced decline of absolute numbers and rates of DALYs due to NTDs was observed in the highly endemic states for Chagas disease in the past , such as Goiás , São Paulo , and Minas Gerais [57 , 58] . With the control of the vector domiciliary transmission of Chagas disease by its principal vector , the kissing bug Triatoma infestans , other types of transmission have become more relevant . These are directly related to the enzootic cycle of infection , such as extra-domiciliary vector transmission and domiciliary without vector colonization , and oral transmission [52 , 59] . In fact , oral transmission was the most frequent infection route of acute cases recorded in Brazil in recent years , mostly in the Amazon region [52] . However , due to its chronic nature , the challenge of recognizing chronic Chagas disease in the health services network through surveillance actions persists in Brazil [39] . Due to impact of surveillance and control program measures based mainly on large-scale treatment of risk populations in endemic areas , morbidity and mortality of schistosomiasis has been reduced , mainly in Northeast Brazil [36 , 48 , 55] . However , the wide geographical distribution of intermediate snail hosts , internal migration , tourism activities , and poor sanitary conditions still favor the persistence and expansion of disease foci [36 , 55] . There was a drastic reduction in human rabies cases in Brazil during the last decades , mainly due to systematic prevention and control activities directed to the control of urban canine rabies and post-exposure prophylaxis after aggression by suspect animals [56] . However , there are still endemic areas in which the urban cycle prevails , especially in the Northeast region [56 , 60] . At the same time , there has been observed an emergence and expansion of sylvatic transmission cycle , with increasing importance of blood-feeding bats in Brazil [56 , 60 , 61] . This highlights the need for improvement and maintenance of surveillance and control of rabies aimed at the urban cycle and its implementation in the sylvatic cycle [56] . Other factors not related to disease-specific control programs , in particular for NTDs without systematic surveillance and control programs or compulsory notification , such as cysticercosis and STHs , may have played an important role in the decline of burden for some NTDs in Brazil , such as improvements in socio-economic and sanitary conditions , increased urbanization , improved health education and access to healthcare services [8 , 9 , 11 , 41 , 49] . In contrast , the age-standardized DALY rates due to dengue and visceral leishmaniasis showed consistent increase . In fact , despite of the efforts of specific control programs , the measures to reduce transmission of these diseases has not proven to be sufficiently effective [40 , 62 , 63] , and the failures in the control of these infectious diseases favored the increase of morbidity and mortality in recent years [26 , 40 , 42 , 44] . There has been an increase in mortality and case fatality from visceral leishmaniasis in the last decades , related to the introduction of the disease in new geographic areas and unfavorable host factors , such as malnutrition , immunosuppression ( mainly due to HIV coinfection ) , and other comorbidities [42 , 64] . The large increase of dengue burden in Brazil is consistent with the wide geographical spread of the mosquito vector , and simultaneous circulation of multiple dengue virus serotypes [26 , 40 , 44] . Despite of the decrease in the age-standardized DALY rates for all NTDs combined between 1990 and 2016 , the absolute number of DALYs remained practically the same in the period . A particular trend has been observed for some causes such as leprosy and hookworm disease , with an increase in the number of DALYs , but with a decrease in age-standardized rates during the period . This pattern may suggests that the increase in absolute numbers is mainly attributable to demographic changes such as population growth and changes in population age structure [12] . GBD 2016 findings also show that the observed burden for the main NTDs , was higher than expected for the country based on SDI . This implies that the income per capita , educational levels and fertility rates were not commensurate with the high burden for some NTDs in Brazil , and despite the decrease in DALY rates , the impact for some diseases are much higher than expected for the socioeconomic development status of the country [65] . In addition , the current political and economic crisis in the country has widened poverty and social inequalities . This has potential significant negative impacts on policies and actions for health care and surveillance , as well as education and research . Together , they can increase the burden of NTDs in Brazil in the future . There was a substantial geographic variation in NTD burden in Brazil , with occurrence of health lost due to NTDs in all 27 Brazilian states . The observed geographic differences in NTD burden in Brazil are due to geographical distribution of human prevalence and incidence , vectors and/or reservoirs associated with each disease , as well as socioeconomic , demographic and environmental conditions , sanitation , quality of health surveillance , and access to healthcare services for diagnosis and treatment . These factors favor the maintenance , transmission and spread of these diseases , with consequently negative impact on morbidity , disability , and premature mortality [8 , 9 , 11] . In 2016 , with the exception of Minas Gerais ( Southeast region ) , the highest age-standardized DALY rates were observed in the states of the Central-West , North , and Northeast regions . This observed pattern reflects mainly the presence of areas highly endemic for important NTDs in the past and present , especially for Chagas disease ( Central-West and Southeast regions ) , schistosomiasis ( Northeast and Southeast regions ) , and leishmaniasis ( Northeast and North regions ) [11 , 36 , 39 , 42 , 58] . In 2016 , dengue was the predominant cause of NTD burden in 13 states located mainly in the North and Northeast regions , reflecting a marked increase of incidence , number of severe forms , and deaths from dengue , contributing to the increase in the loss of healthy years of life due to disease in recent years [26 , 45] . The states of the South region , the most socioeconomically developed region , showed the lowest age-standardized DALY rates due to NTDs in 2016 [8 , 11] . Currently , there is a global and regional effort directed to face the NTDs [2 , 66–68] . The launching of the WHO NTD roadmap and the London Declaration on NTDs contributed significantly to these global efforts [66–68] . In 2015 , NTDs have been included in the Sustainable Development Goals ( SDGs ) , with the goal 3 . 3 to end the epidemics of AIDS , tuberculosis , malaria , and NTDs and combat hepatitis , water-borne diseases , and other communicable diseases by 2030 [39 , 69] . In line with global initiatives , Brazil launched in 2012 an integrated plan of strategic action to eliminate some important NTDs , such as leprosy , filariasis , schistosomiasis and onchocerciasis as a public health problem , trachoma as an important cause of blindness , and more effective control of STHs [70] . In conjunction with states and municipalities , local plans for the elimination of these diseases should be developed throughout Brazil , promoting public health and social inclusion actions , consistent with the principles of the Brazilian Unified Health System ( Sistema Único de Saúde—SUS ) [39 , 70] . In 2016–2017 , Brazil has completed the fourth edition of this national annual campaign–about 6 million school-aged children ( 5–14 years ) had been screened in public schools of Brazilian municipalities with higher social vulnerability and high disease risk for leprosy , trachoma , and schistosomiasis , with treatment of positive cases ( including household contacts ) , and received preventive chemotherapy against STHs [71] . For effective and sustainable control of NTDs , specific control measures should be developed in conjunction with other integrated inter-sectoral public policies , such as human rights , improvements in social conditions , access to adequate water and sanitation , improved access to health care services and health education [9 , 11 , 36 , 39 , 48 , 72–74] . A higher priority should be given to research and expansion and improvement of health technologies ( drugs , vaccines , diagnostics , and control methods ) for NTDs [72] . Both financial and technical management will have to be decentralized even more to state and municipal governments , for structuring the capacity to implement interventions for surveillance , control and prevention of cases and deaths by NTDs in endemic areas [11 , 48 , 70] . There is a clear need to integrate care and attention to these conditions into the network of health care in the SUS with a high priority to primary health care [35 , 39] . Integrated access and quality of health care should also be guaranteed for the diagnosis and management of chronic comorbidities ( e . g . hypertension and diabetes mellitus ) and coinfections ( e . g . HIV/AIDS ) , since the presence of these NTD patients could aggravate the disease evolution , increasing the morbidity and mortality [26 , 47 , 75] . In addition , the implementation and sustainability of appropriate surveillance and control mechanisms and a mandatory reporting system for some important NTDs , such as cysticercosis , STHs and chronic Chagas disease , throughout the national territory could provide more accurate epidemiological data on the population prevalence and would allow geographical mapping of the affected areas [39 , 41 , 49] . In fact , the knowledge of true NTD burden is essential to track health progress , assess the impact of public health interventions , and inform evidence-based policy decisions [16] . Overall limitations of GBD 2016 study have been published in detail elsewhere [12–14 , 22] . Some specific limitations related to NTDs estimated in GBD studies were described in previous publications , such as coverage , quality , and availability of epidemiological data used to estimate the disease burden [5 , 16] . Despite considerable improvements of quality since the 1990s , mortality data differ in coverage and quality among Brazilian states , which may have cause underestimations especially in the Northeast region [10 , 37 , 46 , 76] . The GBD study used comparable and standardized methods for correction of underreporting of deaths and redistribution of garbage codes [13] . In addition , because of the lag time between mortality data reporting and the availability of databases , estimates for 2016 are mainly based on data and trends from recent years [77] . For non-fatal estimation , epidemiological data available for some NTDs , especially for diseases without national surveillance and control programs and/or mandatory reporting system such as cysticercosis , are scarce [41 , 78] . When data are of poor quality or unavailable for a location ( subnational unit , country , or region ) , GBD estimates are based on model covariates and data available from neighboring locations with a similar health profile , which may be less precise [14 , 15 , 30 , 79] . ICD coding rules allow only a single underlying cause of death and the GBD assumes that some NTD causes ( e . g . cutaneous leishmaniasis and leprosy ) have no mortality or are considered a rare event ( total DALYs for theses causes are equal to the total YLDs ) , possibly leading to underestimation [13 , 16 , 32 , 80] . The underlying causes of death may have been coded as a complication or aggravation associated with some NTDs ( such as gastrointestinal bleeding , portal hypertension , and esophageal varices for schistosomiasis , and heart failure for Chagas disease ) [10 , 11 , 16 , 36 , 47 , 48] . For some NTDs in which death is considered rare or with few records , the non-inclusion of these in the YLL calculation may substantially underestimate the total DALYs in higher endemic locations . GBD estimates are intended to only capture the direct health loss due to a specific cause in an individual [5 , 81] . They do not consider the social and economic impact and stigmatization of NTDs in the affected individuals , their families and communities [5 , 32 , 33 , 81] . Thus , the estimates of disease burden are partial measures of the impact and consequence of NTDs for the society [81] . NTDs continue being an important cause of disability and premature death in Brazil , since most diseases are preventable and/or treatable with highly efficient interventions . NTDs contribute considerably to the loss of health in individuals of all ages in all Brazilian states , with a higher burden among males , youngest ( children under 1 year ) and oldest age groups ( aged 70 years and older ) . Our findings call for renewed and comprehensive efforts to control and prevent the NTD burden in Brazil through evidence-based interventions . Integrated control and surveillance measures should focus on vulnerable population groups and geographic areas with highest NTD burden .
|
Neglected Tropical Diseases ( NTDs ) are a public health problem in Brazil . We used findings from the Global Burden of Disease Study 2016 ( GBD 2016 ) to explore the burden of NTDs in Brazil by sex , age group , specific causes , and Brazilian states from 1990 to 2016 . In 2016 , NTDs caused 475 , 410 disability-adjusted life-years ( DALYs ) ( 95% uncertainly interval [UI]: 337 , 334–679 , 482; equating to an age-standardized rate of 232 . 0 DALYs/100 , 000 population ) in Brazil . Chagas disease was the main cause of DALYs among all NTDs . The disease burden was higher among males , in the youngest and eldest ( children under 1 year and elderly aged 70 years and older ) , and in Brazilian states considered endemic for the major NTDs . There was a consistent reduction in overall age-standardized DALY rates for all NTDs combined ( overall reduction of 45 . 7% ) and most NTD causes between 1990 and 2016 , but with a pronounced increase for dengue and visceral leishmaniasis . Despite of the remarkable progress in reducing the DALY rates , NTDs remain as important preventable and neglected causes of disability and premature death in Brazil . GBD 2016 results call for intensified and comprehensive efforts to prevent and reduce the burden of NTDs in Brazil , with special emphasis on less developed areas and vulnerable populations .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"tropical",
"diseases",
"geographical",
"locations",
"parasitic",
"diseases",
"age",
"groups",
"global",
"health",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"south",
"america",
"protozoan",
"infections",
"brazil",
"people",
"and",
"places",
"helminth",
"infections",
"schistosomiasis",
"chagas",
"disease",
"population",
"groupings"
] |
2018
|
The burden of Neglected Tropical Diseases in Brazil, 1990-2016: A subnational analysis from the Global Burden of Disease Study 2016
|
Echinococcus granulosus is characterized by high intra-specific variability ( genotypes G1–G10 ) and according to the new molecular phylogeny of the genus Echinococcus , the E . granulosus complex has been divided into E . granulosus sensu stricto ( G1–G3 ) , E . equinus ( G4 ) , E . ortleppi ( G5 ) , and E . canadensis ( G6–G10 ) . The molecular characterization of E . granulosus isolates is fundamental to understand the spatio-temporal epidemiology of this complex in many endemic areas with the simultaneous occurrence of different Echinococcus species and genotypes . To simplify the genotyping of the E . granulosus complex we developed a single-tube multiplex PCR ( mPCR ) allowing three levels of discrimination: ( i ) Echinococcus genus , ( ii ) E . granulosus complex in common , and ( iii ) the specific genotype within the E . granulosus complex . The methodology was established with known DNA samples of the different strains/genotypes , confirmed on 42 already genotyped samples ( Spain: 22 and Bulgaria: 20 ) and then successfully applied on 153 unknown samples ( Tunisia: 114 , Algeria: 26 and Argentina: 13 ) . The sensitivity threshold of the mPCR was found to be 5 ng Echinoccoccus DNA in a mixture of up to 1 µg of foreign DNA and the specificity was 100% when template DNA from closely related members of the genus Taenia was used . Additionally to DNA samples , the mPCR can be carried out directly on boiled hydatid fluid or on alkaline-lysed frozen or fixed protoscoleces , thus avoiding classical DNA extractions . However , when using Echinococcus eggs obtained from fecal samples of infected dogs , the sensitivity of the mPCR was low ( <40% ) . Thus , except for copro analysis , the mPCR described here has a high potential for a worldwide application in large-scale molecular epidemiological studies on the Echinococcus genus .
Historically , four species have been recognized within the genus Echinococcus: E . multilocularis , E . oligarthrus , E . vogeli and E . granulosus [1] . E . shiquicus and E . felidis are two newly discovered additional species isolated from small Tibetan mammals and African lions , respectively [2] , [3] . Extensive research on genetic variation , intermediate host affinities as well as morphological , biological and biochemical differences resulted in a more sophisticated classification of the dog tapeworm E . granulosus into ten genotypes/strains [4]–[6]: sheep strain ( G1 ) , Tasmanian sheep strain ( G2 ) , buffalo strain ( G3 ) , horse strain ( G4 ) , cattle strain ( G5 ) , camel strain ( G6 ) , pig strain ( G7 ) , cervid strain ( G8 ) , pig/human strain ( G9 ) and Fenno-Scandian cervid strain ( G10 ) . The poorly characterized strain G9 is closely related to E . canadensis ( G7 ) [7] and the existence of G9 as a separate genotype remains still controversial [8] , [9] . More recently , new data obtained from phylogenetic analysis have shown an even more pronounced genetic divergence between these ten E . granulosus genotypes [5] , [10] . Based on sequences of the complete mitochondrial genome [11] and several nuclear markers [8] , [12] , the phylogeny for E . granulosus was reconstructed . Data obtained from nuclear protein-coding genes resulting in two nuclear alternative phylogenies: ( i ) nuclear phylogeny [8] is supported by morphological data , whereas ( ii ) nuclear phylogeny [12] is in agreement with mitogenome phylogeny [13] . Thus , E . granulosus became considered as a complex consisting of four species: E . granulosus sensu stricto ( G1/G2/G3 ) , E . equinus ( G4 ) , E . ortleppi ( G5 ) and E . canadensis ( G6–G10 ) . The phylogenetic relations within the latter group remain unresolved and are still under controversial discussion , since the E . canadensis cluster was proposed to be divided into the two species E . canadensis ( G8/G10 ) and E . intermedius ( G6/G7 ) [14] , [15] . This proposal gained further support from nuclear phylogeny [8] , but mitogenome phylogeny analyses contradicted this assumption by showing that E . canadensis ( G6/G7/G10 ) form a subgroup and E . canadensis ( G8 ) is a closely related sister taxon [16] . The adult worms of E . granulosus complex reside in the small intestine of their definitive hosts , principally wild or domestic canids . Infective eggs are shed with feces into the environment and are orally ingested by intermediate hosts where they develop into the metacestode ( larval ) stage , known as the aetiological agent of cystic echinococcosis ( CE ) in humans and predominantly ruminants , pigs and horses . Due to its success to undergo its life cycle in domesticated animals during both definitive and intermediate stages , E . granulosus constitutes an important worldwide public health problem with significant economic impact [17]–[19] . Human susceptibility to CE depends largely upon the infecting species or genotype of the E . granulosus complex . Worldwide molecular epidemiological studies revealed that E . granulosus s . s . ( G1 ) is most commonly found in humans , but also a high prevalence of E . canadensis ( G6 ) [20]–[25] and E . canadensis ( G7 ) [26] , [27] was reported . E . ortleppi ( G5 ) has a very marginal impact on human health with only two reported cases [20] , [28] . One major factor behind the worldwide spreading of many zoonoses can be the introduction of the parasite by host animals , as it happened in Australia , where E . granulosus was imported with domestic livestock about 200 years ago [29] . The worldwide distribution of CE reveals a geographic heterogeneity of E . granulosus species in many overlapping areas . Some examples are the co-existing genotypes E . granulosus s . s . ( G1 ) and E . canadensis ( G6 ) in North African countries [23] , [30]–[32] , E . granulosus s . s . ( G1/G2 ) , E . ortleppi ( G5 ) and E . canadensis ( G6/G7 ) in Argentina [20] , [33] , [34] or E . granulosus s . s . ( G1 ) , E . canadensis ( G6 ) and E . equinus ( G4 ) in Kyrgystan [35] . In these areas co-infections with more than one E . granulosus species/genotype might occur in the intermediate or definitive hosts . In addition , the not yet confirmed hypothesis of an eventual genetic exchange by sexual reproduction between E . granulosus species/genotypes is still discussed [36] . The knowledge about the distribution of the E . granulosus complex is important e . g . in the context of any control or eradication program . Thus , regular molecular epidemiological surveys provide key information on the spatio-temporal dynamics of parasite populations . Knowledge about the transmission and prevalence of E . granulosus in humans and animals , including dogs , is a basic step before and during control and/or surveillance strategies . Different methods for genotyping genetic variants of the E . granulosus complex have been developed so far . Based on PCR amplified sequences of the mitochondrial cytochrome c oxidase subunit 1 ( cox1 ) or the NADH dehydrogenase subunit 1 ( nad1 ) , genotyping can be performed in a relative time and/or cost intensive way by sequencing [37] , RFLP ( Restriction Fragment Length Polymorphism ) [38] , [39] , fingerprinting [40] or SSCP ( Single Strand Conformation Polymorphism ) [41] . More recently , pure PCR based methods that simplify the genotyping have been designed . With a consecutive PCR approach a part of the E . granulosus complex ( G1 , G5 , G6/G7 ) can be genotyped [42] and by applying four parallel PCRs the discrimination between E . multilocularis , E . granulosus s . s . ( G1 ) and an E . ortleppi ( G5 ) /E . canadensis ( G6/G7 ) cluster is possible [43] . Parallel PCR approaches can be combined in a multiplex PCR setup and became rapidly and successfully applied worldwide in many aspects of DNA analyses , especially in the field of molecular diagnosis of infectious diseases such as bacterial [44] , viral [45] and fungal [46] infections . For cestode infections , a 3-plex-PCR approach was already established to distinguish between E . multilocularis , E . granulosus complex and Taenia [47] . However , the potential of such an approach has not yet been evaluated for the specific detection and/or genotyping of different isolates within the E . granulosus complex . Based on the identification of a number of discriminating polymorphism sites in nuclear and mitochondrial genes of the Echinococcus genus , we established a single-tube multiplex PCR ( mPCR ) approach that allows a rapid and simultaneous detection and discrimination among the following members of the E . granulosus complex: E . granulosus s . s . ( G1/G2/G3 ) , E . equinus ( G4 ) , E . ortleppi ( G5 ) , E . canadensis ( G6/G7 ) and E . canadensis ( G8/G10 ) . We assessed the performance of the mPCR assay by re-identifying reference DNA panels ( 42 samples ) and by genotyping 153 unknown DNAs from human and animal Echinococcus cyst samples isolated from infected intermediate hosts in Tunisia , Algeria and Argentina . Finally , we assessed the feasibility of applying mPCR for the detection and genotyping of E . granulosus complex in fecal egg samples , and directly in frozen or fixed parasite material ( hydatid fluid or protoscoleces ) .
Based on known mitochondrial or nuclear DNA sequences , polymorphisms between Echinococcus strains/genotypes were identified and used for strain/genotype specific primer design . Each primer pair was first applied on its respective genotype-specific DNA , and if one clear PCR product was amplified , it was applied on DNA samples of all other genotypes/strains in order to exclude non-specific amplicons . Finally , 11 primer-pairs resulting in genotype/strain/genus specific targets were used for the mPCR . The mPCR was set up with normalized known template DNAs in a sequential approach by starting with one specific primer pair in the PCR mix , followed by the incorporation of other primer pairs . The PCR was run with every additional new primer pair on all genotype/strain specific DNA samples to confirm specificity . Simultaneously the molar amount of primers was adjusted in order to achieve comparable amplicon intensities . To reduce variable parameters and to allow comparison between experiments the basic mPCR conditions using GoTaq DNA polymerase from Promega were defined as followed: 94°C for 3 min , 25 cycles of 94°C for 30 sec , 56°C for 30 sec , 72°C for 30 sec and a final extension step for 5 min at 72°C . With this setup the sensitivity range was determined by adding different amounts of template DNA into the mPCR mix . The specificity of the mPCR was tested by ( i ) adding more PCR cycles , ( ii ) using mixed DNA templates derived from different Echinococcus genotypes/strains , ( iii ) using template DNAs of closely related genus Taenia or ( iv ) by the addition of foreign DNA derived from bovine thymus or dog feces . To exclude lab-specific conditions , 13 samples were genotyped by mPCR in two different laboratories . To assess potential problems with materials derived from different suppliers , the system was tested with DNA polymerases from different companies . The mPCR performance was further validated by genotyping 42 E . granulosus complex samples derived from known origin and genotype , and subsequently 153 unknown DNA samples were genotyped . Furthermore , the mPCR was assessed on DNA derived from Echinococcus eggs isolated from feces of infected dogs . Finally , approaches were developed to perform the mPCR directly on fresh protoscoleces , either frozen or fixed , or on hydatid fluid . Information on the complete mitochondrial genome sequences containing the genes cytochrome oxidase subunit I ( cox1 ) , cytochrome oxidase subunit 2 ( cox2 ) , ATP synthase subunit 6 ( atp6 ) and NADH dehydrogenase subunit I ( nad1 ) as well as mRNA sequences of the nuclear genes RNA polymerase II ( rpb2 ) , DNA polymerase delta ( pold ) , ezrin-radixin-moesin-like protein ( elp ) , elongation factor 1 alpha ( el1a ) and calreticulin ( cal ) were obtained from the databases of the National Center of Biotechnology Information ( NCBI ) for E . granulosus s . s . ( G1/G2/G3 ) , E . equinus ( G4 ) , E . ortleppi ( G5 ) , E . canadensis ( G6/G7 ) , E . canadensis ( G8/G10 ) , E . multilocularis , E . vogeli and E . oligarthrus . The respective sequences were retrieved via GenBank [http://www . ncbi . nlm . nih . gov/] and were aligned with BioEdit 7 . 0 . 9 to detect polymorphic sites . The accession numbers of the used Echinococcus sequences are listed at the end of the manuscript . The primers were designed on the assumption that one specific 3′-base will be sufficient to result in genotype-specific amplification since Taq-polymerases lack a 3′–5′ proofreading activity . In consequence , primers were chosen such as to strain-specifically bind to the targets described above . If possible , primers were selected that contained more than one specific 3′-base , but five primers of the final set that were targeted to nuclear sequences matched this one base difference . Because genotyping based on single nucleotide polymorphisms ( snips ) is error-prone due to mutations [48] , [49] , we chose two genotype/strain-specific probes for all E . granulosus complex members . The exception was E . canadensis ( G8/G10 ) , where only one probe was selected due to its rare occurrence and close relationship to E . canadensis ( G6/G7 ) . Two additional probes were chosen: a common one for all E . granulosus complex members , and one for the overall detection of all known Echinococcus species: E . granulosus complex , E . multilocularis , E . vogeli , E . oligarthrus and E . shiquicus . Therefore , three levels of differentiation were obtained for each sample by determination of ( i ) the genus Echinococcus , ( ii ) the affiliation to the E . granulosus complex and ( iii ) the specific strain or genotype within the complex . For all primers , a Tm of approximately 55°C was selected , and for each primer-pair a PCR product of distinct size was anticipated , in order for the amplicons to be easily discriminated by 2% agarose gel-electrophoresis . Table 1 shows the complete list of the final 22 primers used in this study , including names , molar concentrations in the mPCR mix , the final product sizes , the specificities ( genotypes ) , the primer sequences ( including the polymorphic sites ) , the primer lengths , the target genes ( gene marker ) , the accession numbers of the published DNA target sequences , and the corresponding positioning of the primer sequences within their targets . The reaction mix for the final mPCR was composed of 100 µM dNTPs and 0 . 05 units µl−1 GoTaq DNA polymerase in 1× PCR Buffer ( all Promega ) and contained the 22 primers specific for 11 targets in the molarities shown in Table 1 . For standard genotyping 5 ng template DNA were added into the PCR mix . Each reaction was performed in single tubes in a volume of 20 µl PCR mix . The cycling conditions were as follows: an initial denaturation step at 94°C for 3 min , 25 cycles ( 94°C–30 s , 56°C–30 s , 72°C–1 min ) and a final extension step lasting 5 min at 72°C . 10 µl of the PCRs were separated by electrophoresis in a 2% agarose gel and visualized by ethidium bromide staining and subsequent UV excitation . The genotype specific amplicon profile is shown in Figure 1 . The mPCR conditions were a result of pre-experiments described below , and these conditions were used throughout if not indicated otherwise . Ethical statement: For the parasite samples of animal origin , these were taken from animals in abattoirs being processed as part of the normal work of the abattoirs , in the frame of conventional meat inspection . For the parasite samples of human origin , these were obtained for and thus part of the normal diagnostic investigation to determine the etiology of the biopsied tissue for clinical purpose . Thus the present investigation was part of the conventional diagnostic procedure used in clinical practice . Samples were all anonymized for carrying out data evaluation . ( A ) For establishment of the mPCR and all evaluations concerning the sensitivity and the specificity of the method , a test panel of E . granulosus complex chromosomal DNAs was used . Genomic DNA specimens used for the test panel were: E . granulosus s . s . ( G1 ) , E . equinus ( G4 ) , E . canadensis ( G6 ) , E . canadensis ( G7 ) , and E . canadensis ( G8 ) . These were obtained from institutional DNA-collections in Berne/Switzerland , Zürich/Switzerland and Tartu/Estonia . Genomic DNA extracted from E . ortleppi ( G5 ) was kindly provided by Dr . Karen Haag ( Departamento de Genétic , Instituto de Biociências , Universidade Federal do Rio Grande do Sul/Brazil ) and protoscoleces from E . canadensis ( G10 ) were kindly provided by Prof . Thomas Romig ( Institute of Parasitology , University of Hohenheim/Germany ) . All samples had been genotyped conventionally by sequencing cox1 and/or nad1 . The genomic DNA of the E . canadensis ( G10 ) protoscoleces was isolated using a standard phenol-chloroform protocol [50] , using RNAse A ( Sigma-Aldrich ) , Proteinase K ( Sigma-Aldrich ) and a subsequent isopropanol precipitation followed by multiple washes in 75% EtOH prior to drying and dissolving in ddH2O . For most genotyped samples used in these parts of the study , the original extraction method for genomic DNA could not be retrospectively determined . A general problem in the usage of genomic DNA prepared by multiple methods ( e . g . column based nucleic acid purification , phenol/chloroform extraction , presence or absence of RNAseA or proteinase K treatment ) arises when quantifying the DNA concentration , e . g . by Nanodrop ND-1000 measurement . Therefore , an E . granulosus s . s . ( G1 ) DNA amount ( selected upon the most intense PCR amplification product when using the Echinococcus specific primers Echi-Rpb2 F and Echi-Rpb2 R , 1 µM , see Table 1 ) , was defined as a reference measure point . The DNAs of all other species/genotypes were normalized to this sample by comparative PCR using the same primers . The PCRs were performed under the following conditions: 94°C for 3 min followed by 25 cycles of 94°C for 30 s , 56°C for 30 s and 72°C for 1 min and a final extension step of 5 min at 72°C . ( B ) For the evaluation of specificity in the context of cross binding of the primers , DNA derived from Echinococcus species outside of the E . granulosus complex ( E . multilocularis and E . vogeli ) as well as DNA of the closely related Taenia saginata , T . solium , T . crassiceps , T . taeniaformis and T . pisiformis were obtained from the institutional DNA-collection at the University of Berne/Switzerland . ( C ) For the evaluation of specificity in the context of contaminating DNA , bovine thymus DNA was obtained commercially from Serva , and dog feces DNA was isolated as described above by phenol/chloroform extraction from feces of a helminth-free dog that was obtained from the Small Animal Clinic of the Vetsuisse Faculty , University of Berne , Switzerland . ( D ) For the assessment of the mPCR genotyping performance on DNA derived from metacestodes and/or protoscoleces , two panels of known ( reference ) and unknown Echinococcus metacestode DNAs were used . Known/genotyped materials were 20 reference DNA samples originating from Bulgaria [51] and 22 samples from Spain ( unpublished ) obtained from the institutional DNA-collection at the University of Berne/Switzerland . Unknown/non-genotyped materials were 13 DNA samples harvested from slaughterhouses in Buenos Aires/Argentina . Protoscoleces fixed in 95% ( v/v ) ethanol were obtained from 101 animal cysts harvested from slaughterhouses in Tunisia ( 75 samples ) and Algeria ( 26 samples ) . Human isolates were collected after surgery from human patients in Tunisia ( 39 samples ) . Chromosomal DNA was prepared as described above . For more detailed information e . g . on host animal species , see Table 2 . A part of these samples were used for the reliability and reproducibility tests . These 66 samples are marked with an asterisk in Table 2 . ( E ) For the assessment of the mPCR genotyping performance of feces , eggs were isolated according to Mathis et al . [52] from 28 dog fecal samples ( Sample collection Zürich/Switzerland: 20 samples from a study in Kyrgyzstan [35] and 8 samples from a study in Lithuania [53] ) . DNA extraction was performed as previously described [54] , and the DNA was characterized by a multiplex PCR for the simultaneous detection of E . granulosus ( G1–G10 ) , E . multilocularis and Taenia spp . [47] . Echinococcus was identified in all samples; 18 out of 28 with E . granulosus ( 10 from Kyrgyzstan , mainly sheep strain G1 ) and 8 from Lithuania where only E . canadensis G7 occurs and 10 with E . multilocularis ( 10 from Kyrgyzstan ) . These preselected samples were used to assess the potential of the mPCR as a molecular diagnosis tool for canine infection with adult Echinococcus . ( F ) To evaluate the mPCR directly on parasite material , none genotyped Echinococcus samples obtained from the institutional sample collection of Berne/Switzerland were used: ( i ) frozen hydatid fluid , ( stored at −20°C ) and ( ii ) solid E . granulosus complex germinal layers and protoscoleces , used natively ( frozen ) or fixed in either 95% ( v/v ) ethanol or 4% PBS-buffered formaldehyde solution . The mPCR conditions described above were a result of 3 preliminary sets of experiments . Used samples are described above in sample section A . To specify the amount of template DNA which can be used in the mPCR , the sensitivity of the method was determined by varying the template concentrations of normalized test panel E . granulosus complex DNAs in the standard mPCR mix containing all 22 primers . Therefore 0 . 1 , 0 . 5 , 1 , 2 . 5 , 5 , 10 , 25 , 50 , 100 , 250 , 500 ng and 1 µg normalized DNA from E . granulosus s . s . ( G1 ) , E . equinus ( G4 ) , E . ortleppi ( G5 ) , E . canadensis ( G6 ) , E . canadensis ( G7 ) , E . canadensis ( G8 ) or E . canadensis ( G10 ) were tested individually by mPCR employing the conditions described above ( sample origin is described in sample section A ) . For the readout of this experiment low amounts of the different template DNAs had to result in clearly visible bands and high amounts of template should not yield additional or smeary products . With these preconditions/definitions a usable template range resulting in clear genotyping patterns was determined . To test the influence of additional PCR cycles ( more than 25 ) , the mPCR was performed individually with 5 and 250 ng template DNA of the different Echinococcus strains ( see sample section A ) . The mPCRs were run with 25 , 30 and 35 amplification cycles and after gel electrophoresis the amplicons were screened for smeary or unspecific products to detect the cycle number range which resulted in clear genotyping patterns . To determine the detection limit of a specific E . granulosus complex strain in a dual-strain DNA mixture , normalized test panel DNA from E . granulosus s . s . ( G1 ) and E . canadensis ( G6 ) were mixed and applied in the standard mPCR in total amounts of 5 ng ( ratios; 80∶20 , 60∶40 and 50∶50 ) , 50 ng ( ratios; 97 . 5∶2 . 5 , 95∶5 , 90∶10 and 80∶20 ) and 250 ng ( ratios; 99 . 37∶0 . 63 , 98 . 75∶1 . 25 , 97 . 5∶2 . 5 , 95∶5 and 90∶10 ) . Samples are described in sample section A . For the readout , clearly visible amplicons of the E . granulosus complex DNA applied in lower ratios indicated a successful detection . Depending on the applied template amount , different ratios were detected . Additionally a DNA cocktail containing 5 ng of normalized test panel DNA from each member of the E . granulosus complex was used as template for the mPCR to verify that all 11 targets could be amplified simultaneously in one tube . To exclude unspecific cross binding of the primers on the closely related Taenia genus , 10 ng template DNA derived from different Taenia species were applied in individually performed standard mPCRs . The samples employed for assessment of cross-binding are described above in sample section B . To assess the mPCR specificity in the presence of host-derived contaminations in individually performed mPCRs , 5 ng of normalized E . granulosus s . s . ( G1 ) test panel DNA ( sample section A ) were mixed with different amounts ( 1∶1 up to 1∶200 ) of two types of foreign DNA ( sample section C ) . Clearly visible specific amplicons combined with a lack of unspecific PCR products indicated successful genotyping . To confirm the reliability of the mPCR , a set of 66 samples ( sample section D ) were genotyped first according to the PCR-sequencing technique described by Bowles et al . 1992 [37] , using the cox1 primers JB3 ( 5′-TTTTTTGGGCATCCTGAGGTTTAT-3′ ) and JB4 . 5 ( 5′-TAAAGAAAGAACATAATGAAAATG-3′ ) . The PCR products were purified with the High Pure PCR Product Purification kit ( Roche Applied Science ) according to the manufacturer's instructions and subsequently sequenced using an automated DNA sequencer ( Applied Biosystems , ABI 3130× I Genetic Analyzer Sequencer ) . Sequence data were analyzed and compared with existing sequences derived from GenBank [http://www . ncbi . nlm . nih . gov/] . In a second step these 66 samples were used as templates in the standard mPCR setup using ∼20 ng DNA . Finally the results of both genotyping approaches were compared . The reproducibility of the mPCR was assessed by performing the test in two different qualified laboratories and using the same mPCR protocol and test samples ( see sample section D ) . Therefore , 13 samples from Argentina were genotyped in parallel by mPCR in the laboratories of Berne/Switzerland and Buenos Aires/Argentina . The mPCRs were performed with 20 ng template DNA as described above and the results were compared between the laboratories . Additionally , all 13 samples were genotyped by cox1 sequencing ( see above ) . Since the mPCR was set up with GoTaq DNA polymerase from Promega and the DNA polymerases from different suppliers can influence the mPCR performance , a panel of DNA polymerases was tested in a second reproducibility test by replacing the GoTaq polymerase and GoTaq PCR buffer by other products in the standard mPCR setup . For the mPCR , 5 ng of normalized E . granulosus s . s . ( G1 ) template DNA was used ( Sample section A ) . DNA polymerase systems , which clearly yielded the 4 expected products , were designated as “useful” and the others yielding unspecific products , smears or missing amplicons were designated as “needing optimization” . The tested DNA polymerases and the performance results are listed in Table S1 . In total 195 E . granulosus complex DNA samples were tested . The DNA concentrations in all metacestode derived samples were measured and 1 µl ( ∼20 ng ) of the DNA samples was used as template . The mPCR was performed with the standard settings described above . Information on the samples tested is given above in sample section D . In order to investigate whether the mPCR is suitable as a molecular diagnostic tool to detect Echinococcus eggs in canine fecal samples , a panel of positively preselected DNA samples prepared from Echinococcus eggs was investigated . Since contaminating DNA can be present , 2 µl of the DNA samples ( 150–350 ng DNA ) were used for mPCR , which was first performed under standard conditions as described above , and subsequently with 35 instead of 25 cycles and with up to 1 µg of template DNA per reaction . Information on the samples is given above in sample section E . To simplify the genotyping procedure , we elaborated protocols that allow omitting DNA extraction procedures for mPCR amplification by using frozen or fixed E . granulosus materials ( Sample section F ) . Many Echinococcus samples contain high amounts of calcium corpuscles that could interfere with the mPCR . These calcium corpuscles form a relatively solid pellet at the lowest bottom of the tube after centrifugation and by using the upper cellular part of the pellet a carry-over can be avoided . Frozen hydatid fluid ( HF ) ( stored at −20°C ) was thawed at room temperature and 1 ml was heated to 100°C for 30 min , centrifuged at 13 , 000 rpm for 10 min , and different volumes ( 0 . 25 , 0 . 5 , 1 , 1 . 5 , 2 , 2 . 5 , 3 , 10 µl ) of the resulting supernatant were used as templates for mPCR . Additionally , 1 and 2 µl none heated HF were applied in the mPCR . Solid E . granulosus complex germinal layers ( cut into small pieces ) and protoscoleces were used either natively ( frozen ) or fixed , either in 95% ( v/v ) ethanol or 4% PBS-buffered formaldehyde solution . The material was prepared either by boiling or by alkaline lysis . In both cases , frozen material was used directly , and fixed material was pre-washed twice with PBS . For the preparation of the material by boiling , 10 µl solid sedimented Echinococcus material was resuspended in 90 µl H2O and incubated in a shaking heater ( 1 , 200 rpm , 100°C ) for 30 min . Shaking is important in this step and if no shaking heater is available , the samples have to be vortexed from time to time , or must be intensively resuspended by pipetting . After centrifugation at 13 , 000 rpm for 10 min , different volumes ( 0 . 25 , 0 . 5 , 1 , 1 . 5 , 2 , 2 . 5 , 3 , 10 µl ) of the supernatant were used for the mPCR . For alkaline lysis , 10 µl solid Echinococcus material was incubated in 50 µl of 0 . 4 M NaOH and 2 µl of 1 M dithiothreitol ( Sigma ) and the mixture was heated for 15 min at 65°C in a shaking heater ( 1200 rpm ) . The suspension was neutralized by adding 50 µl of 0 . 4 M HCl and 1 µl 1 . 5 M Tris-HCl pH 8 , and centrifuged for 10 min at 13 , 000 g . Shaking is important at this step ( see above ) . For the mPCRs , 2 µl of different supernatant dilutions ( 1∶1 , 1∶2 , 1∶4 , 1∶6 , 1∶8 , 1∶10 and 1∶25 ) were used in 20 µl setups . Furthermore , 1 and 2 µl undiluted supernatant were applied in the mPCR .
The mitochondrial genome and different nuclear genes were aligned and analyzed for sequence differences appearing specifically within in the genes of the individual E . granulosus complex members: E . granulosus s . s . ( G1/G2/G3 ) , E . ortleppi ( G4 ) , E . equinus ( G5 ) , E . canadensis ( G6/G7 ) and E . canadensis ( G8/G10 ) . Specific primer-pairs were designed and tested individually for sensitivity and specificity . In these preliminary experiments , primer concentrations were 0 . 5 µM , but template DNA amounts varied between 10 pg and 5 ng , and different numbers of amplification cycles ( 25 , 30 or 40 ) were assessed . Primer pairs yielding specific and clear PCR products were combined to a set of 22 primers , which allowed the amplification of 11 different size-specific PCR products . This set of primers was used for the mPCR and the final concentrations of the primers in the mPCR mix were adjusted in order to achieve similar amplicon quantities . In this optimization step , 5 ng template and 25 amplification cycles were used , because by keeping the template DNA amount and amplification cycle numbers constant , the procedure for optimization of the final mPCR primer concentrations was simplified . In addition , keeping the numbers of cycles low reduced non-specific amplification and would speed up the procedure . The results of the single primer-pair tests that might be used for specific single primer-pair PCRs are depicted in Table S2 and all information about the chosen primers and their final concentrations used in the mPCR are shown in Table 1 . These pre-experiments resulted in a standard setup for the mPCR , which applies 22 primers at different concentrations . The mPCR was performed with GoTaq DNA polymerase in a final reaction volume of 20 µl and 25 amplification cycles . As template , 5 ng of normalized DNA of the different E . granulosus species were used . All reactions yielded a highly specific and clearly distinguishable banding pattern ( Figure 1 A and B ) , allowing the discrimination among E . granulosus s . s . ( G1 ) , E . equinus ( G4 ) , E . ortleppi ( G5 ) , E . canadensis ( G6/G7 ) and E . canadensis ( G8/G10 ) . The smallest band ( 110 bp ) was designated to specifically indicate all members of the E . granulosus complex and was clearly present in all 5 species . The upper band ( 1232 bp ) specifically identified the genus Echinococcus ( Figure 1B ) and detected the E . granulosus complex as well as E . multilocularis and E . vogeli . The sensitivity of the mPCR was investigated by applying different concentrations of E . granulosus complex template DNA ( 0 . 1 ng–1 µg ) , and the results showed 5–250 ng template DNA are required for a successful detection of all members . When lower or higher amounts of DNA were employed , some PCR products were missing or non-specific amplification occurred . Out of the recommended amounts of template DNA , the detection limits depend largely on the species; E . granulosus s . s . ( 0 . 1 ng–1 µg ) , E . equinus ( 2 . 5 ng–250 ng ) , E . ortleppi ( 0 . 5 ng–250 ng ) , E . canadensis ( G6/G7 ) ( 1 ng–500 ng ) and E . canadensis ( G8/G10 ) ( 5 ng–250 ng ) . Thus , in several experiments lower amounts of DNA ( 0 . 1–5 ng ) were sufficient , but this occurred only when DNA of high quality was used ( Figure 3A ) . The specificity of the mPCR assay was investigated in four ways: ( i ) increasing the numbers of PCR cycles; ( ii ) employing mixed template DNA derived from different Echinococcus genotypes/strains; ( iii ) applying template DNAs of the closely related genus Taenia; ( iv ) addition of non-related DNA derived from bovine thymus or dog feces . Increasing the cycle numbers had an influence on the specificity of the mPCR . In the case where up to 250 ng normalized template DNA was applied , a specific banding pattern was achieved at 25 amplification cycles , but as shown in Figure 2 , increased numbers of cycles still allowed genotyping based on the most prominent bands . However , in some genotypes , application of 30 cycles or more resulted in smeary or unspecific amplicons . Thus , for mPCR 25 amplification cycles are recommended . To test the specificity of the mPCR with mixed template DNA derived from different Echinococcus species , two experiments were performed . First , a DNA cocktail containing 5 ng of normalized DNA from each member of the E . granulosus complex was used as template for the mPCR , and this resulted in a clear and simultaneous expression of all specific amplicons . Additionally , this experiment showed that all specific PCR products could be amplified in parallel , without interference or non-specific amplification ( Figure 3B ) . Since E . granulosus s . s . ( G1/G2/G3 ) and E . canadensis ( G6/G7 ) have been reported to co-exist in several areas , these two species were selected to determine the detection limit of a specific genotype in a dual-genotype DNA mixture . Thus , DNA from E . granulosus s . s . ( G1 ) and E . canadensis ( G6 ) were mixed in different ratios and analyzed by mPCR . When 5 ng of the mixed DNA was used as template , one genotype could be detected when it was present in a concentration of 20% ( Fig . 3A , lanes 11 and 15 ) . By using 50 ng template DNA one genotype was detectable in a concentration of 2 . 5% ( Fig . 3A , lane 6 ) and if 250 ng template DNA were used , the detection of one genotype was possible at a concentration of 5% ( Fig . 3A , lane 4 ) . Both experiments showed that two or more genotypes can be detected in parallel by mPCR . To test the cross-reactivity with closely related Taenia species , mPCRs were performed with 10 ng template DNA of T . saginata , T . solium , T . crassiceps , T . taeniaformis and T . pisiformis . As shown in Figure 1C ( lanes 4–8 ) no products indicative for non-specific primer binding were amplified . To mimic contaminations occurring during the isolation of DNA from metacestodes or E . multilocularis eggs , 5 ng normalized E . granulosus s . s . ( G1 ) DNA and different amounts of DNA from bovine thymus or canine feces were mixed with at different rations ( 1∶1–1∶200 ) . As shown in Figure 4 , mPCR tolerated a 200-fold excess of foreign DNA ( Figure 4 ) . To test the reliability of the mPCR , 66 unknown samples were genotyped by cox1-sequencing [37] and mPCR in parallel and both methods obtained identical results ( Table 2 , used samples marked with an asterisk ) . The interlaboratory reproducibility of the mPCR was evaluated by genotyping 13 samples in parallel , namely in Berne/Switzerland and Buenos Aires/Argentina , respectively . Both laboratories employed GoTaq DNA polymerase , but otherwise worked independently from each other . Identical results were obtained; seven of the samples contained E . canadensis ( G6/G7 ) and six contained E . granulosus s . s . ( G1/G2/G3 ) isolates ( data not shown ) . In order to investigate whether the type of DNA polymerase used in mPCR could influence the results , a panel of DNA polymerases derived from different suppliers was tested . The GoTaq polymerase ( Promega ) originally used for the development of the mPCR yielded optimal results . However , similar results were obtained employing the 5× Multiplex PCR mix from New England Biolabs as well as AmpliTaq DNA Polymerase from Applied Biosystems . Other DNA polymerases failed to provide useful results , leading to non-specific amplicons , smeary products or missing amplification . A list showing the tested DNA polymerases is depicted in Table S1 . The newly established mPCR was applied on previously characterized metacestode DNA , and on metacestode DNA samples of unknown origin . A total of 195 hydatid cysts , 149 isolated from animals and 46 obtained in humans , and all originating from different regions and/or continents , were genotyped by mPCR ( for details on the samples , see Table 2 ) . The mPCR amplified the corresponding genotype-specific banding patterns , and in no case unspecific amplicons or mixed genotypes were detected ( data not shown ) . All 46 human CE cases and 135 of the 149 animal CE cases clustered within E . granulosus s . s . ( G1/G2/G3 ) . Furthermore , the mPCR detected 7 European E . equinus ( G4 ) cases isolated from 6 horses and 1 donkey from Spain , and 7 pig-derived E . canadensis ( G7 ) cases from South American samples ( Table 2 ) . In this experiment 28 preselected Echinococcus egg DNA samples extracted from dog feces were used: 10 E . granulosus s . s . ( G1 ) , 8 E . canadensis ( G7 ) and 10 E . multilocularis samples . Employing mPCR and 150–350 ng template DNA , only 5 out of 10 E . granulosus s . s . ( G1 ) samples , 0 out of 8 E . canadensis ( G7 ) samples , and 4 out of 10 E . multilocularis samples could be positively identified . Increasing the number of amplification cycles up to 35 and/or employing increased amounts of template DNA ( up to 1 µg ) did not result in any improvement ( data not shown ) . In order to avoid time-consuming DNA extraction steps , the mPCR was performed directly on hydatid fluid ( HF ) and protoscoleces ( Figure 5 ) . The mPCR failed when these samples were used directly without any pre-treatment . However , heating HF followed by centrifugation and subsequent mPCR with 1–3 µl of the supernatant resulted in amplification of the entire E . equinus ( G4 ) specific banding profile . Inclusion of lower or higher amounts of boiled HF supernatant , or inclusion of fresh , frozen or fixed Echinococcus tissue , did not result in mPCR amplification products ( Figure 5A; data not shown ) . However , preparation of the material employing an alkaline lysis protocol resulted in effective genotyping with frozen and/or EtOH fixed samples , but not with protoscoleces fixed in 4% formaldehyde . When 2 µl of a 1∶8 or 1∶10 dilution of the alkaline lysed supernatant derived from frozen protoscoleces was used for mPCR the whole E . granulosus s . s . ( G1 ) specific banding pattern was detected ( Figure 5B ) . Application of 2 µl of a 1∶2 or a 1∶4 supernatant dilution of EtOH fixed protoscoleces resulted in the detection of a clearly amplified E . granulosus s . s . profile ( Figure 5C ) . Conditions outside of these ranges yielded incomplete or lacking amplification of specific targets . It should be noticed at this point that calcium corpuscles interfere with the PCR . Best results were achieved when calcium corpuscles present at the bottom of the tube after centrifugation of the solid Echinococcus material were not included in the boiling or alkaline lysis steps .
A ) The primers Echi Rpb2 F and Echi Rpb2 R used for the detection of all Echinococcus species were designed using the Echinococcus gene RNA polymerase II ( rpb2 ) : E . granulosus s . s . ( G1/G2/G3 ) - FN566850 . 1 , E . equinus ( G4 ) - FN566851 . 1 , E . ortleppi ( G5 ) - FN566852 . 1 , E . canadensis ( G6 ) - FN566853 . 1 , E . canadensis ( G7 ) - FN566854 . 1 , E . canadensis ( G8 ) - FN566855 . 1 , E . oligarthrus - FN658827 . 1 , E . vogeli - FN566847 . 1 , E . multilocularis - FN566845 . 1 . B ) The complete mitochondrial genome sequence was used to design the E . granulosus complex specific primers E . g . complex F and E . g . complex R ( gene marker: cox2 ) , the E . ortleppi ( G5 ) specific primers E . ortp ATP6 F and E . ortp ATP6 R ( gene marker: atp-6 ) as well as E . ortp CoxI F and E . ortp CoxI R ( gene marker: cox1 ) and the E . canadensis ( G6/7 ) specific primers E . cnd G6/G7 NDI F and E . cnd G6/G7 NDI R ( gene marker: nad1 ) : E . granulosus s . s . ( G1/G2/G3 ) - AF297617 . 1 , E . equinus ( G4 ) - AF346403 . 1 , E . ortleppi ( G5 ) - AF235846 . 1 , E . canadensis ( G6 ) - AB208063 . 1 , E . canadensis ( G7 ) - AB235847 . 1 , E . canadensis ( G8 ) - AB235848 . 1 . C ) The ezrin-radixin-moesin-like protein ( elp1 ) was used to design the E . canadensis ( G8/G10 ) specific primers E . cnd G8/G10 F and E . cnd G8/G10 R: E . granulosus s . s . ( G1/G2/G3 ) - EU834886 . 1 , E . equinus ( G4 ) - EU834891 . 1 , E . ortleppi ( G5 ) - FN582298 . 1 , E . canadensis ( G6/G7 ) - EU834893 . 1 , E . canadensis ( G8 ) - EU834894 . 1 , E . canadensis ( G10 ) - EU834896 . 1 . D ) The DNA polymerase delta ( pold ) gene was used to design the E . canadensis ( G6/7 ) specific primers E . cnd G6/G7 pold F and E . cnd G6/G7 pold R: E . granulosus s . s . ( G1 ) - FN568361 . 1 , E . equinus ( G4 ) - FN568362 . 1 , E . ortleppi ( G5 ) - FN568363 . 1 , E . canadensis ( G6 ) - FN568364 . 1 , E . canadensis ( G7 ) - FN568365 . 1 , E . canadensis ( G8 ) - FN568366 . 1 . E ) The calreticulin ( cal ) gene was used to design the E . granulosus s . s . ( G1/G2/G3 ) specific primers E . g ss cal F and E . g ss cal R as well as the E . equinus specific primers E . eq cal F and E . eq cal R: E . granulosus s . s . ( G1 ) - EU834931 . 1 , E . equinus ( G4 ) - EU834936 . 1 , E . canadensis ( G6/G7 ) - EU834937 . 1 , E . canadensis ( G8 ) - EU834939 . 1 , E . canadensis ( G10 ) - EU834940 . 1 . F ) The elongation factor 1 alpha ( ef1a ) gene was used to design the E . granulosus s . s . ( G1/G2/G3 ) specific primers E . g ss Ef1a F and E . g ss Ef1a R: E . granulosus s . s . ( G1 ) - FN568380 . 1 , E . equinus ( G4 ) - FN568381 . 1 , E . ortleppi ( G5 ) - FN568382 . 1 , E . canadensis ( G6 ) - FN568384 . 1 , E . canadensis ( G7 ) - FN568383 . 1 , E . canadensis ( G8 ) - FN568385 . 1 . G ) The cytochrome oxidase subunit I ( cox1 ) gene was used to design the E . equinus ( G4 ) specific primers E . eq cox1 F and E . eq cox1 R: E . granulosus s . s . ( G1/G2/G3 ) - M84661 . 1 , E . equinus ( G4 ) - M84664 . 1 , E . ortleppi ( G5 ) - M84665 . 1 , E . canadensis ( G6 ) - M84666 . 1 , E . canadensis ( G8 ) - DQ144021 . 1 , E . canadensis ( G10 ) - DQ144022 . 1 .
|
The dog tapeworm Echinococcus granulosus ( E . granulosus ) is a cosmopolitan parasite . The adult worms reside in the small intestine of their definitive hosts ( dogs ) . Infective eggs are shed with the feces into the environment and are orally ingested by intermediate hosts where they develop into the metacestode ( larval ) stage , causing cystic echinococcosis ( CE ) in humans and livestock . Ten intraspecific genotypes of E . granulosus ( G1 to G10 ) have been reported from different intermediate host species . Based on the recently established molecular phylogeny , E . granulosus is now considered a complex consisting of four species: E . granulosus sensu stricto ( G1/G2/G3 ) , E . equinus ( G4 ) , E . ortleppi ( G5 ) and E . canadensis ( G6–G10 ) . Simple and highly discriminative molecular epidemiological approaches are needed to explore dynamics , life cycle patterns , and the pathogenicity of the members of this complex . We here introduce a one-step multiplex PCR ( mPCR ) protocol for the genotyping and discrimination of the different members of the E . granulosus complex , allowing three levels of discrimination: ( i ) Echinococcus genus , ( ii ) E . granulosus complex , and ( iii ) genetic variants within the E . granulosus complex . The relatively complicated task of E . granulosus complex speciation and genotyping is clearly simplified by mPCR , and this technique therefore represents a useful tool for routine practice .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"veterinary",
"diseases",
"veterinary",
"parasitology",
"veterinary",
"diagnostics",
"biology",
"microbiology",
"veterinary",
"science",
"parasitology",
"veterinary",
"medicine"
] |
2013
|
A Multiplex PCR for the Simultaneous Detection and Genotyping of the Echinococcus granulosus Complex
|
Chronic atrial fibrillation ( AF ) is a complex disease with underlying changes in electrophysiology , calcium signaling and the structure of atrial myocytes . How these individual remodeling targets and their emergent interactions contribute to cell physiology in chronic AF is not well understood . To approach this problem , we performed in silico experiments in a computational model of the human atrial myocyte . The remodeled function of cellular components was based on a broad literature review of in vitro findings in chronic AF , and these were integrated into the model to define a cohort of virtual cells . Simulation results indicate that while the altered function of calcium and potassium ion channels alone causes a pronounced decrease in action potential duration , remodeling of intracellular calcium handling also has a substantial impact on the chronic AF phenotype . We additionally found that the reduction in amplitude of the calcium transient in chronic AF as compared to normal sinus rhythm is primarily due to the remodeling of calcium channel function , calcium handling and cellular geometry . Finally , we found that decreased electrical resistance of the membrane together with remodeled calcium handling synergistically decreased cellular excitability and the subsequent inducibility of repolarization abnormalities in the human atrial myocyte in chronic AF . We conclude that the presented results highlight the complexity of both intrinsic cellular interactions and emergent properties of human atrial myocytes in chronic AF . Therefore , reversing remodeling for a single remodeled component does little to restore the normal sinus rhythm phenotype . These findings may have important implications for developing novel therapeutic approaches for chronic AF .
Atrial fibrillation ( AF ) , the most common arrhythmia in clinical practice , is a complex disease with multiple etiologies [1] . However , the endpoint can be broadly characterized by two pathophysiological features: a tissue substrate with increased propensity to arrhythmia as well as loss of contractility . These global outcomes are due to adverse remodeling processes , leading to self-perpetuation of the arrhythmia [2] , [3] . Despite its clinical significance the mechanisms of AF-induced contractile dysfunction are still poorly understood , and current drugs for the treatment of chronic AF ( cAF ) increase the risk of life-threatening arrhythmias while featuring only moderate efficacy [4] . In the literature , cAF-related remodeling is typically divided into three categories: ( 1 ) electrical , ( 2 ) contractile and ( 3 ) structural [5] . The first includes decreased conductances of L-type Ca2+ current ( ICaL ) , transient outward K+ current ( Ito ) and ultra rapid delayed rectified K+ current ( IKur ) , and increased conductance of inward rectified K+ current ( IK1 ) , and is considered a typical hallmark of cAF [6] . This electrical remodeling causes , for example , shortening of both the action potential ( AP ) duration and the effective refractory period ( ERP ) . Contractile remodeling , on the other hand , appears to be predominantly a result of impaired intracellular Ca2+ handling , as contractile force can be almost completely restored by increasing the extracellular Ca2+ concentration [7] . Emerging evidence suggests that abnormal Ca2+ handling is a key contributor to atrial remodeling during AF [8] . The third category , structural remodeling , includes changes at both the cellular level ( hypertrophy , glycogen accumulation and modified mitochondrial morphology , among others ) and tissue level ( fibrosis ) [9] . It has been established in both in vitro and in silico experiments that the remodeling of sarcolemmal Ca2+ and K+ channels creates a substrate which supports the maintenance of AF [5] . Recent studies have also demonstrated that remodeled intracellular Ca2+ handling is one of the main causes for the loss of contractility observed in cAF [10] , [11] . Furthermore , cellular hypertrophy has been shown to cause conduction disturbances , even in the absence of increased fibrosis [12] . However , neither how the above mechanisms interact nor how these may contribute as isolated modifications to alter the electrical and contractile function of atrial myocytes in cAF is well understood . To approach this complex problem , we conducted an extensive literature review to form a cohort of virtual cell variants that represent the various cellular components reported as remodeled in cAF . We then analyzed , both in single cells and in tissue , the mechanisms underlying AP shortening , altered intracellular Ca2+ signaling , and changes in excitability in cAF , using a recently developed mathematical model of the human atrial myocyte [13] .
We initially compared the simulation results with in vitro findings for AP and CaT characteristics in cAF vs . normal sinus rhythm ( nSR ) . The model reproduces one of the hallmarks of electrophysiological cAF-remodeling , AP shortening ( Figure 1B ) . Measured as a decrease of APD90 , in simulations AP shortening ( 31 . 9% ) corresponds closely with previous in vitro studies ( Figure 1C ) . In addition , the more negative ( 5 . 5% increase ) resting membrane potential ( RMP ) observed here in cAF as compared to nSR cells is in line with experimental findings ( Figure 1C ) . Cellular cAF-remodeling also causes dramatic changes to Ca2+ dynamics ( Figure 1D ) . In simulations , both reduced diastolic [Ca2+]i ( −29 . 1% for cAF vs . nSR ) and the decreased CaTamp ( −62 . 3% for cAF vs . nSR ) match in vitro findings well ( Figure 1E ) . Additionally , a small reduction in sarcoplasmic reticulum ( SR ) Ca2+ content ( −23% for cAF vs . nSR ) , measured as the integral of INCX during a caffeine pulse ( Figure S2 ) is observed , which corresponds well with the 18% decrease reported in vitro in cAF myocytes [11] . Furthermore , CaT peak is delayed in simulations ( by 49 . 8% for cAF vs . nSR ) , which compares well qualitatively with results obtained from a canine AF model [10] . The CaT decay time constant was also increased by 35 . 6% in cAF vs . nSR , within the reported range for in vitro results ( 28% [14] and 80% [11] ) . The spatiotemporal presentation of CaT in Figure 1 F&G shows there is virtually no rise in [Ca2+]i in the central parts of the cAF-remodeled cell , which also corresponds well to in vitro findings [10] . Simulation results also accurately represent the non-linear nature of cardiac myocyte Ca2+ dynamics: although the maximum conductance of ICaL is decreased by 65% in the cAF model as compared to the nSR model , the total Ca2+ influx is decreased only by 39 . 6% overall , as there is less Ca2+ dependent inactivation of ICaL in cAF . These results are also in line with in vitro findings of 42% and 22% reduction in peak vs . integrated ICaL , respectively [14] . 1D tissue simulations reveal restitution properties that also correspond well to in vivo findings in nSR vs . cAF ( Figure 2 A–D ) . Relative APD90 and conduction velocity ( CV ) changes lie within the measured standard deviation . The model reproduces relative ERP for the nSR case quite well , although this is rather low for the cAF case . The rotor center in a 2D tissue patch for the nSR and the cAF in silico models are depicted in Figure 2 E and F , respectively . The rotor center trajectory of the nSR variant consumes greater area than that of cAF , representing the stabilization of reentrant waves associated with this electrophysiological remodeling . Furthermore , the rotor is meandering comparatively stable during the simulation time of 8 s for the cAF case , whereas the instable rotor center of the nSR case drifts collides with the geometry boundary and vanishes . Movies showing rotor movement for both cases can be found in Supporting Information ( Video S1–S2 ) . Simulated dominant frequencies are 4 . 15 Hz for the physiological case and 11 . 23 Hz for the cAF model , which compares favorably to the range of measured values of 11 . 6±2 . 9 Hz [15] . To evaluate the contribution of individual variables to cAF remodeling , we next simulated changes in each cellular component separately in the model and calculated three resultant biomarkers APD90 , APtri and CaTamp for all cell variants . Each cAF-remodeled component was included one at a time , and the AP and CaT characteristics of all cell variants were compared to the nSR myocyte ( Figure 3 ) . In simulations , reduced ICaL alone decreased APD90 by 17 . 3% , while increased IK1 caused an even greater reduction of APD90 ( by 52 . 7% ) . Increased INCX and cell dilation each had the opposite effect: APD90 increased by 21 . 5% and 7 . 4% , respectively . APtri was substantially decreased in cAF ( −35 . 9% ) vs . nSR , which appeared to be primarily related to increased IK1 and reduced ICaL , as these singular modifications reduced APtri by 58 . 9% and 16 . 6% , respectively . While cell dilation had almost no effect on APtri ( +3 . 1% ) , cAF-remodeled NCX increased APtri quite dramatically ( by 22 . 9% ) . On the other hand , CaTamp was impacted most by reduced ICaL ( −46 . 0% ) , cell dilation ( −20 . 0% ) and increased NCX activity ( −12 . 4% ) . These modifications also hampered the propagation of intracellular Ca2+ waves from sarcolemma to cell center ( right column of Figure 3 ) . The crucial role of increased IK1 and reduced ICaL in cAF remodeling was further demonstrated in tachy pacing ( BCL = 250 ms ) simulations . Results showed that without these two remodeling targets , the virtual cell is unable to recover excitability between stimuli during such a fast pacing regime ( Figure S5 ) . Investigation of restitution properties in 1D tissue simulations ( Figure S7 ) also revealed the dominant effects of changes in ICaL and IK1 on APD , ERP , CV and WL restitution ( as known from measurements and illustrated in Figure 2 A–C ) . Interestingly , the reduction of ICaL alone led to alternans in electrical properties at higher rates . Similar analysis using the same three biomarkers was also carried out for the cAF-remodeled Ito , IKur , SERCA and RyR; the results are shown in Supporting Information ( Figure S3 ) . Surprisingly , the effects of SERCA and RyR remodeling on both CaTamp and AP morphology were very small as compared to the effects of , for example , cAF-remodeled ICaL and NCX . However , it has been shown in animal studies that increased RyR sensitivity has only transient effects on CaTamp , as reduced SR Ca2+ content balances the effect of increased sensitivity [16] . Furthermore , increased PLB and decreased SLN expressions have opposing effects on the Ca2+ affinity of SERCA , so these modifications partially balance one another in the cAF model . To explore putative targets among the remodeled cellular components for reversing the cAF phenotype , we performed simulations in which we excluded each one of these components independently , and then compared AP and CaT characteristics to the full cAF model ( Figure 4 ) . Neglecting the effect of IK1 remodeling caused a substantial increase in APD90 ( +79 . 7% ) , while similarly excluding the effects of ICaL and NCX remodeling , as well as cell dilation had only relatively minor effects ( +2 . 5% , −8 . 7% and +1 . 0% change in APD90 , respectively ) . Interestingly , neglecting the effect of remodeled IK1 renders the virtual cell bistable: depending on initial conditions either a normal or unresponsive/depolarized steady-state is reached via normal pacing ( BCL = 1000 ms; data not shown ) . The second biomarker , APtri , was changed by −12 . 4% , 112 . 7% , −33 . 9% and −17 . 2% in comparison to cAF , when the remodeling of ICaL , IK1 , NCX and cell dilation , respectively , were independently reversed . When compared to nSR , APtri values were not well restored: −43 . 8% , +36 . 4% , −57 . 6% and −46 . 9% for the ICaL , IK1 , NCX and cell dilation , respectively . On the contrary , CaTamp was almost completely restored when the effects of remodeling ( reduced ) ICaL ( +94 . 4% ) were reversed , and enhanced to a smaller extent ( +38 . 2% ) if the virtual cell was not dilated ( Figures 4 B&E , right column ) . The vital role of increased IK1 and reduced ICaL in cAF remodeling was further demonstrated in tachy pacing ( BCL = 250 ms ) simulations; omitting either of these remodeling targets renders the virtual cell unresponsive to pacing at such a rapid rate ( Figure S6 ) . 1D restitution simulations revealed similar results as in single cell simulations ( Figure S8 ) . An increase of APD , ERP , CV and WL was only significant in the cases wherein ICaL or IK1 remodeling were omitted . Interestingly , reduction of IK1 led to alternans at higher rates in this case . Similar analysis was performed for cAF-remodeled Ito , IKur , SERCA and RyR ( results shown in Supporting Information , Figure S4 ) . In a previous study , we showed that SR Ca2+ release is a strong modulator of APD [13] . Here , we used a similar approach to investigate to what extent AP shortening and triangulation in cAF might be reversed if intracellular Ca2+ dynamics were restored to match those in nSR . Figure 5A shows the subsarcolemmal CaT ( CaTss ) clamp used in simulations . Restoring CaTss had substantial effect on AP shape , increasing APD90 by 18 . 9% and APtri by 16 . 1% . Figures 5 C and D illustrate the underlying changes in INCX and ICaL responsible for modifying these late and early stages of repolarization , respectively . As the typical rate of electrical activation of cells in cAF is dramatically faster than in nSR , we next analyzed accumulation of intracellular Na+ and Ca2+ during increasingly fast pacing ( Figure 6 B and C ) . Previous studies have already established that intracellular Na+ accumulation , which is inherently linked to Ca2+ accumulation via NCX , is an important mechanism for AP shortening during fast pacing [11] , [13] . Motivated by the finding that restoring the intracellular CaT appears to impart a beneficial effect on AP shape ( increased APD90; Figure 5 ) , we analyzed the effect of independently reversing remodeling of cellular components affecting CaT properties the most: ICaL reduction , increased NCX , and cell dilation . As results in Figure 6 reveal , reversing remodeling of ICaL affects Na+ and Ca2+ accumulation most dramatically of the three . Interestingly , in addition to increasing the magnitude of ion accumulation , there is also dramatic shift in the ionic dynamics . Specifically , when ICaL is restored to a “healthy level” in a cAF-remodeled virtual cell , the regime of Ca2+ overload is shifted to larger , more physiologically relevant BCLs ( Figure 6A ) . Similar analysis was performed for all the other cell model variants ( results shown in Supporting Information , Figure S10 ) . To show directly that Na+ accumulation is still a mechanism responsible for AP shortening in drastically remodeled cells , we clamped Na+ concentration to its steady-state value when pacing the model at BCL = 1000 ms , while all other variables represent a steady-state at BCL = 167 ms . The late phase of AP repolarization is slowed substantially ( Figure S9E ) during Na+ clamp , as there is less intracellular [Na+] to activate the Na+/K+ ATPase ( NKA ) current ( Figure S9F ) than when Na+ is allowed to accumulate normally . Delayed afterdepolarizations ( DADs ) have been linked to various arrhythmogenic diseases; however their role in cAF has not yet been elucidated [17] . The main mechanism for induction of cellular DADs in human atrial cells has been shown to be NCX [18] . It was thus of special interest to examine , how remodeled Ca2+ handling might affect the inducibility of DADs in these cells . First , we tested whether DADs could be induced via an extra opening of RyRs during diastole ( Figure 7A ) . While it was not possible to induce DADs in the cAF-remodeled virtual cell ( Figure 7B ) , the subsequent activation of NCX ( Figure 7B ) in the nSR model variant did elicit DADs , as the inward current sufficiently depolarized the virtual cell to elicit an AP . A possible explanation for this surprising finding is the reduced SR Ca2+ content in cAF cells . To test this , we employed the same protocol used in cAF versus nSR virtual cells in a cell variant featuring RyR remodeling only ( with all other features identical to nSR model ) . Even maximal opening of the RyR was not enough to activate NCX and induce a DAD in this variant ( cAF: RyR in Figure 7 ) , supporting the hypothesis that it was not possible to induce DADs in cAF cells due to reduced SR Ca2+ content . In this case , however , membrane potential following RyR opening was closer to the AP initiation threshold than in the original cAF cell variant ( Figure 7B ) . An additional mechanism to explain the lack of DADs in the cAF-remodeled cell could be increased IK1 , which might stabilize the membrane potential such that pathological opening of RyRs during diastole would not induce a DAD . To investigate this possibility , we performed simulations in which DADs were induced by current injection during diastole ( Figure 7D ) . A 65% greater current amplitude was needed to induce a DAD in cAF as compared to nSR , suggesting that the cAF model membrane was indeed more stable with respect to depolarization . To further isolate the role of increased IK1 , we ran corresponding simulations in a model variant that included only the remodeling of IK1 ( cAF: IK1 in Figure 7 ) . Compared to nSR , a 69% greater current amplitude was needed to induce a DAD in this model variant ( Figure 7D ) . These results implicate two mechanisms that dramatically reduce the inducibility of DADs in cAF-remodeled virtual cells . First , when SR Ca2+ content is reduced , it is not possible to release a sufficient amount of Ca2+ from the SR to activate NCX to the extent that would elicit a DAD . Second , increased IK1 decreases cell excitability and stabilizes the membrane potential against DADs in cAF , as with a RMP hyperpolarization of −3 . 7 mV , a larger depolarizing current is needed to reach the AP threshold . Thus , reduced SR Ca2+ load together with increased IK1 actually overcompensates for the combined , contradictory effects of increased RyR sensitivity and increased expression of NCX , such that DAD inducibility is reduced rather than enhanced in our in silico model of cAF . Simulation results are summarized in Figure 8 in a heat map-like presentation , where each individual modification is rated as based on its impact in remodeling ( from nSR to cAF ) and in reverse remodeling ( from cAF closer to nSR ) on selected AF biomarkers . It is apparent that individual modifications have differential impact on cell function depending whether each is involved in remodeling or reverse remodeling ( i . e . the isolated modification is aimed at improving function following cAF remodeling ) . For example , while cell dilation affects several functional variables ( CaT , Na+ accumulation , DAD inducibility and cell excitability ) during the remodeling process , when excluded from cAF-remodeled cells , cell dilatation affects only cell excitability . It is also clear that the cAF phenotype is more resistant to modifications than the nSR phenotype ( see larger gray area in Figure 8 ) . In evaluating the impact of isolated features of remodeling , results also show that the increase in IK1 is the most central component in cAF remodeling , since it affects all functional markers .
The importance of abnormal intracellular Ca2+ handling in the pathophysiology of AF is becoming clear [19] , [20] . According to our simulations , Ca2+ signals are severely blunted in cAF ( smaller CaTamp ) , which corresponds well with experimental data [10] , [11] , [14] , [21] . These changes are primarily due to decreased ICaL , and secondarily to the increased activity of NCX . Changes in other remodeling targets involved in Ca2+ handling , such as RyR and SERCA , had only minor impact on Ca2+ signals ( Figures 3 and 4 and Figures S2 and S3 ) . Reduction in ICaL exerts its effect not only by limiting the Ca2+ influx and thus the immediate trigger for Ca2+-induced Ca2+ release during the AP , but also reduces SR Ca2+ content over time , thereby further reducing the strength of SR Ca2+ release . In atrial myocytes , Ca2+-induced Ca2+ release involves two separate phases: the Ca2+ influx first activates RyR clusters in the vicinity of the sarcolemma and Ca2+ release from these junctional release sites triggers a propagating Ca2+ wave activating adjacent RyRs located deeper within the cell [10] , [13] , [22] . However , recent data suggest that larger mammals , including humans , might actually have a more developed network of t-tubules in their atrial cells than previously thought [23] . In fact , t-tubules are present in the ovine atrial myocytes at low density and strongly reduced in AF , leading to gradual loss of synchronization of Ca2+ signals [24] . This feature of cAF-remodeling is a very interesting topic for future study , particularly as increased spatial heterogeneity in Ca2+ diffusion within the cell has been shown to promote the genesis of Ca2+ alternans [25] . Our simulations and in vitro data [10] have shown the vulnerability of the fire-diffuse-fire mechanism to disruption in cAF; suppression of Ca2+ influx during AF remodeling leads to severely compromised transverse propagation of Ca2+ inside the cell . Ca2+ ions are thus only circulated within the volume just beneath the sarcolemma in this case; a defect aggravated by cAF-induced cell dilatation , which increases the volume of the junctional space and further dilutes [Ca2+] in the subsarcolemmal space . This redistribution of Ca2+ is likely to contribute to the suppression of contraction during AF , as with impaired propagation , Ca2+ signals do not reach the contractile elements located centrally within the cell . Furthermore , this alteration likely has a profound indirect effect on energy expenditure of the cAF-remodeled cells , since contractile elements are not activated to same extent and thus consume less energy than in healthy cardiomyocytes [26] . According to experimental data from patients suffering from chronic AF , contractile force of atrial tissue can be restored with increased extracellular [Ca2+] [7] . This implies that remodeling of the cellular contractile elements involved in cAF has a lesser role in depressing contraction in cAF as compared to the impact of altered Ca2+ signaling in the disease . Furthermore , the data also suggest that Ca2+ influx is the single most influential variable when considering cAF-induced contractile dysfunction in the light of electrical remodeling . This view is supported by our simulations , wherein cAF-induced ICaL downregulation alone reduces CaTamp by 46% and induces defects in Ca2+ signal propagation ( Figure 3 ) . A number of remodeling targets in AF have been proposed to contribute to changes in the AP waveform [5] , [17] . Our simulations support the conclusion that the two most important elements leading to AP shortening in AF are increased IK1 and decreased ICaL . Increased IK1 alone reduces APD by 52 . 7% and is also the single most influential factor contributing to AP triangulation ( Figure 3 ) . Furthermore , the membrane potential is hyperpolarized in diastole by the cAF-related remodeling of IK1 ( which is a repolarizing current and one of the main contributors to the maintenance of the RMP in cardiac myocytes ) . Interestingly , some of the remodeling modifications also act to lengthen the AP , e . g . increases in INCX and cell volume . Increase in the NCX current promotes augmented inward current during AP repolarization when it exchanges cytosolic Ca2+ ions for extracellular Na+ ions at a ratio of 1∶3 . The effect of the cell volume on AP is a bit more indirect; the volume increase appears to delay the Ca2+ removal from the cytosol , which in turn increases INCX during late repolarization ( Figure 3 ) . Intracellular Ca2+ signals and the AP are tightly coupled in human atrial myocytes inherently , and this coupling seems to be an essential part of AF remodeling . Hence , as compared to changes in e . g . K+ current densities which have more straightforward effects on AP , changes in variables involved in Ca2+ signaling , like ICaL and INCX , modulate not only the AP directly , but have more adverse consequences through their effects on intracellular Ca2+ signals . In our simulations , remodeling of ICaL alone reduced APD90 by 17 . 3% ( compared to AP shortening by 31 . 9% in cAF , Figure 3 ) , while normalization of CaT in cAF cells lengthened the AP by 18 . 7% ( Figure 5 ) . This suggests that effects of ICaL on AP are mediated only partly by direct impact of the current on Vm and that major effects come via the secondary suppression of CaT . Changes in AP morphology also impact tissue electrophysiology ( see Figure 2 ) . The simulated tissue APD90 is reduced by around 30% , which is in agreement with the available in vivo data [27] . Similarly , the simulated ERP is reduced by about 20% . In this case , the measured data from Yu et al . [28] revealed reduction by a lesser extent ( around 10% ) as compared to simulation data . These differences might be due to different stages of remodeling . CV is not influenced to a great degree in our simulations , as we did not include gap junction remodeling . Feld et al . [29] measured a reduced CV in cAF , suggesting that there might be changes in conductive tissue properties during cAF . The reduced wavelength ( the product of ERP and CV ) in the simulated cAF case suggests the higher chance of the maintenance of AF following rotor initiation . The simulated rotor center trajectories ( Figure 2 ) show that these anchor more easily in cAF , evincing greater stability , whereas in nSR , the rotor core tends to meander and subsequently might be eliminated at a boundary or an anatomical obstacle . The simulated dominant frequencies also demonstrate the higher chance of a permanent fibrillation in the cAF case . Intracellular Na+ accumulation has been established as an important mechanism for AP shortening during fast pacing in previous studies [11] , [13] . In cardiac myocytes , [Na+]i is mainly dictated by the balance between Na+ influx during an AP ( upon activation of INa and INCX ) and Na+ efflux ( through NKA and NCX ) . Therefore , Na+ fluxes are tightly coupled with both [Ca2+]i and activation frequency , both of which are drastically altered in AF . According to our simulations , high frequency activity induces substantial Na+ accumulation in cAF cells , and this accumulation acts to shorten the AP upon activation of NKA , although this mechanism is less prominent than in nSR cells . In all types of cardiac myocytes , Na+ accumulation can result indirectly via Ca2+ overload which itself automatically results from high frequency pacing . In cAF cells , Ca2+ overload is limited by remodeling ( reduced ICaL ) , which drastically suppresses the AP-evoked CaT . Thus , there is less Ca2+ to be extruded by NCX and consequently a lesser degree of Na+ accumulation . To demonstrate the link between [Ca2+]i and [Na+]i , we normalized the ICaL in cAF model and noticed that pacing-induced Ca2+ and Na+ accumulation were both augmented ( Figure 6 ) . It could be hypothesized that altered Ca2+ and Na+ balances are actually among the features of cAF cells that enable sustained high frequency activity . When cardiomyocytes act to restore normal levels of [Ca2+]i and [Na+]i , vast amounts of ATP are consumed by SERCA ( to pump Ca2+ to the SR ) and by NKA ( to pump Na+ to the extracellular space ) . Thus , when ion gradients are smaller , cAF cells can maintain high frequency of activation at lower energy costs [26] . The main mechanism for induction of cellular DADs in human atrial cells has been shown to be activation of NCX [18] . As NCX is overexpressed in cAF , we expected to see a lower threshold for DADs in simulations; however , results revealed the opposite finding . In fact , an extra Ca2+ release from the SR was not enough to trigger DADs in the cAF-remodeled virtual cell . This result appears to contradict the recent in vitro findings that showed increased spontaneous Ca2+ waves in cells of AF patients [30] , [31] . Possible explanations for this discrepancy include the measurement conditions ( experiments carried out at room temperature ) and pooled patient population ( no separation for paroxysmal , persistent and chronic AF ) . Indeed , further analysis of our simulation results showed that the mechanism explaining this surprising finding was the reduced cellular excitability due to increase of inward rectifying K+ currents . Our results suggesting reduced DAD inducibility in cAF contradict the recent finding that enhanced SR Ca2+ leak and NCX function underlie DADs in patients with cAF [14] . In another study , however , DADs were not observed in either nSR nor cAF patient tissue despite the fact that the measurement conditions were in favor of such events , as IKur was blocked with AVE0118 compound [32] . These controversial results suggest that increased propensity for DADs in cAF might depend , for example , on underlying etiologies in the patient population . To summarize , DAD inducibility depends on four factors mechanistically: 1 ) the strength of the input ( SR Ca2+ load ) , 2 ) how this input is transformed into a trigger ( sensitivity of RyR ) , 3 ) how much depolarizing current this trigger induces ( NCX vs . SERCA balance in Ca2+ removal from the cytosol ) , and , ultimately , 4 ) if the depolarizing current is large enough to depolarize the membrane voltage above the threshold for INa activation ( which depends on the dynamic balance of depolarizing and repolarizing membrane currents ) . All four factors are altered in the context of cAF . Because factors #2 and #3 are greater ( increase RyR sensitivity and greater net depolarizing current , respectively ) one might intuitively infer that DAD inducibility would be increased in cAF-remodeled cells . In our in silico cAF model , however , the reduced SR Ca2+ load together with increased IK1 , which reduce the trigger and stabilize the resting membrane potential , respectively , overcompensate for the combined depolarizing effect of increased RyR sensitivity and increased expression of NCX , such that DAD inducibility is actually reduced . Future computational studies , possibly employing stochastic methods and finer spatial resolution , should address factors #2 and #3 in more detail , when in vitro data on the co-localization of RyR and NCX in cAF vs . nSR human atrial cells becomes available . Both AP shortening and loss of contractility are hallmarks of cAF . Our analysis indicates that , at the cellular level , these changes are strongly coupled to the increased IK1 and decreased ICaL conductances , respectively . In fact , without the IK1 modification , the cAF-remodeled cell becomes unresponsive during more rapid pacing due to sustained depolarization of the membrane voltage ( which inactivates fast sodium channels ) . Decreased ICaL conductance , on the other hand , has a more diverse effect . While also contributing to AP shortening , reduced ICaL is the main mechanism for the diminished intracellular CaTamp in cAF . The large impact of the remodeling of ICaL is related to its dual role , since it not only acts as a trigger for Ca2+ release from the SR , but also affects Ca2+ loading of the SR . Remodeled ICaL and INCX work in synergy to adapt the cell to abnormally fast reoccurring activation in cAF . While reduced ICaL and increased INCX both reduce Ca2+ overload during fast pacing , they also shift Ca2+ dynamics from the normal “whole-cell state” to a “subsarcolemmal state” , where Ca2+ cycling is limited primarily to the vicinity of the cell membrane . Myocyte hypertrophy exacerbates the effect of remodeled Ca2+ handling , in that it further reduces CaTamp in cAF in addition to the effects of ICaL and NCX remodeling . The dilation of the cell also increases the delay between the peaks of the AP and the CaT , which may have arrhythmogenic effects in tissue . In fact , Schotten et al . [12] found that myocyte hypertrophy can cause conduction disturbances in the absence of increased fibrosis in a goat model of chronic atrio-ventricular block . As changes in intracellular Ca2+ signaling are centrally involved in normal and pathological regulation of myocyte growth , apoptosis and necrosis [33] , cell dilation warrants further research to elucidate its role in cAF . Anti-arrhythmic drug therapy to counter AF has long concentrated on agents that may delay atrial repolarization . Drug targets have included , for example , IKr and IKs; however , more recently agents blocking IKur have been studied extensively , because of the current's atria-specificity in human myocardium . More recently , intracellular Ca2+ handling has been established as a potential drug target in cAF [19] . As our results showed ( Figure 5 ) , restoration of intracellular CaT could , hypothetically , be used to improve AP shape ( increase APD90 ) in cAF to , for example , lengthen the effective refractory period . The most effective targets for restoring healthy cell properties following cAF-induced electrical remodeling are likely to be those that most impact the cAF phenotype . Our simulations suggest that changes in IK1 and ICaL in isolation induce most of the characteristic features of cAF ( Figure 8 ) . Therefore , restoring either the K+ or Ca2+ conductance could , in theory , be effective in limiting the effects of electrical remodeling in the cAF substrate . However , complete reversal of any single cAF-induced change via pharmacological means is not likely to be feasible . Instead , it might be useful to consider therapies that aim at partial restoration of combinations of targets . In such efforts , however , understanding the full implications of altered cellular electrophysiology on tissue and organ dynamics is absolutely essential . To illustrate , consider the partial inhibition of K+ currents ( for increasing APD90 and thus ERP ) , in combination with drugs aimed at increasing CaTamp ( for restoring contractility ) . Partial block of NKA with digoxin , combined with reduced RyR Ca2+ leak using a calmodulin kinase II inhibitor , appeared to be beneficial in single cell simulations ( Figure S11 ) and may actually become feasible in the near future , as novel specific blockers of IK1 are being developed [34] . The 1D restitution results ( Figure S12 ) also illustrated increases in APD , ERP , CV and WL , which may be desirable in terms of protecting against arrhythmia . However , this model variant developed alternans at higher pacing rates . In 2D simulations , these alternans also led to a break-up of a single rotor into two rotors ( not captured within the geometry , so excitation vanishes; Figure S12E and Video S3 ) . In a realistic geometry , such wavebreak could lead to stable fibrillatory activity . This finding highlights the need to carry out in silico analysis of potential drug targets at different scales ( cell , tissue , organ ) to achieve a more realistic understanding of pharmacological effects . Although the human atrial myocyte model employed here has been shown to be the most internally consistent and physiologically accurate to date , particularly regarding intracellular Ca2+ handling , in a recent comparison , the model has its limitations [35] . Furthermore , a holistic analysis of cAF as effected in this study is inevitably biased to some extent by the fact that the pathophysiology clinically involves multiple etiologies . Some studies group available data based on , for example , whether patients have a valvular disease or not , while other studies pool the data among AF subtypes and etiologies . Finally , our model of the cAF-remodeled cell is by no means exhaustive , as novel mechanisms of electrical remodeling are reported continuously . Instead , we have included those remodeling targets that have been established in more than one study of human atrial electrophysiology . When novel experimental data on these disease mechanisms accumulate , the model should be updated accordingly . The results indicate that , at the cellular level , reduced ICaL and increased INCX contribute synergistically to adapt the cell to fast activation rates of cAF by reducing Ca2+ overload , which additionally causes a drastic decrease in CaTamp at normal heart rates . Furthermore , our findings suggest that an increase of IK1 in cAF is the dominant mechanism responsible for AP shortening in cAF , while the effect of reduced ICaL is less prominent and the role of remodeled Ito and IKur are rather insignificant . Increased IK1 , in synergy with reduced intracellular Ca2+ stores , also stabilizing the cAF-remodeled cell against DADs . The results also show that , in addition to remodeling of ion currents and Ca2+ handling , cellular hypertrophy is an important mechanism contributing to changes in atrial refractoriness , contractility and arrhythmogenicity . Finally , the intrinsic complexity and interdependency of electrophysiological mechanisms are highlighted by our analysis . The presented results thus suggest that instead of targeting a single cellular component a more holistic approach is worth considering when looking for novel therapeutic approaches for chronic AF .
The modeling platform of this study is our recently developed human atrial myocyte model that enables the simulation of emergent spatiotemporal characteristics of intracellular Ca2+ dynamics [13] . Methods for simulation of tissue-level electrophysiology and its analysis are presented in the Supporting Information and are detailed in [35] . Contrary to most previous in silico studies of cAF , we performed a broad literature search on cellular remodeling to define the average remodeled parameter values ( Figure 1A ) instead of using a single in vitro data set or small subset . We have included those remodeling targets that have been established in more than one study . Full sets of referenced human data are shown in Supporting Information ( Tables S2–S4 ) . The modifications of existing model components , as well as the simulation protocols are described in detail in the Supporting Information . Briefly , we reformulated the ICaL to increase the contribution Ca2+-dependent vs . voltage-dependent inactivation of the current , and decreased the time constants based on recent in vitro data [36] , Supporting Information Figure S1 . Parameters of the SERCA pump have been modified according to a previously developed scheme [37] , [38] to enable the representation of changed expression of phospholamban ( PLB ) and sarcolipin ( SLN ) in cAF . In our analysis of cAF-related cellular remodeling , we use the following three biomarkers:
|
Atrial fibrillation is a complex disease which , at the level of individual atrial muscle cells , is a result of changes in a number of ion channels and transporters , as well as in cellular structure . How these alterations , together and separately , affect electrical and contractile function of the atrial cells is not well understood . In this study , we evaluated the effect of these changes using a computational approach . Our results show that abnormal function of both calcium and potassium ion channels at the sarcolemma has the largest impact on the electrical properties of the human atrial myocyte . Changes in intracellular calcium handling and cellular geometry are also significant for cellular function . Finally , our results highlight the interactions and additive effect of these abnormalities , in that a hypothetical restoration of any single modification does not result in recovery of function to a healthy phenotype . These findings have potentially important implications for developing novel treatment options for atrial fibrillation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"physics",
"computer",
"and",
"information",
"sciences",
"medicine",
"and",
"health",
"sciences",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences"
] |
2014
|
In Silico Screening of the Key Cellular Remodeling Targets in Chronic Atrial Fibrillation
|
The protozoan parasite Entamoeba histolytica causes a wide spectrum of intestinal infections . In severe cases , the trophozoites can breach the mucosal barrier , invade the intestinal epithelium and travel via the portal circulation to the liver , where they cause hepatic abscesses , which can prove fatal if left untreated . The host Extra Cellular Matrix ( ECM ) plays a crucial role in amoebic invasion by triggering an array of cellular responses in the parasite , including induction of actin rich adhesion structures . Similar actin rich protrusive structures , known as ‘invadosomes’ , promote chemotactic migration of the metastatic cancer cells and non-transformed cells by remodeling the ECM . Recent studies showed a central role for Rab GTPases , the master regulators of vesicular trafficking , in biogenesis of invadosomes . Here , we showed that fibronectin , a major host ECM component induced actin remodeling in the parasite in a Rab21 dependent manner . The focalized actin structures formed were reminiscent of the mammalian invadosomes . By using various approaches , such as immunofluorescence confocal microscopy and scanning electron microscopy , along with in vitro invasion assay and matrix degradation assay , we show that the fibronectin induced formation of amoebic actin dots depend on the nucleotide status of the GTPase . The ECM components , fibronectin and collagen type I , displayed differential control over the formation of actin dots , with fibronectin positively and collagen type I negatively modulating it . The cell surface adhesion molecule Gal/GalNAc complex was also found to impose additional regulation on this process , which might have implication in collagen type I mediated suppression of actin dots .
The protozoan parasite , Entamoeba histolytica , is the causative agent of amoebic dysentery . Approximately 35–50 million cases of clinical amoebiasis are reported worldwide with 100 , 000 deaths per year . Despite substantial improvement in the global sanitary facilities ( WHO Water Sanitation and Hygiene Data ) , amoebiasis still poses as a burden on the health care system of the developing economies . The infection can soar up with variable outcomes manifesting in diarrhea , invasive colitis or metastatic infection . The invasive disease pathologies are usually associated with massive destruction of the host tissue . This is caused partly due to invasion of the parasite through the intestinal epithelium and it’s migration to the extra intestinal sites and pro inflammatory responses of the host . Both these factors contribute to amoebic hepatic abscess [1] . Although , amoebiasis poses as a major health risk to the developing countries , the studies reporting molecular basis of tissue invasion are very limited [1] . The actin cytoskeleton organization and dynamics plays an important role in motility related functions of the parasite , many of which are relevant for the observed invasive pathologies . During colonization and infection , the parasite comes in the contact with the host extracellular matrix . The host milieu presents the pathogen with overwhelming stimuli in the form of various extracellular matrix ( ECM ) components . The ECM cues in turn can activate an array of signaling pathways which lead to remodeling of actin cytoskeleton in the parasite [2] . Furthermore , adhesion to fibronectin ( FN ) , a ubiquitous ECM component , is known to induce secretion of amoebic proteases which locally degrade the ECM and promote random and directed motility of the parasite[3 , 4] . The fibronectin receptor ( β1EhFNR ) has been identified and characterized in E . histolytica and is antigenically similar to the human β1 integrin [5] . FN is also shown to induce actin rich dots in E . histolytica , structures that visually resemble the mammalian invadosomes [6] . Thus , fibronectin act as a potent signal for the parasite , inducing actin remodeling in various ways . Though the involvement of FN in amoebic cytoskeletal remodeling is well presented in the literature , identity of the molecular players acting downstream of fibronectin remains mostly unknown . Eukaryotic cells constantly converse with their surrounding by forming various actin rich structures . These , prominently include focal adhesion , lamelliopodia , filopodia and invadosomes [7 , 8] . The invadosomes found exclusively in transformed cells and cells of monocytic origin are formed as a result of an interplay between the structural proteins viz vinculin , paxillin and actin regulatory and nucleating factors like Arp2/3 ( actin related protein 2/3 ) , N-WASP ( Neuronal Wiskott Aldrich syndrome protein ) , Rho GTPases[9] . These are dynamic structures of rapidly polymerizing actin which are jointly regulated by cues from the ECM and growth factors . Invadosomes allow a cell to integrate remodeling and degradation of the ECM with cell motility [10] . They act as hubs for the activity of the integrins [11 , 12] and are the major sites for ECM degradation due to focalized secretion of matrix metalloproteinases ( MMPs ) [13–15] . The importance of Rab GTPases , a class of small G proteins , which principally coordinate vesicular trafficking in eukaryotic cells , has been recently realized in formation and maturation of invadosomes [16 , 17] . Rabs act as molecular timers for host of cellular processes in a nucleotide dependent manner [18 , 19] . Dysregulated expression of Rab GTPases has been associated with various cancers [20 , 21] . Change in expression of Rabs could lead to abnormal trafficking of the growth factor receptors , cell surface integrins and matrix metalloproteinases which could further cause increased cell proliferation and migration . Active invadosomes are marked by the presence of MT1MMP ( membrane type1 matrix metalloproteinase ) , a principal membrane anchored collagenolytic protease . Recently Rab5a , Rab8a and Rab14 have been identified as regulators of vesicular trafficking of MT1MMPin mammalian cells [22 , 23] . While the above subset of Rabs regulate matrix degradation , Rab21 and Rab25 govern cell adhesion and migration by interacting with integrin heterodimers [24 , 25] . Here , we showed that upon FN stimulation , Entamoeba histolytica forms unique actin richdots on its ventral surface in a Rab21 dependent manner . Rab21 could induce de novo amoebic actin dots in a nucleotide ( GDP/GTP ) dependent manner . Interestingly , additional regulation was observed over the process by cell surface adhesion molecule , the Gal/GalNAc ( galactose/N-acetyl galactosamine ) lectin complex which may have implication in collagen type I mediated suppression of the actin dots . Over expression of Rab21CA ( constitutive active or GTP hydrolysis deficient mutant ) increased the invasive index of the trophozoites as was evident from the transwell matrigel invasion and fluorescent matrix degradation assays . Taken together , our results indicate that Rab21 along with the host ECM components plays a potential role in tissue invasion and thereby may contribute to the severity of amoebic infection .
E . histolytica has a dynamic cytoskeleton , allowing it to respond to the signaling inputs from the environment . Host ECM forms an intricate part of the parasite surroundings . Therefore we wanted to study the effect of ECM components on the parasite’s cytoskeleton . Matrigel is a reconstituted basement membrane preparation of heterogeneous composition , an ideal mimic for the host ECM [26] . On the other hand , fibronectin is one of the most ubiquitous and abundant ECM proteins . We used matrigel and fibronectin to seek for changes in the actin cytoskeleton of the parasite . When stimulated with matrigel or fibronectin , more than 60% of the total trophozoites formed focalized actin rich dots on the substratum contacting side ( Fig . 1A ) , referred now onwards as amoebic actin dots . These dots had a mean area of 2 . 5±0 . 21μm2 and extended over 2–3μm in depth ( Fig . 1B , orthogonal view of actin dots ) . We observed a minimum of 2 actin dots per cell to maximum of 42 actin dots per cell with a modal value ( Nmode ) of 10 for fibronectin coated surface . In case of matrigel coated surface , we observed a minimum of 3 actin dots per cell to maximum of 40 actin dots per cell with a modal value ( Nmode ) of 6 . The actin dots visually resembled ECM degradative structures , the invadosomes [27] . This intrigued us to carry out immunostaining for an invadosomal marker , vinculin . Vinculin , an actin binding protein has been previously reported to be a part of the signaling complex assembled in response to the FN exposure in the pathogen [28] . The co-localization of vinculin and actin ( S1 Fig . ) , suggest that the amoebic actin dots also share architectural molecules with mammalian invadosomes . Since collagen type I , the other major ECM component is known to induce a novel class of linear invadosomes in mammalian cells [29] , we further investigated its effect on the actin cytoskeleton of the parasite . Interestingly , unlike fibronectin and matrigel , collagen type I did not promote the formation of actin dots . The actin cystoskeleton remained largely unperturbed in the presence of collagen type I ( Fig . 1A ) . Invadosomes have a dynamic F- actin core with a rapid turnover [9] . Cytochalasin D is a drug widely used against actin polymerization which inhibits the process by binding to the barbed ends of a growing actin polymer . Interestingly , it has also been shown to promote invadopodia formation through activation of Tks5 ( tyrosine kinase substrate with five SH3 domains ) , a Src substrate with a scaffolding role in the assembly and organization of invadopodia [9 , 30] . Thus , the action of cytochalasin D on actin polymerization , especially on invadopodia is counter intuitive . To test how the drug modulates the structure of the amoebic actin dots , we treated the FN stimulated trophozoites with increasing concentrations of cytochalasin D and stained for actin . The drug , at increasing concentration clearly abrogated the morphology of actin rich structures . The most dramatic effect was observed at 10μM of cytochalasin D , where 90–95% of cells were marked by the presence of a long actin thread with bulbous projections at regular intervals ( S2 Fig . ) , suggestive of either a hyper polymerization or lack of depolymerization of actin filaments . Rab GTPases , member of Ras superfamily , are master regulators of vesicular trafficking . They control the route traversed by the cargo inside the eukaryotic cells . They also regulate the function of invadosomes by controlling the trafficking of matrix metalloproteinases [22 , 23 , 31] . Here , we tested a subset of Rab GTPases vizRab5 , Rab21 , Rab7A and Rab7B , some of the amoebic homologues of mammalian endocytic Rabs , to decipher their role in amoebic actin dot formation . We also included amoebic homologue of the recycling Rab i . e . Rab11B . Stable amoebic transformants of the candidate Rab GTPases and vector control ( transfected with pEhExHA/pEhExMyc ) were generated . The amoebic cysteine synthase promoter was used for expressing the genes of interest . The stable transfectants were selected and maintained at 20μg/ml G418 . The different Rab expressing cell lines were plated on glass coverslips , incubated at 37°C for an hour to attach and stained with Alexa 568Phalloidin for visualizing actin . As shown in Fig . 2A , over-expression of Rab5 , Rab7A and Rab21 induced de novo formation of actin dots on glass surface . Interestingly , the actin dots were observed in the moderately expressing trophozoites which accounted for 25–30% of the total trophozoites . For calculating the number of actin dots formed in a given cell line , we analyzed approximately 300 trophozoites over 15 images with 20–22 trophozites per image . The quantification for the same has been depicted in Fig . 2B . Interestingly , we observed actin rich adhesion plates for vector control ( pEhExHA ) , Rab7B and Rab11B expressing trophozoites . Although , Rab5 , Rab7A and Rab21 induced actin formation in the parasite ( Fig . 2A ) , the primary knowledge on the role of human Rab21 ( hRab21 ) being involved in integrin dependent cell migration and adhesion [24] and complete lack of information of the cellular functions of amoebicRab21 prompted us to continue further studies with Rab21 . Receptor-mediated endocytosis and macropinocytosis are essential cellular functions required for nutrient acquisition . hRab21 regulates both these processes [32] . Therefore , to determine whether amoebicRab21 is involved in endocytosis , we carried out classical cargo uptake experiments ( manuscript under review , PNTD-D-14-01300 ) . The endocytosis of holo-transferrin and dextran were assayed in trophozoites expressing wild type Rab21 ( Rab21WT ) , constitutive active or GTP hydrolysis deficient mutant ( Rab21CA ) and dominant negative or GTP binding deficient mutant ( Rab21DN ) using flow cytometery and confocal microscopy . We did not observe any co-localization betweenRab21 and transferrin or dextran positive compartments . Moreover , over-expression of wild type or mutant Rab21 did not show any detectable effect on the uptake of these cargos . Phagocytosis plays an important role in nutrient uptake by E . histolytica and hence is important for its survival and virulence . Rab5 is an important regulator of erythrophagocytosis in amoeba . Since , Rab21 is a member of the Rab5 subfamily , we decided to study whether it is also involved in erythrophagocytosis [33] . Cell tracker labeled RBCs were used to measure the phagocytic efficiency of the trophozoites expressing Rab21WT and mutants ( S2 Method ) . As shown in S3 Fig . , we did not observe any effect of over-expression of Rab21 on RBC phagocytosis . In the current study , we observed that over-expression of Rab21 lead to amoebic actin dot formation . Moreover , these actin rich structures were also formed in the presence of ECM , specifically fibronectin . Therefore to further understand whether Rab21 is involved in this ECM mediated process , we made use of the CA and DN mutants of Rab21 . These mutants are widely accepted tools to understand whether a cellular process is driven by the nucleotide state of the GTPase . As shown in Fig . 3A , the wild type and CA mutant of Rab21 formed actin dots whereas Rab21DN did not . Further we checked for the effect of fibronectin stimulation on the Rab21WT and mutant expressing trophozoites . As shown in Fig . 3B , the effect of Rab21WT and CA on the actin cystoskeleton was enhanced in presence of the stimuli as compared to unstimulated trophozites . Therefore , based on our observations we suggest that the fibronectin mediated actin rearrangement involves Rab21 and it further depends on the nucleotide bound state ( GDP/GTP ) of the GTPase . We further extended our study to investigate the effect of collagen type I , the other major ECM component , on the actin cytoskeleton of the Rab21CA transformants . In contrast to fibronectin , collagen type I suppressed the actin dot formation in Rab21CA transformants ( Fig . 3C ) . Interestingly , smaller actin dots were observed ( mean area = 0 . 35±0 . 024μm2 , n = 180 actin dots , 70 cells , Fig . 3D ) in comparison to the glass surface ( mean area = 2 . 47±0 . 2μm2 , n = 142 actin dots , 50 cells , Fig . 3D ) . Based on the opposing effect observed for the ECM components , we propose the possibility of differential regulation of the actin cytoskeleton . To further characterize the structures and to study the surface topology of the trophozoites at high resolution , we did scanning electron microscopy on the Rab21CA expressing trophozoites sandwiched between thin layers of matrigel . The rational for the sandwich culture was to mimic the 3D environment present inside the host upon infection and to induce the protrusive actin rich structures on the dorsal surface of the trophozoites . The trophozoites were incubated for 6–8hrs in the matrigel sandwich and subsequently processed for acquiring images . As shown in the Fig . 3E , Rab21CA upon incubation in the matrigel sandwich formed numerous surface projections . Similar surface morphology was also observed for the vector control ( pEhExHA ) trophozoites ( Fig . 3E , upper panel ) . In contrast , Rab21CA trophozoites when sandwiched between layers of collagen type I did not form any such protrusive structures but rather had a rough surface devoid of any membrane extensions ( Fig . 3E , lower panel ) . We also carried out an ultrastructural analysis of the trophozoites expressing Rab21CA and DN mutants along with vector control trophozoites plated on glass . Expression of both the mutants resulted in an overall smooth surface . Interestingly , Rab21CA was well spread out and attached to the glass surface ( S4 Fig . , middle panel ) , whereas Rab21DN had a rounded morphology and failed to attach ( S4 Fig . , lower panel ) . Metastatic tumor cells can invade and migrate through the matrigel barrier by secreting matrix metalloproteinases [26 , 34] . The widely used Transwell Matrigel Invasion Assay utilizes the above property to quantify the invasive capacity of the metastatic cells . To understand the functional significance of Rab21 induced actin dots in Entamoeba histolytica we carried out the invasion assay for Rab21WT , Rab21CA , Rab21DN as well as the knock down strain of Rab21 ( Rab21KD ) . Rab21KD was generated using G3 trophozoites transfected with silencing plasmid psAP2Gunma Rab21 ( S1 Method ) [35] . We observed86 . 6% reduction in the expression level of Rab21 in the Rab21KD strain as compared to the vector control ( pSAP2Gunma ) transfected G3 trophozoites ( S5 Fig . ) . Cells were plated in the upper well in serum free medium with the lower well containing adult bovine serum as a chemoattractant . Forty eight hours post-incubation , cells that had invaded were detached from the lower well , harvested and counted . The Rab21CA expressing cells showed increased migration with 47 . 2±8 . 6% of the total cells plated in the upper chamber invading through the matrigel barrier outperforming the Rab21DN which showed 17 . 6±3 . 7% invasion similar to vector control ( 21 . 6±2 . 6% ) whereas the Rab21WT showed 52 . 7±11 . 8% invasion in similar setup ( Fig . 4A ) . The Rab21KD showed least invasion with only 9 . 4±3 . 4% of total cells invading through the matrigel . The vector control for the Rab21KD strain ( psAP2Gunma transfected trophozoites ) showed 14 . 0±3 . 8% invasion . The above results suggest that Rab21modulates the process of migration and invasion . Though matrigel invasion proves to be a useful assay for measuring invasive capacity of the cells , the possibility of amoeboid movement through the barrier cannot be ruled out . Therefore , to confirm that the observed matrix invasion was due to active protease secretion , we carried out fluorescent matrix degradation assay . The surface degradation of Hilyte488 Fibronectin was analyzed with time . Cells expressing Rab21CA and DN were plated on Hilyte488 Fibronectin coated coverslips and incubated for 48hrs under dark , after which they were fixed and stained for actin . The Rab21CA expressing trophozoites showed heavy degradation of the fluorescent matrix , with cells showing a halo ( i . e . loss of fluorescence ) around them ( Fig . 4B ) . Whereas , no matrix degradation was observed for the Rab21DN mutant indicating that actin dots formed as a consequence of the Rab21CA over-expression conferred the cells with increased degradative capacity . Invasion through the matrix barrier relies on the successful attachment of the parasite to the host ECM components . The binding to the host glycoproteins is mediated by the Gal/GalNAc lectin complex , a major amoebic surface receptor [36] . Interestingly , the complex also mediates binding of the parasite to collagen type I [37 , 38] . Therefore , we decided to study the role of Hgl on biogenesis of the actin dots . The essential role of Carbohydrate Recognition Domain ( CRD ) of heavy subunit ( hgl ) of the lectin complex in adherence and cytolysis of target cells has been demonstrated by carbohydrate and monoclonal antibody mediated inhibition assays [36 , 39] . Therefore to study the role of Hgl in biogenesis of actin dots , we blocked the ligand binding site with increasing concentration of GalNAc or monoclonal antibody , αHgl 3F4 , raised against the recombinant CRD [39] . The monosaccharide , GalNAc showed an inhibitory effect . As shown in Fig . 5A and 5B , cells expressing Rab21CA mutant when plated on glass showed a dramatic decrease in the number and the size of the actin dots in the presence of GalNAc ( mean area on glass = 2 . 41±0 . 36μm2 , n = 95 actin dots , 70 cells; mean area in presence of 100mM GalNAc = 0 . 36±0 . 183μm2 , n = 130 actin dots , 94 cells ) . In agreement with the GalNAc inhibition assay performed using sugar , the αHgl3F4 antibody also suppressed the formation of the dots on the glass surface {mean area = 2 . 05±0 . 15μm2 ( untreated ) , n = 283 actin dots , 152 cells; mean area = 0 . 28±0 . 05μm2 ( treated with αHgl 3F4 ) , n = 155 actin dots , 91 cells} ( Fig . 5E and 5F ) . We further extended our assay to the fibronectin coated surface using both , the sugar and the αHgl 3F4 . Interestingly , the inhibitory effect of GalNAc and αHgl 3F4 on actin dots was not observed on fibronectin coated surface ( Fig . 5C and 5G ) . In presence of GalNAc , Rab21CA expressing trophozoites plated on FN coated surface formed actin dots {mean area on FN coated surface = 1 . 38±0 . 15 μm2 , n = 130 actin dots , 60 cells; mean area in presence of 100mM GalNAc = 1 . 59±0 . 20 μm2 , n = 138 actin dots , 65 cells; Fig . 5D} . Similarly , we observed no suppression of actin dots on FN coated surface in presence of αHgl 3F4 {mean area = 1 . 20±0 . 30 μm2 ( untreated ) , n = 184 actin dots , 134 cells; mean area = 1 . 48±0 . 36 μm2 ( treated with αHgl ) , n = 130 actin dots , 90 cells; Fig . 5H} . Based on our above observations , we speculate that the lectin complex via its CRD region regulates the biogenesis of actin dots in the parasite .
Entamoeba histolytica , a professional phagocyte displays varying responses to the extra cellular matrix ( ECM ) . Interaction with the ECM is important for amoebic invasion of the host tissue and therefore is a key step in the pathogenesis . Both , host and pathogen factors contribute to the invasion process [40 , 41] . The ECM is a highly diverse and dynamic protein network , providing structural support and signaling cues which regulate cell behavior [42 , 43] . The cell responds to these signaling inputs by employing various actin rich structures [7] . Invadosomes are a class of actin rich protrusive structures which degrade the underlying matrix by secreting a battery of zinc regulated matrix metalloproteinases ( MMPs ) [10 , 27] . Invadosome dependent cell invasion is often seen in cancers but it is also observed during normal animal development[44]and angiogenesis as well as in immune surveillance by leukocytes [45] . Here , we demonstrate that upon exposure to ECM , the protozoan parasite forms similar structures in a Rab21 dependent manner . The host ECM is a blend of various proteins with fibronectin ( FN ) and collagen being the major constituents . The physical , topological and biochemical composition of the ECM is highly tissue specific and heterogeneous , varying even within the same tissue and with age of the animal [42 , 43] . In mammals , adhesion to FN is mediated by the α5β1 integrin heterodimer [25] and similar receptors are thought to exist in Entamoeba histolytica . At least two amoebic FN binding proteins have been identified: a 37kDa receptor and a 140kDa receptor ( β1EhFNR ) . β1EhFNR has been shown to have antigenic similarities to the mammalianβ1 integrin [5] . They are thought to regulate the binding of the pathogen to the host ECM[5 , 46] . Binding of the β1EhFNR to fibronectin leads to phosphorylation of several proteins , including FAK ( focal adhesion kinase ) , paxillin and vinculin [28] . Fibronectin can also induce major actin rearrangement in E . histolytica through Rho1 and ROCK dependent pathway[6] . Here , we showed that it promotes the formation of focalized actin dots in the parasite in a Rab21 dependent manner ( Fig . 2A ) . In contrast , the other major ECM component , collagen typeI suppresses the Rab21mediated actin dot formation ( Fig . 3C and 3D ) . The molecular details for the process would require further in depth studies . The possibility of both the ECM components triggering different signaling pathways in the pathogen cannot be overlooked . Thus , the variable ECM composition between individuals may have an implication in their relative susceptibility towards amoebic infection . Binding of GalNAc and MAb αHgl 3F4 also lead to suppression of Rab21 induced actin dots , in a similar manner as observed for collagen type I . Collagen type I is a bonafide ligand for the Gal/GalNAc lectin complex [37 , 38] . Therefore , based on our observations we propose that the inhibitory effect of collagen type I may be mediated via its ability to bind to the CRD of the Hgl subunit . Possibly , the binding of ligands including collagen type I to the Hgl leads to conformational changes which in turn lead to suppression of actin dots . Interestingly , in the past αHgl3F4 has been shown to enhance binding of the parasite to the host glycoproteins and mucin [39]; in our current experimental setup , effect of αHgl 3F4 on adhesion of the parasite to collagen type I and FN coated surface is not completely understood . Moreover , the Hgl mediated inhibition was not observed on the FN coated surface . The possibility of the collagen receptor being involved in the inhibition process cannot be overlooked as well [47] . In the recent past , Rab GTPases have taken centre stage as regulators of invadopodia in the mammalian cells , governing the trafficking of the matrix metalloproteinases [22 , 23 , 31 , 48] . They both , directly or indirectly , associate with different integrin heterodimers and spatially control their distribution along the migratory axis of the cell [12 , 24 , 25] . hRab5 , the principal early endocytic Rab , was shown to regulate formation and maturation of invadosome [23 , 48] . hRab21has been reported to regulate cell adhesion and migration in a β1 integrin dependent fashion [24] . hRab21 was also shown to regulatematrix remodeling by Cancer Associated Fibroblast ( CAFs ) and thereby augmenting cell invasion [49] . Although hRab21 has been shown to be associated with cancer cell migration and invasion , its role in formation or maturation of invadosomes is yet to be investigated . Entamoeba histolytica has an elaborate network of vesicular trafficking as reflected by the presence of more than 90 members of Rab family GTPases [50] . Till date , only few of Rab proteins have been characterized . Rab5 and Rab7A are known to regulate erythrophagocytosis [33] . Additionally , Rab7A is also involved in retromer dependent recycling of the hydrolase receptor [51] . Rab7A and Rab7B are also shown to jointly regulate the biogenesis of the phagolysosomal compartment in the parasite [52] . Rab11B plays a central role in the secretion of cysteine proteases , a major virulence factor of the parasite [53] . But like most other amoebic Rab family members the biological function of Rab21 still remains unknown . AmoebicRab21 unlike its mammalian homologue did not show any effect on the uptake of classical endocytic or phagocytic cargos ( manuscript under review , PNTD-D-14-01300 ) . In this study , we showed that Rab21acts as a downstream cue under the FN stimulated cytoskeletal remodeling in the parasite . The possibility of Rab21 regulating the trafficking of the EhFNR cannot be ruled out . In a previous report [46] , it was shown that FN stimulates protease secretion by the parasite which was further proposed to alter its pathogenic behavior . Here , in the current study , we demonstrated that the ECM degradative activity associated with the actin rearrangement is governed by Rab21 in a nucleotide ( GDP/GTP ) dependent manner ( Fig . 4A and 4B ) . Therefore , based on our observations , we hypothesize that Rab21 may lead to an overall increase in the proteolytic activity , thereby conferring an advantage of enhanced invasive capacity to the parasite . Taken together , our results show that fibronectin induces actin dots in E . histolytica in a Rab21 dependent process while collagen type I suppresses the process perhaps in Hgl dependent manner ( Fig . 6 ) . Further , the amoebic actin dots were found to be associated with ECM degradation . Hence , we propose that Rab21 may play an important role in in vivo tissue invasion and thereby modulate the virulence of the pathogen .
Entamoeba histolytica trophozoites of HM-1: IMSS cl6 and G3 strain[35] ( the G3 strain was generously gifted by Prof . David Mirelman , Department of Biochemistry , Weizmann Institute of Science , Israel ) were cultured axenically in BI-S-33 medium supplemented with 15% heat inactivated Adult Bovine Serum ( Sigma Aldrich ) at 35°C as described previously[54] . The cloning of Rab21 has been described previously ( Pathema ID: EHI_129330 , manuscript under review , PNTD-D-14–01300 ) . Briefly , total RNA was extracted using RNA easy kit ( Qiagen ) and cDNA prepared using the High Capacity RNA to cDNA kit ( Applied Biosystems ) . Rab21 was amplified from the cDNA pool of Entamoeba histolytica using the following primer set; forward 5-ATGGAAAACGAATTTAAAGTAGTTTTGTTG-3 and reverse 5-TTAACAACAATCAGATTTTGCTTGACGAGT-3 and cloned into InsaTA cloning vector ( Thermo Fisher ) . It was then further subcloned into the amoebic expression vector pEhExHA using SmaI/XhoI . For generation of the Rab21 knockdown strain the following primer set was used; forward 5-TTCAGGCCTATGGAAAACGAATTTAAA-3 and reverse 5-TTCGAGCTC AGCTTCTTCTTTTGAAAT-3 . Briefly , total RNA was extracted from HM-I: IMSS cl6 strain using RNA Easy kit ( Qiagen ) and cDNA was prepared using High Capacity RNA to cDNA kit ( Applied Biosystems ) . Further , using standard PCR procedure a 400bp segment of Rab21 was amplified using the above set of primers from the cDNA pool and cloned intopsAP2Gunma using StuI/SacI to generate the gene silencing plasmid psAP2GunmaRab21[55] . The control strain used for the study was G3 transfected with empty psAP2Gunma . The nucleotide hydrolysis deficient , Rab21CA ( Q64L ) and the nucleotide binding deficient , Rab21DN ( T18N ) mutants were generated using Site Directed Mutagenesis strategy ( SDM ) using the following primer sets; forward 5-GGGATACTGCAGGACTAGAAAAATACCAAGC-3 and reverse 5-GCTTGGTATTTTTCTAGTCCTGCAGTATCCC-3; forward 5-GAAGGAAAAGTTGGAAAGAATTCGATGATATTAAGA-3 and reverse 5-TCTTAATATCATCGAATTCTTTCCAACTTTTCCTTC-3 , respectively . pEhExHA Rab21WT was used as the template for the SDM . The following plasmids pEhExHA , pEhExMyc , pEhExHARab5WT , pEhExMyc Rab7AWT[52] , pEhExHA Rab11BWT[53] , pEhExHA Rab7BWT[52] and the gene silencing plasmid psAP2Gunma[55]were described previously . Logarithmic phase trophozoites ( HM1:IMSS for over-expression and G3 strain for gene silencing ) were electroporated with the following plasmids; pEhExHA/pEhExMyc ( vector control ) , Rab21WT , Rab21CA , Rab21DN , Rab5WT , Rab7AWT , Rab7BWT , Rab11BWT , psAP2Gunma and psAP2GunmaRab21 using Biorad MX Cell electroporator using the standard protocol . Briefly , logarithmic phase trophozoites were harvested and washed with chilled 1XPBS , followed by incomplete cytomix buffer ( 10mM K2HPO4 /KH2PO4 ( pH 7 . 6 ) , 120mM KCl , 0 . 15mM CaCl2 , 25mM HEPES ( pH 7 . 4 ) , 2mM EGTA , 5mM MgCl2 ) . The washed trophozoites were resuspended in chilled complete cytomix buffer ( incomplete cytomix buffer supplemented with 4mM ATP , 10mM reduced glutathione ) and 100μg of the respective plasmids were added and cells were electroporated at 500V , and 500μF and immediately transferred to warm complete medium . The transfected trophozoites were selected by adding 4μg/ml G418 ( Sigma Aldrich ) after 48hrs of transfection . The concentration of G418 was gradually increased to 20μg/ml in the following two weeks . All the transformants were maintained at 20μg/ml of G418 for all the experiments carried out . For generating the Rab21KD strain , logarithmic phase G3 trophozoites were electroporated with psAP2GunmaRab21 . The detailed methodology has been described previously[55] . The concentration of G418 was gradually increased to 20μg/ml in the following two weeks . All the transformants were maintained at 20μg/ml of G418 . Anti HA ( cat . no . sc-7392 ) and anti Myc ( cat . no . sc-40 ) mouse monoclonal antibodies were purchased from Santa Cruz . Anti Vinculin ( hVIN1 , cat . no . V9131 ) mouse monoclonal antibody was purchased from Sigma Aldrich . Alexa Fluor488 anti mouse and Alexa Fluor568 Phalloidin were obtained from Molecular Probes ( Invitrogen ) . Mouse αHgl antibody ( 3F4 ) was a generous gift from Dr . William Petri ( Department of Microbiology , University of Virginia ) . Fibronectin from human plasma ( cat . no . F2006 ) was purchased from Sigma Aldrich and collagen type I from rat tail ( cat . no . A10483-01 ) was obtained from Life Technologies . G418 ( cat . no . 1720 ) was purchased from Sigma Aldrich . GalNAc ( cat . no . A2795 ) was purchased Sigma Aldrich . All the chemicals used for the experiments were purchased from Sigma Aldrich . Amoeba transformants in logarithmic phase were harvested and transferred to an eight well glass slide and incubated at 37°C water bath for 30 min to let the trophozoites attach to the slide . Cells were fixed with 4% PFA at room temperature for 15 min , permeabilized with 0 . 1% tritonX100 , blocked with 5% FCS in PBS and stained with anti HA ( 1:250 ) or anti Myc ( 1:250 ) at room temperature for an hour , followed by Alexa fluorophore conjugated secondary antibodies ( 1:500 ) and Alexa568 Phalloidin ( 1:40 ) at room temperature for an hour . The cells were subsequently washed and then mounted in mowiol ( mounting medium ) and the slides were left overnight at RT for drying . Samples were examined on a Zeiss LSM780 confocal laser scanning microscope using 63x , 1 . 4 NA oil immersion objective . Images acquired were further quantified ( image based quantification ) by using Motion Tracking software freely available at the website http://motiontracking . mpi-cbg . de[56 , 57] . Amoeba transformants in logarithmic phase were harvested and transferred to an eight well glass slide and incubated at 37°C water bath for 30 min to let the trophozoites attach to the slide . The medium was removed and fresh medium containing different concentrations of GalNAc was added . Maltose at 100mM was included as a negative control . The cells were incubated for an additional hour at 37°C . After an hour , the cells were fixed , permeabilsed , blocked and stained for actin using Alexa568 Phalloidin ( 1:40 ) at room temperature for an hour . The samples were mounted in mowiol and dried overnight at RT; they were imaged using Zeiss LSM780 . The images obtained were analyzed and quantified using Motion Tracking software . Similar setup was used for GalNAc inhibition assay with MAb against Hgl except that the cells were incubated with medium containing αHgl 3F4 ( 1:30 ) for an hour at 37°C . Fibronectin and collagen type I was coated onto an eight well glass slide at a concentration of 100μg/ml under the laminar hood for 1 hr at room temperature . The excess of the FN and collagen type I was aspirated and the wells were washed gently with sterile PBS at RT . The slides were dried under the hood and used for further experiments . Rab21CA overexpressing cells and vector control were harvested at 400xg , 3min at RT and plated on the coated slides . The cells were further incubated at 37°C for an hour and then fixed , permeabilsed and stained with Alexa568 Phalloidin to visualize actin . Fluorescent matrix-coated coverslips were prepared and the assay carried out as described . Briefly , thin layer of Hilyte488 Fibronectin at 100μg/ml concentration ( cat . no . FNR02 , Cytoskeleton Inc . ) was coated on coverslips precoated with 0 . 01% of poly-L-lysine ( cat . no . P8920 , Sigma Aldrich ) which was cross-linked with 0 . 5% ice cold glutaraldehyde ( cat . no . G5882 , Sigma Aldrich ) for 15 min at 4°C . The coverslips were incubated in the dark for 3 hours at 37°C . Further , the coverslips were immersed in 5 mg/ml NaBH4for 3 min at room temperature . Finally , after a wash and short 10-min incubation in 70% ethanol , coverslips were quenched with complete TYI medium containing 15% adult bovine serum for 1 h at 37°C before cells were plated . Cells were then cultured on ECM-coated coverslips over periods ranging from 36 to 48 h under anaerobic conditions ( Gas Pak EZ cat . no . 260683 , BD Biosciences ) . Matrigel invasion assay was performed essentially as described previously [34] . Briefly , amoeba in the logarithmic growth phase were detached on ice and harvested , suspended in serum free TYI growth medium , and approximately 75 , 000 cells loaded in the upper chamber of a Transwell migration chamber ( 8μm pore size , cat . no . 353097 , Falcon ) coated with 5mg/ml of matrigel ( cat . no . 354277 , Corning ) . The lower chamber contained growth medium supplemented with 15% adult bovine serum . Transwell plates were incubated at 37°C for 48–50 hrs under anaerobic conditions ( Gas Pak EZ , cat . no . 260683 , BD Biosciences ) . Migrated trophozoites attached to the lower chamber wall were detached on ice , harvested , and counted manually using a heamocytometer . Rab21CA transfomants were plated on different ECM coated surfaces and incubated for 4–6 hrs at 35°C and processed for SEM analysis . Briefly , log phase trophozoites were harvested and washed with complete medium and then resuspended in warm complete medium containing 15% adult bovine serum . The trophozoites were plated on glass coverslips in a four well plate placed in a BD EZ Gas Pak and incubated at 37°C for an hour for attachment . Further the medium in the wells was replaced with medium containing 2 mg/ml of collagen type I and 1mg/ml of matrigel and trophozoites were incubated for additional 6–8hrs at 37°C in the GasPak . Thereafter , the medium was removed and the cells were briefly washed with warm 0 . 1M phosphate buffer ( pH 7 . 4 ) and fixed using 2 . 5% EM grade gultaraldehyde in 0 . 1M phosphate buffer ( pH 7 . 4 ) at 4°C for overnight . Following day , the cells were dehydrated in a graded series of alcohol ( 25% , 50% , 75% , 95% ) for 15min each at RT followed by 100% for 15 min at RT for three times . The samples were then left for drying at RT for 48–56hrs covered with an aluminum foil . The dried samples were sputter coated with gold using Quorum Q150R ES and were examined and photographed with Zeiss ULTRA PLUS field emission scanning electron microscope operating at 5kV . Unpaired two tailed Student’s t-test was performed using GraphPad Prism5 software . P values from student’s t-test: *P<0 . 05 , **P<0 . 01 and ***P<0 . 001 .
|
Entamoeba histolytica is one of the major causes of morbidity and mortality in developing economies . Severe amoebic infection leads to metastatic spread of the pathogen to extra intestinal sites , especially the liver , causing hepatic abscess . The migratory ability of the pathogen contributes to the spread of the disease . Here , we report that Rab21 , a Ras superfamily GTPase , promotes actin dot formation under the fibronectin induced signal in E . histolytica . The amoebic actin dots share structural components and functional properties to a class of actin rich degradative counterparts found in higher eukaryotes called “invadosomes” . Invadosomes are associated with both normal animal development and diseased state such as cancer . We also show that collagen type I , another major ECM component , suppresses the genesis of actin dots , possibly through interaction with amoebic cell surface adhesion molecule , the Gal/GalNAc lectin complex . Based on our observations , we propose that Rab21 may play an important role in in vivo tissue invasion by the parasite which may have further implication in its virulence .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Small GTPase Rab21 Mediates Fibronectin Induced Actin Reorganization in Entamoeba histolytica: Implications in Pathogen Invasion
|
Learning in a stochastic environment consists of estimating a model from a limited amount of noisy data , and is therefore inherently uncertain . However , many classical models reduce the learning process to the updating of parameter estimates and neglect the fact that learning is also frequently accompanied by a variable “feeling of knowing” or confidence . The characteristics and the origin of these subjective confidence estimates thus remain largely unknown . Here we investigate whether , during learning , humans not only infer a model of their environment , but also derive an accurate sense of confidence from their inferences . In our experiment , humans estimated the transition probabilities between two visual or auditory stimuli in a changing environment , and reported their mean estimate and their confidence in this report . To formalize the link between both kinds of estimate and assess their accuracy in comparison to a normative reference , we derive the optimal inference strategy for our task . Our results indicate that subjects accurately track the likelihood that their inferences are correct . Learning and estimating confidence in what has been learned appear to be two intimately related abilities , suggesting that they arise from a single inference process . We show that human performance matches several properties of the optimal probabilistic inference . In particular , subjective confidence is impacted by environmental uncertainty , both at the first level ( uncertainty in stimulus occurrence given the inferred stochastic characteristics ) and at the second level ( uncertainty due to unexpected changes in these stochastic characteristics ) . Confidence also increases appropriately with the number of observations within stable periods . Our results support the idea that humans possess a quantitative sense of confidence in their inferences about abstract non-sensory parameters of the environment . This ability cannot be reduced to simple heuristics , it seems instead a core property of the learning process .
Many animals , human adults and even human babies possess remarkable skills to cope with the pervasive uncertainty in their environment [1 , 2] . Learning processes are attuned to uncertainty . They enable one to capture the stochastic characteristics of the environment , as when one learns how often a probabilistic cue leads to a reward [3] . The environmental uncertainty actually occurs at several nested levels , as the stochastic characteristics themselves may also vary suddenly and without warning . The human learning is sophisticated enough to quickly adapt to such higher-order changes: the probabilities and characteristics that subjects learn are adequately fitted by statistical models [4–7] . However , in such tasks and environments flooded with uncertainty , subjects not only estimate the characteristics of the outside world , they also evaluate the degree of certainty that their estimates are accurate . This more subjective aspect of learning , the “feeling-of-knowing” , has received little attention so far . Here , we attempt to provide a formal account of this feeling and its origin . The feeling-of-knowing , or the sense of confidence , has been primarily demonstrated in memorization tasks [8] and in perceptual decision-making tasks in humans , monkeys and rodents [9–11] . By contrast , evidence from probabilistic learning tasks is currently limited . Many learning models actually simply do not consider feeling-of-knowing as a component of the learning process . Most share a common logic , according to which each parameter of the environment is represented at any given moment by a single numerical estimate and is continuously updated based on new observations . Rescorla and Wagner suggested a simple update rule: the point estimate should be shifted in proportion of the prediction error , i . e . the extent to which the estimate deviates from the new observation [12] . Such models therefore only provide point estimates , and they are devoid of any sense of uncertainty . It has been recognized more recently that the learning rate could actually be modulated as a function of an internal estimate of the environmental uncertainty , e . g . volatility [4 , 6] and that learning could even be fully reset when an environmental change is detected [7] . However , the normative Bayesian approach of learning suggests that there is a principled distinction between this environmental uncertainty and the uncertainty in the internal knowledge of what has been learned [13] . We term this second kind of uncertainty , the 'inferential uncertainty' . Despite evidence that the inferential uncertainty could affect learning in humans [5] , how humans perceive this uncertainty remains largely unexplored . Here , we suggest that the feeling-of-knowing , or subjective confidence , corresponds formally to the inferential uncertainty and that it derives from the inference that underpins the learning process itself . Indeed , the fact that humans have distinct degrees in their feeling-of-knowing suggests that they do not keep track of point estimates of environmental parameters , but instead of a set of estimates , each with its own degree of plausibility . Supporting this idea , some models assume that the brain infers full probability distributions [14] . The hypothesis was initially introduced for sensory representations , but it may be extended to higher-level tasks [15–17] , possibly including the learning of any numerical parameter . Following this hypothesis , learning in an uncertain world would be underpinned by a probabilistic inference that provides , not a single parameter value , but a distribution of possible values—and therefore affords an estimation of “feeling-of-knowing” based on the concentration of this inferred distribution onto a single value [18 , 19] . To test this idea , we examined whether humans can provide not only accurate estimates of environmental probabilities , but also accurate confidence ratings in those estimates . Such a finding would imply that the brain not only computes a point estimate , but also , at a minimum , the uncertainty in inferring its value , and perhaps even its full distribution . We designed a challenging probabilistic learning task with two nested levels of environmental uncertainty . Fig 1 shows how we generated the random sequences of visual or auditory stimuli and Fig 2A shows an example session . First , at any given moment , the sequence depends on two parameters: P ( A|B ) and P ( B|A ) , i . e . the transition probabilities between stimuli A and B . Second , these transition probabilities themselves remain stable only for a limited time , then change abruptly to a new random value , thus delineating ‘chunks’ in the sequence separated by ‘jumps’ . These jumps were aimed at inducing fluctuations in the inferential uncertainty over time . Subjects were asked to detect the jumps and , occasionally , to report their estimate of the transition probability to the next stimulus and their confidence in this estimate . Subjective estimates of transition probabilities can be compared to the true generative probabilities . However , this comparison is not completely fair because the generative parameters are not available directly to the subject , but can only be inferred from the specific stimuli received . Furthermore , confidence is simply not a characteristic of the generative process , but solely of the inference process . This highlights the need to derive both the estimates of transition probabilities and confidence levels in a principled manner from the inference itself . We therefore compare subjects' answers with the inference generated by an Ideal Observer endowed with the mathematically optimal inference process . This normative solution formalizes the link between the inference on the one hand , and the probability estimates and confidence levels on the other . Indeed , the optimal inference returns a distribution of likelihood over the transition probabilities , given the specific stimuli received . Both a point estimate and a confidence level in this estimate can be derived from this distribution . The distribution can be averaged to obtain a single best estimate of the transition probability . Confidence should reflect how precise this estimate is: whether the distribution is spread ( low confidence ) or concentrated ( high confidence ) around this estimate . We thus formalized confidence as the precision of the distribution ( its inverse variance ) , as previously suggested [19] . The Ideal Observer being normative , it provides a reference to assess the accuracy of the single point estimates and the fluctuations in confidence levels reported by subjects . In addition , since the Ideal Observer formalizes how single point estimates and confidence levels should derive from the inference process , it affords a series of predictions serving as tests of whether the reported estimates and confidence levels indeed derive from a common inference . And last , if confidence levels derive from an accurate inference , then they should reveal several specific properties of this efficient inference system .
We first asked whether subjects could detect when the characteristics of the sequence changed suddenly . We assessed the accuracy of their detection in comparison to the actual position of jumps with a Receiver Operative Characteristic analysis . Subjects reported more jumps when transition probabilities were indeed changing ( hit ) than when they were stable ( false alarm ) : the difference of hit minus false alarm rates was 0 . 23 ( standard error = ± 0 . 03; t-test against 0: p<10–5 ) . To show that this difference is positive not because of chance , but instead because the detection is based on the actual evidence provided by the observed sequence , we used a more conservative test . Comparison with surrogate data indicates that the observed difference between the hit and false alarm rates is significantly higher than expected from a random detection process ( p<0 . 01 , see Methods ) . Although the detection of jumps by subjects is better than chance , it is not perfect: some jumps were missed , and some others falsely reported . However , some of these errors precisely further demonstrate that subjects based their detection on the actual level of evidence received . Indeed , in principle , not all jumps can be detected equally easily: for instance , when changes in transition probabilities are small and frequent , the sequence may not provide enough evidence for the presence of each jump . The Ideal Observer provides a principled way of quantifying the likelihood of a jump at each position in the observed sequence . We tested whether subjects are sensitive to such fluctuations in evidence by analyzing their errors ( misses and false alarms ) from the Ideal Observer perspective . A significant difference in jump likelihood at the time of the subjects' Hits vs . Misses ( p = 0 . 003 ) indicates that subjects were more likely to miss a jump when the apparent jump likelihood was misleadingly low . Similarly , a difference between False Alarm vs . Correct Rejection ( p = 0 . 001 ) reveals that the subjects' false alarms were more likely to occur when jump likelihood was high ( see Fig 3 ) . Altogether , these results indicate that subjects partially managed to track the jumps in the objective generative process , and their responses give evidence of an efficient statistical use of the available information . We next examined whether subjects could estimate the characteristics of the sequence despite their unpredictable changes in time . The sequence was paused every 12 to 18 stimuli and participants were asked to report the probability that the next stimulus would be A or B . Subjects’ responses were correlated across trials with the true generative probabilities ( t17 = 8 . 8 , p<10–7 ) , indicating that subjects' probability estimates , although imperfect , consistently followed the generative probabilities . The deviations could reflect that the transition probabilities are inferred from the specific and limited amount of stimuli received . We therefore compared the subjects’ estimates of transition probabilities with the optimal values that could be inferred from the data , i . e . the parameter estimates inferred by the Ideal Observer . The subjects' responses were tightly correlated with the optimally inferred probabilities ( t17 = 8 . 5 , p<10–6 , see Fig 4A ) . When both predictors were included in a multiple linear regression , significantly higher regression weights were found for the optimal estimates than for the generative values ( paired difference of weights: t17 = 4 . 6 , p<10–3 ) . Given that the Ideal Observer and the subjects are both asked to estimate a probability , we can not only test whether their estimates are correlated , but also whether they are identical . Fig 4A reveals a remarkable match , although somehow imperfect: the observed slope is actually slightly below the identity . This deviation could reflect the distortion of subjective probabilities classically reported [20] . However , this pattern could also reflect differences in accuracy across trials and subjects . Indeed , the average of ideal estimates should be perfectly aligned on the diagonal , but the average of random estimates would form a flat line at 0 . 5; therefore a mixture of both should result in an intermediate slope . Supporting this view , the inspection of individual data revealed that the regression slopes were significantly larger than 0 in most subjects ( p>0 . 009 for 16 out of 18 subjects ) , but they were significantly equal to 1 in only 3 subjects ( in these subjects , Bayes factor > 9 , see [21] for the computation of this 'Bayesian t-test' ) . Together , these results show that subjects were able to infer the transition probabilities generating the observed sequence of stimuli despite their sudden changes in time . Not surprisingly , subjects were outperformed by the Ideal Observer endowed with the best inference scheme . However , the comparison to the optimum reveals a remarkable accuracy of the subjects' estimates . Subjects were also asked to rate their confidence in their probability estimates . They provided confidence ratings on a bounded qualitative continuum ( see Fig 1 ) . The absolute position of a given 'feeling-of-knowing' on this continuum is a matter of subjective representation , not a property of the inference process . However , if the confidence judgment reflects the certainty of the inferred probability estimate , then distinct confidence ratings should correspond systematically to distinct levels of evidence . Therefore , we assessed the accuracy of the fluctuations in confidence judgment with a regression against a principled measure of the level of evidence . Again , we used the Ideal Observer to this end . Intuitively , confidence should be high if and only if the estimated distribution of transition probability is concentrated on the reported value . This corresponds formally to the notion of precision , the inverse variance of the estimated distribution . Thus , we defined the Ideal Observer confidence as the negative log of the variance of the distribution . We used the log scale because it is the natural space for variance [22] . Note that the log variance and log standard deviation are strictly proportional , therefore the choice of one or the other provides the exact same significance levels in the regression analyses . We found a strong positive correlation between this principled measure of confidence and the subjective confidence ( t17 = 3 . 94 , p = 0 . 001; see Fig 4B ) . In addition , given that the experiment presented visual and auditory stimuli in separate blocks , we checked the robustness of the previous results by testing each modality separately . The regression of subjective estimates against the Ideal Observer was significant within each modality ( for probability estimates: both p<10–5; for confidence: both p<0 . 004 ) . Interestingly , these regression weights were positively correlated between modalities ( for probability estimates , Pearson ρ16: 0 . 55 , p = 0 . 017 , for confidence ρ16: 0 . 59 , p = 0 . 010 ) , supporting the idea that inferential capabilities vary between observers and are not tied to one modality but instead characterize a supramodal level of processing . Another way to evaluate the accuracy of confidence judgments is to ask whether they predict performance . Computing confidence can be useful when it serves as a proxy for the accuracy of performance—high subjective confidence should predict an objectively low rate of errors . We verified that this is true for the normative Ideal Observer: across trials , the magnitude of the error separating the Ideal Observer estimates from the true generative probabilities was negatively correlated with the Ideal Observer confidence ( t17 = -8 . 67 , p<10–6 ) . Crucially , a similar relationship linked the subjects’ confidence with the objective error in their probability estimates ( t17 = -2 . 27 p = 0 . 037 ) . We used simulations to check that this link derives from the normative nature of the subjects' estimates and not from biases in their probability estimates or confidence ratings . We used three separate simulations to reassign randomly one variable ( probability estimates , confidence ratings or true generative probabilities ) , while keeping the two others unaffected . Each simulation disrupts specific links between the generative probabilities and the subjective estimates to capture potential response biases . The simulations showed that the negative relationship observed between confidence and objective error is unlikely to emerge by chance on the sole basis of response biases ( all p<0 . 019 ) . Altogether , the findings indicate that confidence estimation is accurate: it relates linearly to the principled inference made by the Ideal Observer , and it is also correlated with objective performance . Our hypothesis is that estimates of transition probabilities and confidence ratings jointly derive from a single inference process . In other words , there is a common substrate for both estimates . An alternative hypothesis would be that confidence ratings are derived from the estimates of transition probabilities . Our hypothesis leads to several testable predictions that also rule out the alternative . We predict that probability estimates and confidence ratings should be partly related: when information is scarce , the optimal default estimate for transition probability is around 0 . 5 and confidence is low . Extreme estimates ( toward 0 or 1 ) are achieved only when there is substantial evidence and hence when confidence is high . Confidence should thus increase when the probability estimates depart from 0 . 5: this is a fundamental and inescapable property of probabilistic reasoning . The Ideal Observer estimates robustly showed this U-shape pattern ( quadratic weight: t17 = 16 . 1 , p<10–11 ) , and so did our subjects ( t17 = 9 . 77 , p<10–7 ) . This effect was actually significant within every subject ( all p<0 . 0025 ) . However , we also predict that this U-shape relationship should be only partial , since in principle , one may be more or less confident in any probability estimate , depending on the number of observations that support it . To illustrate this property , we binned the participants' confidence ratings by their subjective probability estimates , and within each bin , we then sorted trials by high and low Ideal Observer confidence with a median split ( Fig 5A ) . Subjective confidence reflected the Ideal Observer confidence on top of the general U-shape pattern . To quantify this additional effect , we performed a multiple regression of subjective confidence , without binning , including as predictors both the subject's U-shape transformed probability estimates and the optimal confidence . The data revealed that the Ideal Observer confidence indeed captures aspects of subjective confidence ( t17 = 3 . 12 , p = 0 . 006 ) that are not accounted for solely by a quadratic effect of probability estimates . In our experiment , subjects reported their probability and confidence estimates sequentially . We therefore ran a control experiment to check that the nested relationship between probability estimates and confidence ratings ( as shown in Fig 5A ) is a general property of human reasoning which cannot be attributed to sequential reporting . New subjects performed a variant of the task in which the first question about the probability estimate was omitted ( see Methods ) . Subjective confidence rating still followed the Ideal Observer confidence ( t20 = 7 . 00 , p<10–5 ) . As expected from a probabilistic inference , subjective confidence also showed a quadratic effect of the optimally inferred probability ( t20 = 6 . 97 , p<10–5 ) . In addition , subjective confidence still co-varied with the Ideal Observer confidence on top of the quadratic effect of the optimal probability ( multiple regression: t20 = 3 . 01 , p = 0 . 007 , t20 = 5 . 13 , p<10–5 respectively ) . Our main experiment enables to further test the predictions of our hypothesis concerning the common origin of probability and confidence judgments . If probability estimates and confidence ratings both derive from the same inference , then we also expect that subjects who perform the inference accurately should perform accurately in both estimating probabilities and rating confidence . We defined how accurate subjects were in estimating probabilities and rating confidence with respect to the Ideal Observer . In both cases , accuracy was summarized as the correlation coefficient between the subjects’ response and the optimal response . We found a positive correlation across subjects between the accuracies of probability estimates and confidence ratings ( Pearson ρ16 = 0 . 67 , p = 0 . 002 , Fig 5B ) . We also tested whether this correlation was significant within subjects . On each trial , we computed the accuracy of probability estimates ( or confidence ratings ) as the distance between the Ideal Observer and the subject's responses ( see Methods ) . Again , the two accuracies were significantly correlated across trials ( t17 = 3 . 27 , p = 0 . 005 ) . Note that these correlations are also consistent with the alternative hypothesis that confidence ratings are derived from probability estimates . However , the within-subject data disprove this alternative . Indeed , we controlled that the within-subject correlation we found is not confounded by an effect of an estimation-to-confidence mapping ( be it quadratic or not ) by comparison with two shuffled data sets ( see Methods and Fig 5C ) . Altogether , these results show that probability estimates and confidence ratings are likely to derive from a common inference . In particular , the accurate confidence ratings reflect additional features of the inference that are not reflected in the probability estimates . We now examine whether the data provide cues as to how confidence is computed . The inference should use the incoming data to constantly update an internal model of the hidden process that could have generated the observed sequence of stimuli . There are normative principles ruling this update process . Therefore , any efficient algorithm should have specific characteristics . We show that confidence ratings reveal three properties expected from an efficient information processing system . First , whenever the probability estimates change a lot , indicating a severe revision of the internal model ( for instance , after a jump ) , then confidence should be low; conversely , when estimates are stable , confidence in the seemingly 'good' value should be high . Questions being asked only occasionally to the subjects , the subjective model revision cannot be estimated from their reports . Instead , we estimated the degree of model revision from the Ideal Observer . This ensures in addition that subjective confidence is regressed against a normative estimate in every subject . We observed the predicted negative correlation between subjective confidence and the amount of revision in the probability estimates relative to the previous observation ( t17 = 3 . 67 , p = 0 . 002; Fig 6A ) , indicating that subjective confidence tracks the revision of an internal model . Second , the number of data samples accumulated since the last detected jump should affect the level of confidence: more samples should lead to more precise estimations . We counted the cumulative number of samples between the optimally detected jumps . As predicted , we observed a positive correlation between subjective confidence and the number of samples since the last jump ( t17 = 3 . 51 , p = 0 . 003; Fig 6B ) , indicating that subjective confidence increases with the accumulation of evidence . Again , using the Ideal Observer to estimate the number of samples in the current chunk provides a normative comparison across subjects . Instead , using the subjects' jump detection entangles several factors , e . g . whether subjects are accurate and conservative in reporting jumps . The same analysis based on the subjects' jump detection however also revealed a positive correlation ( t17 = 2 . 26 , p = 0 . 037 ) . Third , confidence should be lower when the estimation of the model is made more difficult by decreasing the predictability of the sequence . Formally , the unpredictability of a sequence is characterized within a chunk by the entropy of the generative transition probabilities: it is maximal when the transition probability is 0 . 5 and it decreases as the transition probability goes toward 0 or 1 . Note that we quantify here the generative environmental uncertainty , not its subjective estimate ( as in Fig 5A ) . We therefore examined if confidence was negatively correlated with this entropy . As predicted , a negative correlation was observed ( t17 = -5 . 58 , p<10–4; Fig 6C ) . As a control , we examined if a similar effect occurred when computing the entropy of the other , currently irrelevant transition probability ( transition from the stimulus which was not presented on the previous trial ) . No significant effect was found ( t17 = -0 . 76 , p = 0 . 50; Fig 6D ) . The results therefore indicate that subjects keep a distinct record of the confidence attached to each of the two transition probabilities that they are asked to estimate . We checked that the results presented in Fig 6 survive correction for multiple comparisons and partial correlations by including the four regressors into a multiple regression of confidence levels . The three factors of interest were still significant ( amount of model revision needed: p = 0 . 006; number of samples received: p = 0 . 042; entropy of the relevant transition: p = 10–5 , and not the irrelevant one: p = 0 . 3 ) . We also confirmed that these results coincide with the normative theory by running the same analysis on the Ideal Observer confidence ( effect of the 3 factors of interest: |t17|>8 . 7 , p<10–7 , no effect of the irrelevant transition entropy: t17 = -0 . 3 , p>0 . 7 ) . These results support the idea that confidence ratings derive from a rational process that approximates the optimal probabilistic inference . Our hypothesis is that confidence and probability estimates both derive from the probabilistic inference itself . An alternative is that subjective confidence is derived independently with a valid heuristic [23 , 24] . In the current experiment , the probability estimate for instance is a rational cue for confidence: as discussed with Fig 5A , there is a strong and principled correlation between confidence and how much the probability estimate departs from 0 . 5 . However , we showed ( in Fig 5A and 5C ) that the accuracy of confidence judgment goes beyond this kind of mapping . This therefore precludes that subjective confidence derives only from a heuristic based on the probability estimate . An example of such heuristic would be to count the number of correctly predicted stimuli in the immediately preceding trials to determine a confidence level . We then showed that confidence is also systematically impacted by the entropy of the generative transition probability , the amount of samples accumulated in the current chunk and the degree of revision of the probability estimates ( Fig 6 ) . At a minimum , these results imply that confidence arises from a sophisticated heuristic that combines the above factors . However , we can prove here that human confidence ratings are more accurate than such a heuristic would predict: even after regressing out the effect of the above factors , the residual subjective confidence still co-varied significantly with the Ideal Observer confidence ( t17 = 2 . 89 , p = 0 . 01 , Fig 6E ) . What additional features of the inference process could explain this finding ? In deriving the “number of samples” heuristic , we assumed that subjects discretize the incoming sequence into discrete chunks separated by jumps , and that this process allows them to track how much evidence they received since the last jump . This heuristic is suboptimal , however: the optimal inference avoids any discrete decision , but computes with the full probability distribution that a jump occurred at any moment , and uses it to weight recent evidence . To evaluate whether human subjects integrate jump likelihood into their confidence estimates , we computed , on each trial , the current uncertainty on the location of the last jump . We quantified it as the variance of the current chunk length estimated by the Ideal Observer , normalized by its mean value ( over similar positions in the sequence across sessions and subjects ) so that values higher than 1 indicated that it was less clear than average when the last jump occurred . Subjective confidence correlated negatively with this uncertainty on jump location ( t17 = -3 . 12 , p = 0 . 006 ) , exactly as expected from a normative viewpoint ( same analysis with Ideal Observer instead of subjective confidence: t17 = -5 . 71 , p<10–4 ) . It therefore seems that subjects are able to factor an estimate of jump probability in their confidence judgments . Altogether , these results suggest that the inference underpinning learning in this task is a probabilistic computation .
Several classifications of uncertainties have been proposed [25] . Our distinction between environmental and inferential uncertainties is close to Kahneman and Tsersky’s [23] classical division of external uncertainty ( stochastic nature of the environment ) versus internal uncertainty ( state of knowledge ) . A similar distinction is also made in recent computational works , e . g . in [13] , the environmental uncertainty would correspond to the 'risk' and 'unexpected uncertainty' , the inferential uncertainty to 'estimation uncertainty'; in [5] a similar distinction is made . Internal uncertainty is sometimes called ambiguity , in particular in economics , when it characterizes the absence of knowledge [25 , 26] . Our terminology ( environmental vs . inferential uncertainties ) , stresses that these two kinds of uncertainties differ in their epistemic nature . By operationalizing this distinction , our study revealed how they are only partially related . We built upon previous paradigms that manipulated environmental uncertainty [4 , 7] in order to induce frequent variations in inferential uncertainty . We showed how a first-order environmental uncertainty ( probabilistic transitions between stimuli ) increases the inferential uncertainty , and how a second-order environmental uncertainty ( unexpected changes in these transition probabilities ) produces additional fluctuations in inferential uncertainty over time . The fact that environmental and inferential uncertainties are only partly related is particularly salient in our task when a transition probability is 0 . 5 . Such probability produces the least predictable outcomes ( high environment uncertainty ) and a precise estimation of this probability needs more samples than any other probabilities ( hence , a high inferential uncertainty ) . However , with a large number of observations , one can get quite confident that the outcomes are indeed completely unpredictable . All these effects were observed in a normative Ideal Observer model , and subjects' confidence faithfully tracked ideal-observed confidence . Thus , human adults possess sophisticated mechanisms for tracking their inferential uncertainty . Juslin & Olson [27] made a different distinction , separating Brunswikian uncertainty , independent from us and in that sense 'external' , and Thurstonian uncertainty , due to the imprecision of our information-processing systems . While Thurstonian uncertainty may have contributed to the small deviations that we observed between subjective confidence and the optimal observer , we stress here that learners are uncertain , not only because they are faulty , but primarily because inference from stochastic inputs is by essence uncertain . The Ideal Observer quantifies this irreducible level of inferential uncertainty that any learner must face in our task . It is an open question whether and how humans may combine this core inferential uncertainty with the additional uncertainty arising from their cognitive limitations . Broadly defined , confidence indexes a degree of belief in a particular prediction , estimation or inference [19 , 23 , 25] . What confidence is about may thus vary drastically , from mere detection ( feeling of visibility , e . g . [28] ) , to accuracy in perceptual tasks [9 , 10 , 29] , in memory retrieval [8] , or in response to general-knowledge questions [30 , 31] . Mathematical concepts clarify how the present work differs from these previous studies . In most studies , confidence can be formalized as the likelihood of some binary variable e . g . the posterior probability that a response is correct/incorrect , a stimulus is seen/unseen , etc . [9] . By contrast , here we investigated confidence in a continuous numerical quantity ( the inferred transition probability ) , so that a principled and natural formalization for the strength of evidence is , as suggested previously [19] , the precision of this variable ( its inverse variance ) . This computational distinction , in comparison with most previous studies , entails a noticeable difference in practice . In typical binary decision tasks , the accuracy of subjective confidence is estimated by comparison with the actual performance of the subject . This estimation may be more or less susceptible to biases [32] . In our task , confidence is defined as the precision of the variable inferred , and is therefore amenable to a principled quantification with the Ideal Observer . Therefore here , the accuracy of subjective confidence can be estimated by comparison with this optimal confidence . Crucially , this estimation is independent from the performance in the primary estimation task , which may even remain unknown to the experimenter . One could disagree with our particular formalization of confidence , and suggest alternative mathematical quantities such as the inverse variance ( not its log , as we did ) , or the posterior probability of the mean or of the maximum of the inferred distribution , or the entropy of this posterior distribution . All these metrics roughly quantify the same notion: they are highly correlated with the one we used , and running the analyses with these other metrics led to similar ( although less significant ) results . The tight correlation between the ideal-observer precision and human subjective confidence therefore strongly suggests that humans possess a remarkable capacity to extract and use probabilistic information . We assessed the accuracy of the subjective precision estimates based on their relative variations between trials . The metacognition literature however makes a classical distinction between whether the accuracy of confidence is only relative or also absolute [31] . Absolute confidence levels , and thus the identity between the subjective and the optimal levels , cannot be investigated in our design: indeed , mapping confidence onto a qualitative scale is subjective , not principled . Subjects may produce absolute confidence measures for binary variables , e . g . they may estimate the fraction of correct or seen trials , but asking them a numeric estimate of subjective precision seemed too difficult , which is why we resorted to a qualitative confidence scale . This aspect of our study leaves open the question of whether there is an internal scale for precision that could be sufficiently calibrated to be transferred between tasks [33] or even individuals [34] , as previously shown for binary judgments . Our estimation of the accuracy of subjective confidence relies on a comparison with an Ideal Observer . However , the literature on the perception of probabilities have evidenced frequent deviations from optimality , e . g . the over and under estimation of small and large probabilities [35 , 36] , and a bias toward the detection of alternation vs . repetition [37 , 38] . Whether adjusting the Ideal Observer to these biases could provide a tighter fit to subjective data is an open issue and a matter for further research . Different options are available to include these biases in the ideal observer model . One possibility is that only the report of the probability is distorted . In that case , the inference , and hence the confidence levels , would remain unaffected . By contrast , the bias could affect a particular component of the inference itself . Potential targets for such distortions include ( 1 ) the likelihood of the current observation given some inferred probability estimate , which serves to update the posterior knowledge; ( 2 ) the posterior estimate itself , which serves to evaluate the likelihood of future observations; ( 3 ) the prior about the generative probabilities , which biases the inference at the beginning of each new sequence , but also at any time a jump in probabilities is suspected . These different potential sources of bias may result in quantitative differences in confidence levels , which could help to arbitrate between these scenarios . Our results reveal some characteristics of the computation of confidence in humans . One possibility is that second-order estimates occur independently from the first-order estimates , by relying on indirect cues or heuristics such as reaction time in the first-order task [23 , 24] . However , several aspects of our results contradict this view . First , the sophisticated heuristics we tested did not fully account for confidence reports; similar results were reported in the perceptual domain [39] . Second , the accuracies of the first and the second-order estimates were tightly correlated across trials and subjects which contradicts that confidence levels occur independently . The alternative view is that first and second-order processes are related , e . g . the second-order process relies on a readout of the same single-trial inferential data available to the first-order process [40–42] . Signal detection theory formalized this readout process in perceptual decisions , postulating that the second-order estimate corresponds to a statistical quantity ( d-prime ) characterizing the first-order process [32] . Our hypothesis extends this idea to the learning domain: learning could be supported by a probabilistic inference [17 , 43] , resulting in a posterior distribution whose mean and precision would yield , respectively , the first-order and second-order estimates . The terms first-order and second-order estimates may indeed be unfortunate , as they suggest a sequential process . It is in fact an open issue whether the primary response and the confidence in this response arise in parallel or serially , and from a single brain circuit or not [11 , 40] . Parallel extraction by distinct circuits could account for the fact that confidence and performance are often correlated , but still dissociable [44 , 45] , for instance in situations of speeded judgment [29] , overconfidence [46] , or when the accuracy of confidence is impaired while performance is preserved . By revealing some characteristics of the computation of confidence , our results may reveal some characteristics of the learning process itself . Indeed , if both the learned estimates and the assigned subjective confidence levels derive from the same inference , then investigating subjective confidence could provide critical insights on the learning process . It should be the case if subjective confidence levels reveal something more than what the learned estimates already reveal by themselves . We showed that it is the case: the accuracy of subjective confidence cannot be reduced to the accuracy of the learned estimates . This implies that the classic view of learning , exemplified by the Rescorla Wagner rule , according to which learning simply consists in updating parameter estimates , does not suffice—the brain also keeps track of the uncertainty associated with each value . Recent computational works have already started to revisit this classic learning model so as to incorporate notions of uncertainty [5 , 13] . Our results emphasize the need to investigate confidence as part of the learning algorithm . Future work should determine whether learning relies on simplified computations involving only summary statistics such as mean and variance [5] , on sampling schemes [17 , 47] , or on full computations over distributions [15] .
The study was approved by the local Ethics Committee ( CPP n°08–021 Ile de France VII ) and participants gave their informed written consent prior to participating . 18 participants ( 9 females , mean age 23 , sem: 0 . 74 ) were recruited by public advertisement . The task was delivered on a laptop using Matlab ( Version R2013a ) and PsychToolBox ( Version 3 . 0 . 11 ) . The experiment was divided into 4 blocks , each presenting a sequence of 380 stimuli ( denoted A and B ) . On alternated blocks , A and B were either auditory or visual stimuli perceived without ambiguity , see Fig 1 for a description and the timing . A fixation dot separated the visual stimuli and remained present during the auditory blocks . The modality used in the first block was counterbalanced over subjects . The sequence was generated randomly based on predefined transition probabilities between stimuli , e . g . an 80% chance that A is followed by A and a 30% chance that B is followed by A . These values are thus called 'generative transition probabilities' . The sequence was structured into chunks: transition probabilities were constant within chunks and changed from one chunk to the next at so-called 'jumps' . Chunk lengths were sampled from a geometric distribution , with an average chunk length of 75 stimuli . To avoid blocks without jumps , chunks longer than 300 stimuli were discarded . In each chunk , transition probabilities were sampled independently and uniformly in the 0 . 1–0 . 9 interval , with the constraint that , for at least one of the two transition probabilities , the change in odd ratio p/ ( 1-p ) relatively to the previous chunk should be at least 4 . The sequence was paused occasionally ( every 15 stimuli , with a jitter of ± 1 , 2 or 3 stimuli ) to ask subjects about their probability estimates and confidence ( see Fig 1 ) . Probing subjects more often would have provided more information on their internal estimates; however it would also have disrupted more their effort to integrate serial observations , which is critical to estimate transition probabilities . Asking every 15 stimuli is thus a compromise . The raw data are provided as Supporting Information ( S1 Dataset , see S1 Text for a description ) . 20 participants ( 12 females , mean age 25 , sem: 0 . 76 ) were recruited for the control experiment . The key difference compared to the main task was that subjects were only asked the confidence question . The other task parameters were identical , excepted a minor modification: subjects used a four-step scale instead of a continuous scale to report their confidence level . Subjects first performed one session of the main experiment which served as training . Then , they performed four sessions of the modified task . All participants received detailed explanation about how the sequences are generated . An interactive display made intuitive the notions of transition probabilities , jumps and randomness . Transition probabilities were framed as state-dependent probabilities: e . g . if the current stimulus is A , there is an 80% chance that it is repeated and a 20% chance that it changes for B . For each state ( 'after A' and 'after B' ) these contingencies were presented as pie-charts . Random sampling from these contingencies was illustrated as a 'wheel of fortune': a ball moved around the pie chart , with decreasing speed , and the final position of the ball determined the next stimulus ( A or B ) . Participants could repeat this process and simulate a sequence of stimuli until they felt familiar with the generative process . To introduce the concept of jump , a dedicated key press triggered a change in the pie-chart ( hence , in transition probabilities ) . During the task , subjects were instructed to report jumps . They could press a key at any moment to pause the sequence and access the bottom right-hand screen shown in Fig 1 . By adjusting the counter displayed , they specified when the jump occurred ( e . g . '13 stimuli ago' ) . It was made clear that 1 ) the estimation and confidence questions would be prompted automatically , 2 ) the occurrence of questions and jumps was predefined and independent so that it was unlikely that a question prompt would coincide with a jump and 3 ) answers in the task had no impact on the actual generative transition probabilities . We used two methods to analyze the accuracy of jump detection . The first is the classic approach of the Receiver Operating Characteristic ( ROC ) : the reported jumps were compared to the actual , generative jumps . The second approach is a follow-up of the ROC analysis , benefiting from the Ideal Observer perspective: the binary subjective reports ( there is a jump vs . there is not ) were compared with the continuous , normative posterior probability of a jump . For both approaches , we sorted the subjects' responses into hits and false alarms . Given the stochastic nature of the task , it is difficult to detect exactly when a jump occurred . Consider for instance the sequence: A1 B2 A3 A4 A5 A6 B7 A8 A9 A10 B11 B12 B13 B14 A15 B16 B17 B18 Subscripts indicate stimulus position and the italic font indicates the second chunk . These chunks were generated from the following transition probabilities: low for AB and high for BA from stimulus 1 to 9; high for AB and low for BA from stimulus 10 to 17 . The true generative jump occurred at stimulus 10 , yet it seems more likely to have occurred at stimulus 11: A9A10 better fits in the first chunk in which the AA transition rate is high . To circumvent this issue , we tolerated some approximations in the jump detection by counting a hit when there was a true generative jump within a window of ±5 stimuli around the reported jump location , and a false alarm otherwise . This same window size was used throughout our data analysis , and other choices did not change the qualitative findings . In line with the ROC approach , we computed , for each subject , the difference in hit rate minus false alarm rate , known as the informedness index . Informedness is bounded between -1 and 1 , with values higher than 0 denoting a detection better than chance; and lower than 0 a detection worse than chance . A t-test on informedness revealed that the mean value was significantly larger than zero . However , to make sure that such a result was unlikely to emerge by chance from the detection characteristics of our subjects and the generative structure of our sequences , we adopted a more conservative permutation-based approach . We computed a null ( chance-level ) t-value distribution for informedness by keeping subject reports unchanged but randomly regenerating ( 10000 times ) the stimulus sequence . The p-value reported in the text corresponds to the probability of observing a t-value equal or higher under the null distribution , indicating how likely it is that the result is due to chance . We followed up the results of the ROC analysis by inspecting the posterior probability of jump estimated by the Ideal Observer in trials corresponding to the subjects' hits , misses , false alarms and correct rejections . More precisely , since we tolerated a margin of ±5 stimuli in the subjects' jump detection , we compared the subjects' report with the posterior probability that a jump occurred in a window of ±5 stimuli around each observation , see Fig 2E for an example session . For hits and false alarms , we took the posterior probability of a jump at the position reported by subjects , given the sequence they had observed when they reported it . It is less straightforward for misses and correct rejections since , precisely , jumps were never reported at these positions . We thus estimated for each subject the typical latencies of jump report and we averaged over this list of latencies to compute the posterior probability of jump at each position corresponding to a correct rejection or miss . To assess the accuracy of the subjects' probability estimates and confidence ratings , we used several regressions against predictor variables . The significance of these regression analyses was estimated by computing regression coefficients at the subject-level as a summary statistic and then comparing these coefficients against zero with a two-tailed t-test at the group level ( t and p-values are reported in the text ) . All regression models included a constant and the z-scored regressors of interest . The multiple regressions corresponding to Fig 6 deserves more details . In Fig 6A the estimation revision is the absolute difference of the Ideal Observer probability estimates between two consecutive similar transitions . Consecutive transitions are not necessarily consecutive stimuli ( e . g . the transition 'from A' in ABBBBAA ) . In Fig 6B , the jump-wise count of samples was also made per transition type . For this count , a log-scale was used since it is an analytical result that , on average , confidence ( the Ideal Observer log-precision ) should increase linearly with the log-number of samples . We based this count on the Ideal Observer . However , the Ideal Observer does not estimate a binary variable ( there is a jump vs . there is not ) , instead it computes the continuous posterior probability that a jump occurs at each position of the observed sequence , and it revises this estimate each time a new observation is made . We therefore transformed the posterior probability estimates ( a two-dimensional matrix , see Fig 2E for an example ) into discrete jumps . The thresholded ( two-dimensional ) posterior probability serves to identify when the sequence should be interrupted to report a jump and what should be the location of the reported jump , e . g , report at trial W that a jump occurred at position Z ( thus W-Z trials ago ) . The posterior jump probability being relatively smooth ( e . g . in Fig 2E ) , the thresholding forms patches . Each of these patches corresponds to a jump; the reported W and Z corresponds to the coordinates of the upper limit of each patch . We used a Receiver Operating Characteristic to identify the threshold ( posterior probability = 0 . 25 ) that maximized the accuracy of this discretization , with respect to the actual generative jumps: we searched the threshold that resulted in the maximal difference between hit and false alarm rates . We took as an estimate of single-trial accuracy , the un-signed error ( i . e . the distance ) between the subject estimate and the Ideal Observer estimate . The probability estimates in both the subjects and the Ideal Observer are expressed on the same probability scale: they can be compared directly . This is not the case for confidence: the scale for the Ideal Observer is normative , it is the log-precision which can be potentially infinite; by contrast for subjects the scale was bounded and qualitative , the mapping between confidence levels and the scale is thus highly subjective . To express the Ideal Observer and the subject confidence on a common scale , we adjusted their offset and scaling based on a linear fit . For each subject , the single-trial accuracies in probability estimates and confidence ratings were taken into a Pearson correlation over trials . The resulting correlation coefficients could then have been taken into a classical t-test; however , we wanted to estimate to what extent the correlation would be positive due to a systematic mapping between probability estimates and confidence ratings . We thus devised two permutation-based estimations , each corresponding to a null-hypothesis distribution of the correlation of accuracies between probability estimates and confidence ratings . Shuffling #1 ( Fig 5C , middle ) preserved the mapping but disrupted the sequence , by keeping pairs of probability estimates—confidence ratings and shuffling their order in the sequence separately for the Ideal Observer and the subjects . Shuffling #2 ( Fig 5C , right ) disrupted both the mapping and the sequence by shuffling the trials independently for probability estimates and confidence ratings , thus removing any correlation between them . 10000 distinct permutations were used to estimate each null distribution . Given that the shuffling was applied within-subject , we computed the null t-distribution for the paired differences between 'Observed data' and 'Shuffling keeping pairs' . The 'Full shuffling' resulted in values close to 0 for all participants so that the estimated null t-distribution was equivalent to the parametric t-distribution; tests against the 'Full shuffling' null were thus classical t-tests against 0 . P-values in Fig 5C correspond to one-tailed t-test . We derived mathematically the optimal observation-driven estimates of the transition probabilities and jump locations: the so-called Ideal Observer . This optimal inference relies on Bayesian principles and returns a distribution of estimates p ( θ | y ) , i . e . the posterior distribution of the transition probability , θ , at each time step in the experiment , given the observed sequence of stimuli , y . From this distribution , we derive the expected value of the inferred transition probability: μ = ∫ θp ( θ | y ) dθ and the confidence in that estimation , which we defined as its log-precision: -log ( ∫ ( θ − μ ) 2 p ( θ | y ) dθ ) . We designed two algorithms for this Ideal Observer: a sampling approach and an iterative approach . The iterative approach was used to double check the sampling approach: both provided numerically similar values of probability estimates , confidence levels and jump location . The sampling approach explicitly computes the likelihood of possible decompositions of the sequence into chunks , whereas the iterative approach computes the likelihood that a jump occurred at any given position , independently from the other potential positions . The sampling approach is computationally slower but it allows a straightforward estimation of jump-related statistics used here: 1 ) The likelihood that a jump occurred around a given position , e . g . within a window of ±5 stimuli; 2 ) The variance of the estimated length of the current chunk , which reflects the precision of the knowledge of the observer about the last jump location . The derivation of each algorithm is presented in detail below . Computations were performed numerically in Matlab using regular grids . If we assume that the transition probabilities generating the sequence are stable over time , then the inference can be computed analytically: the posterior distribution is a function of the number of transitions observed in the sequence . The formula is derived in the first sub-section below . However , sequences in the task were generated with jumps . For a given partition , the inference of transition probabilities can be made chunk-wise using the above-mentioned formula . Such an inference is conditional in the sense that it is computed given a particular partition . However , the partition itself is unknown and must be inferred from the sequence observed . The estimation of the transition probabilities must therefore factor out the uncertainty in the partition , which is achieved by marginalizing the conditional inference over all partitions: p ( θ|y1 , … , yt ) =∑πp ( θ|y1 , … , yt , π ) p ( π|y1 , … , yt ) ( 1 ) Where y is the sequence of A and B stimuli , θ = [θA|B , θB|A] are the transition probabilities 'from B to A' and 'from A to B' respectively , and π is a partition describing the location of jumps . The 1st term of the sum is thus the conditional posterior distribution of transition probabilities given a particular partition of the sequence; the second term is the posterior probability of this partition . The sequence length being 380 , there are 2380 possible partitions of the data . The exact inference would require that we compute the sum over these 2380 partitions . It is computationally intractable and actually not necessary: most partitions are very unlikely and contribute little to the sum . The posterior distribution of transition probabilities can thus be approximated numerically by averaging the conditional posterior distributions of transition probabilities over a subset of partitions sampled uniformly [22] . The second subsection below shows how to sample uniformly from the posterior distribution of partitions . It is not necessary to decompose the sequence explicitly into a partition to compute the posterior θ distribution given the stimuli observed . Indeed , if we know θ at position t in the sequence , then at position t+1 , θ should remain the same if no jump occurred , or be different if a jump occurred . In case a jump occurred , the new θ is sampled from the prior distribution and the likelihood can be assessed given the ( t+1 ) -th stimulus . In that case , the observations made before t become no longer needed to estimate θ after t . This so-called Markov property makes it possible to estimate θ iteratively , by going forward: at stimulus t+1 , we update the estimate made at time t , based on the new observation . In the following we derive the forward algorithm to estimate θ , the transition probabilities . We also derive a backward algorithm to estimate the likelihood of jumps in the observed sequence . Note that both algorithms are provided with the exact same observations as those presented to the subject . In particular , the backward sweep does not benefit from extra stimuli not yet observed by the subject , but simply processes the information received by moving backward in time .
|
Learning is often accompanied by a “feeling of knowing” , a growing sense of confidence in having acquired the relevant information . Here , we formalize this introspective ability , and we evaluate its accuracy and its flexibility in the face of environmental changes that impose a revision of one’s mental model . We evaluate the hypothesis that the brain acts as a statistician that accurately tracks not only the most likely state of the environment , but also the uncertainty associated with its own inferences . We show that subjective confidence ratings varied across successive observations in tight parallel with a mathematical model of an ideal observer performing the optimal inference . Our results suggest that , during learning , the brain constantly keeps track of its own uncertainty , and that subjective confidence may derive from the learning process itself . Our results therefore suggest that subjective confidence , although currently under-explored , could provide key data to better understand learning .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
The Sense of Confidence during Probabilistic Learning: A Normative Account
|
Gibberellins ( GAs ) are a class of important phytohormones regulating a variety of physiological processes during normal plant growth and development . One of the major events during GA-mediated growth is the degradation of DELLA proteins , key negative regulators of GA signaling pathway . The stability of DELLA proteins is thought to be controlled by protein phosphorylation and dephosphorylation . Up to date , no phosphatase involved in this process has been identified . We have identified a dwarfed dominant-negative Arabidopsis mutant , named topp4-1 . Reduced expression of TOPP4 using an artificial microRNA strategy also resulted in a dwarfed phenotype . Genetic and biochemical analyses indicated that TOPP4 regulates GA signal transduction mainly via promoting DELLA protein degradation . The severely dwarfed topp4-1 phenotypes were partially rescued by the DELLA deficient mutants rga-t2 and gai-t6 , suggesting that the DELLA proteins RGA and GAI are required for the biological function of TOPP4 . Both RGA and GAI were greatly accumulated in topp4-1 but significantly decreased in 35S-TOPP4 transgenic plants compared to wild-type plants . Further analyses demonstrated that TOPP4 is able to directly bind and dephosphorylate RGA and GAI , confirming that the TOPP4-controlled phosphorylation status of DELLAs is associated with their stability . These studies provide direct evidence for a crucial role of protein dephosphorylation mediated by TOPP4 in the GA signaling pathway .
Gibberellins ( GAs ) are a class of major plant hormones mediating almost all physiological events during normal plant lifespan , including seed germination , leaf formation , cell elongation and flowering time control , etc [1]–[3] . In recent decades , several molecular components essential for GA signal transduction have been characterized using genetic and biochemical approaches [4] . One group of these components is nuclear-localized DELLA proteins . These proteins belong to a subset of GRAS family of putative transcriptional regulators that contain specific DELLA motifs at their N-termini and conserved GRAS domains at their C-termini . They are key repressors of the GA signaling pathway [2] , [5] . In Arabidopsis genome , there are five DELLA proteins , designated as GA INSENSITIVE ( GAI ) , REPRESSOR OF ga1-3 ( RGA ) , REPRESSOR OF ga1-3-LIKE protein ( RGL ) 1 , RGL2 , and RGL3 , respectively [6]–[9] . Genetic analyses indicated that these DELLAs have overlapping and sometimes distinctive roles in regulating plant growth and development . For example , GAI and RGA are important for stem elongation [10] , [11]; RGL2 regulates seed germination [9]; whereas RGA , RGL1 , and RGL2 are involved in floral development [12]–[14] . A major GA signaling cascade has recently been elucidated . In the nucleus , GA is perceived by its receptor , GIBBERELLIN INSENSITIVE DWARF 1 ( GID1 ) [15] , [16] . The formation of the ligand-receptor complex enhances the interaction of GID1 with the DELLA domain of DELLA proteins [17] , leading either to their direct inactivation [18] or ubiquitination by SCFSLY1/GID2 ( Skp1-Cullin-F-box protein complex ) E3 ligase [19]–[22] . Ubiquitinated DELLA proteins are subsequently degraded by the 26S proteasome system , triggering GA responses . In the absence of GA , on the other hand , DELLAs are stably localized in the nucleus where they interact with other transcription factors to inhibit the transcription of GA-responsive genes [23]–[27] , restraining growth and development processes in Arabidopsis [6] , [9] , [11] , [13] . DELLAs also promote the transcription of GID1b by interacting with other transcription factors , and maintain GA homeostasis by up-regulating the expression of GA biosynthetic genes GA 20-oxidases 2 ( GA20ox2 ) and GA 3-oxidases 1 ( GA3ox1 ) [28] . Moreover , DELLAs are important integrators of other phytohormones , including auxin , ethylene , abscisic acid ( ABA ) , brassinosteroid ( BR ) , and jasmonate ( JA ) [29]–[31] , and environmental factors , such as light [24] , [25] , cold [32] , and salt [33] . Very recently , these proteins were found to regulate cortical microtubule organization [34] . Besides ubiquitination and glycosylation [19] , [22] , [35] , limited evidence also suggested that DELLA proteins are regulated by reversible protein phosphorylation and dephosphorylation [36] , [37] . The detailed molecular mechanisms , however , are poorly understood and the protein phosphatases involved in this process have not been reported . Protein phosphatases 1 ( PP1s ) are a major group of serine/threonine ( Ser/Thr ) protein phosphatases . They are expressed ubiquitously in eukaryotes [38] , regulating diverse cellular processes in animals [39] , although their functions in plants are uncertain . In Arabidopsis , PP1s are referred to as type-one protein phosphatases ( TOPPs ) [40] . Previous studies indicated that they regulate embryonic development and blue light-dependent stomatal opening [41] , [42] . In general , molecular mechanisms of TOPPs in regulating plant growth and development are not well studied . Using a forward genetic approach , we identified that one of the nine TOPPs in Arabidopsis , TOPP4 , is involved in GA signal transduction . Biochemical analyses revealed that TOPP4 directly interacts with and dephosphorylates DELLA proteins RGA and GAI , promoting the GA-induced destabilization of these two proteins . A novel regulatory mechanism for protein dephosphorylation in the GA signaling pathway via TOPP4 is proposed .
We isolated an extremely dwarfed mutant from a 2000 M2 ethyl methane sulfonate ( EMS ) -mutagenized Arabidopsis population . The dwarfed plant was back-crossed three times with wild-type Col-0 and the resulting mutant was used in all studies presented . The mutant exhibits lack of apical dominance and aberrant leaf phyllotaxy ( Figure 1A–C ) . Compared to wild-type plants , the mutant has tiny , curled , and dark-green rosette leaves ( Figure 1A–C ) , delayed flowering ( Figure S1 ) , smaller flowers with irregular and narrow sepals ( Figure 1D ) , partially twisted petals and siliques ( Figure 1D–E ) , reduced mature pollen grains in anthers ( Figure 1F ) , and fewer seeds in mature siliques ( Figure 1G ) . This mutant resembles GA deficient or signaling mutants , since the dwarfism , reduced rosette radius , delayed flowering time , and high chlorophyll content of the mutant are similar to those of ga1-3 , ga4 , gai-1 , and gid1a/b/c [16] , [43] , [44] ( Figure S1 ) . The mature heterozygous mutant plants were semi-dwarfed with clustered siliques and no apical dominance , suggesting that the traits were inherited in a semi-dominant manner ( Figure 1B ) . When the mutant was back-crossed to Col-0 , the F2 population resulting from self pollination had a segregation ratio of 97∶257∶118 ( normal plants∶semi-dwarf plants∶dwarf plants ) , close to the expected 1∶2∶1 segregation ratio for a semi-dominant single locus . Map-based cloning was employed to identify the gene responsible for the mutant phenotype ( Figure 1H ) . The mutant in Col-0 ecotype background was crossed to Landsberg erecta-0 ( Ler-0 ) . The extremely dwarfed homozygous plants were selected from the F2 population for mapping . The locus was roughly mapped to a site close to a known marker nga168 on chromosome 2 . Eight newly developed insertion/deletion ( In/Del ) and cleaved amplified polymorphic sequence ( CAPS ) markers were then used for fine mapping ( Table S1 ) . The corresponding locus was eventually mapped to a 90-Kb region between markers T5I7-29008 and T28M21-47168 , with three and two recombinants for each marker , respectively , in the population of about 1000 individuals . This 90-Kb region contains 32 gene loci according to the gene annotation data obtained from the Arabidopsis genome database . We sequenced all 32 genes , and found a G to A single-nucleotide substitution in At2g39840 that resulted in the conversion of threonine ( Thr ) to methionine ( Met ) in amino acid 246 which is near the C terminus of the predicted protein sequence ( Figure 1I ) . At2g39840 encodes a protein previously named TOPP4 [40] . The mutant was therefore designated as topp4-1 . The single nucleotide substitution of topp4-1 did not influence TOPP4 gene transcription and its protein level ( Figure S2 ) . Our genetic result suggested that topp4-1 protein should have a dominant-negative effect . To confirm that , we made transgenic plants by introducing a construct containing the full-length cDNA of topp4-1 driven by a cauliflower mosaic virus ( CaMV ) 35S promoter ( 35S-topp4-1 ) into wild-type plants . More than 20 independent transgenic lines were obtained . All of them showed dwarfed phenotypes similar to those of the topp4-1 mutant plants , such as lacking apical dominance , abnormal leaf phyllotaxy and curled rosette leaves , and reduced sterility ( Figure 2A ) . The severity of the defective phenotypes appeared to be positively correlated with the expression levels of the topp4-1 gene ( Figure 2A–B ) . Most of the transgenic seedlings died before flowering . We also generated a construct containing the cDNA of topp4-1 driven by its own promoter ( pTOPP4-topp4-1 ) and transformed it into wild-type plants . The pTOPP4-topp4-1 transgenic plants showed topp4-1 mutant-like phenotypes ( Figure 2A ) . Furthermore , we constructed two different artificial microRNA ( amiRNA ) vectors that specifically target the TOPP4/topp4-1 gene ( amiR-TOPP4-1 and amiR-TOPP4-2 ) [45] . The topp4-1 mutant could be partially rescued regarding to inflorescence height , rosette leaves , and flowering time when the mutated topp4-1 gene was knocked down by amiRNA ( Figure 2C ) . The recovery effect was positively correlated with the knocked down level of the topp4-1 gene ( Figure 2D ) . These results clearly indicated that the mutated topp4-1 protein caused a dominant-negative effect on plant growth . To confirm whether the single nucleotide substitution of TOPP4 was responsible for the defective phenotypes , we transformed topp4-1 plants with a 1 . 6-Kb TOPP4 genomic fragment under the control of CaMV 35S promoter ( 35S-TOPP4 ) using an Agrobacterium tumefaciens-mediated floral dipping method [46] . Ten independent T1 transgenic lines were obtained , all of which showed semi-dwarfed phenotypes . Five independent lines apparently containing a single insertion were selected for further studies . Approximately 25% of the plants from each of the four lines ( #2–#5 ) had obviously complemented phenotypes in the T2 generation: they were significantly taller than the topp4-1 background but still a little shorter than wild-type plants ( Figure 3A ) . However , one line ( #1 ) had a weak complemented phenotype ( Figure 3A ) . Subsequent quantitative reverse transcription-polymerase chain reaction ( qRT-PCR ) revealed that this line had relatively low expression level of TOPP4 compared to the other four lines ( Figure 3B ) . Therefore , it seemed that the inflorescence heights of the transgenic plants were positively correlated with the expression levels of TOPP4 ( Figure 3A–B ) . The point mutation of TOPP4 was responsible for the extremely dwarfed phenotype of topp4-1 . In addition , we transformed TOPP4 gene under the control of its own promoter ( pTOPP4-TOPP4 ) into the topp4-1 mutant . The T2 transgenic lines showed increased rosette width than the topp4-1 mutant , but the height and the curled leaf phenotypes were not altered ( Figure S3 ) . In order to analyze the phenotype of the loss-of-function mutants , we searched the SALK and GABI-Kat T-DNA insertion databases for T-DNA insertion alleles of AT2g39840 . Two independent T-DNA lines , SALK_090980 and N466328 , were identified by PCR-based analyses ( Figure S4A–C ) . Neither of them had obvious mutant phenotypes . In SALK_090980 , the T-DNA is inserted 92 nucleotides upstream of the initiation codon ATG ( Figure 1I ) . This insertion did not alter the transcription level of TOPP4 ( Figure S4D ) . In N466328 , the T-DNA is located in the 3′ untranslated region , 23 bp after the stop codon TGA ( Figure 1I ) . qRT-PCR analysis showed that , the expression level of TOPP4 is decreased in this mutant ( Figure S4D ) . N466328 did not show obvious phenotype , possibly because it is a knock-down rather than a null mutant . It still expresses about 40% of the wild-type level of TOPP4 . To test whether further reducing the expression level of TOPP4 can finally result in a dwarfed phenotype , we transformed the two amiR-TOPP4 vectors into wild-type plants . Four plants from 86 T2 transgenic lines of amiR-TOPP4-1 and five plants from 128 T2 transgenic lines of amiR-TOPP4-2 exhibited dwarfed phenotypes . Three representative plants , amiR-TOPP4 #1-1 , #2–7 and #2–38 , were selected for subsequent analyses . They showed shorter inflorescences , curled leaves , decreased fertility , and retarded growth ( Figures 3C–D and S5 ) . Compared to wild type , the expression level of the TOPP4 gene in these amiRNA transgenic lines was decreased to about 30% of their wild type counterpart ( Figure 3E ) . Correspondingly , TOPP4 protein also dramatically declined in these lines ( Figure 3F ) . To investigate the effect of TOPP4 on plant growth and development , we transformed wild-type plants with the 35S-TOPP4 construct . More than 20 independent lines were obtained . The constitutive expression of TOPP4 in these lines was confirmed by qRT-PCR ( Figure 2E ) . Interestingly , all of them had enlarged organs compared to wild-type plants ( Figure S6A ) . One of the representative transgenic lines ( #7 ) with the highest expression level of TOPP4 was selected for subsequent analyses ( Figure 2F ) . Overexpression of TOPP4 in wild-type plants resulted in elongated hypocotyls , increased plant height , thickened stems , and enlarged rosette leaves , inflorescences , flowers and siliques ( Figures 2F and S6B–H ) . To understand the expression patterns of TOPP4 , a 2-Kb fragment upstream of the translation initiation codon ATG of the TOPP4 gene was fused to the β-glucuronidase ( GUS ) reporter gene in the binary vector pCAMBIA 1300-GUS ( pTOPP4-GUS ) and this vector was transformed into wild-type plants . More than 20 transgenic lines were obtained , and T2 generation transgenic plants were used for GUS staining analyses . Because all transgenic lines showed consistent expression patterns , only one representative line , pTOPP4-GUS #8 , was used for further analyses . In young seedlings , GUS staining was detected in the stele of roots and hypocotyls ( Figure 4A–C ) , the vascular bundles of cotyledons ( Figure 4B ) , and newly emerging leaves ( Figure 4C ) . In mature leaves of 3-week-old plants , GUS expression was observed mainly in tips , blades ( Figure 4D–E ) , stomata ( Figure 4F ) , and the base of trichomes ( Figure 4G ) . Cross sections of the rosette leaves also revealed GUS activity in vascular bundles and mesophyll cells ( Figure 4H ) . Furthermore , GUS staining was observed in pistil and stamen filaments of flowers ( Figure 4I–J ) , as well as the apex and the base of elongating siliques ( Figure 4K ) . In conclusion , TOPP4 is ubiquitously expressed in various organs throughout development , suggesting its diverse and crucial functions in plant developmental processes . qRT-PCR analyses were consistent with the GUS staining results and showed relatively higher expression levels of TOPP4 in stems , rosette leaves , and young siliques ( Figure 4L ) . To reveal subcellular localization of the TOPP4 protein , a yellow fluorescent protein ( YFP ) -tagged TOPP4 was transiently expressed in mesophyll protoplasts from wild-type plants . TOPP4-YFP protein was ubiquitously distributed in cells . It was mainly localized in the nucleus and at the plasma membrane ( Figure 5A ) . TOPP4-YFP signals were also found in cytoplasm ( Figure 5A ) . Analyses of the transient expression pattern of a green fluorescent protein ( GFP ) -tagged TOPP4 in Nicotiana benthamiana leaves confirmed the subcellular distribution of the TOPP4 protein ( Figure 5B ) . The localization of TOPP4 to the plasma membrane was also verified by plasmolyzing roots of 10-day-old 35S-TOPP4-GFP plants with 0 . 8 M mannitol for 1 h . Confocal analysis of these roots revealed that TOPP4-GFP was associated with the plasma membrane ( Figure 5C ) . This result was further confirmed by an immunoblotting assay using purified plasma membrane fraction ( Figure 5D ) . The morphological similarity of topp4-1 plants to GA deficiency or signaling mutants prompted us to examine whether the GA signal transduction is altered in topp4-1 . Therefore , responses of the mutant to exogenously applied GA3 were analyzed . Although GA3 rescued the dwarfed phenotype of ga1-3 as previously reported [47] , it did not affect the stem elongation of topp4-1 ( Figure 6A ) . Moreover , GA3 increased the transcription level of GA responsive genes EXPANSIN A8 ( EXP8 ) and PACLUBUTRAZOL RESISTANCE 1 ( PRE1 ) in wild type and N466328 backgrounds , but almost had no effect on these genes in topp4-1 and amiR-TOPP4 #1-1 ( Figure 6B ) . These results indicated that topp4-1 is insensitive to GA and that the GA signaling pathway is blocked in the mutant . Thus , TOPP4 is likely involved in GA signal transduction . The level of TOPP4 protein after GA3 treatment was also examined by immunoblotting using an anti-TOPP4 antibody . A band with a molecular mass between 37 and 55KD was detected , and it was weak in N466328 ( Figure 6C ) , demonstrating that it is the corresponding band of TOPP4 and the anti-TOPP4 antibody is specific . After exogenous GA3 treatment , TOPP4 protein was increased in wild-type plants but was almost unaltered in topp4-1 ( Figures 6C and S7 ) . This result suggested that GA may promote TOPP4 protein accumulation in wild-type plants , and this effect is apparently disturbed in topp4-1 . This might be one of the reasons that topp4-1 is insensitive to GA . However , the GA3-induced TOPP4 protein accumulation was not caused by the increased transcription level of TOPP4 ( Figure 6D ) . We also detected the TOPP4 protein level in the gid1a/b/c mutant treated with or without GA3 . Compared to wild type , TOPP4 protein was significantly decreased in the gid1a/b/c mutant ( Figure 6E ) , probably attributed to that GA could not be perceived in it . After GA3 treatment , the TOPP4 protein level was not obviously changed in the gid1a/b/c mutant ( Figure 6E ) . These results indicated that GA enhances the TOPP4 protein level through a GA-GID1 pathway . RGA and GAI are the main repressors of GA signaling and their accumulation causes severe dwarf phenotypes in plants [44] . Given the fact that topp4-1 also showed severe dwarfism , genetic interactions between TOPP4 and RGA or GAI were studied . We screened rga-t2 topp4-1 , gai-t6 topp4-1 , and rga-t2 gai-t6 topp4-1 from offsprings of a cross between topp4-1 and the DELLA penta mutant ( gai-t6 rga-t2 rgl1-1 rgl3-1 SGT625-5-2 , Ler background ) . Then , the rga-t2 topp4-1 , gai-t6 topp4-1 , and rga-t2 gai-t6 topp4-1 double and triple mutants were back-crossed with Col six times for subsequent analyses . At the same time , the rga-t2 and gai-t6 single mutants and the rga-t2 gai-t6 double mutant were screened from the same genetic cross as control ( Figure 7A ) . Phenotypic analyses revealed that the loss-of-function mutants rga-t2 and gai-t6 could partially reverse the defective phenotype of topp4-1 ( Figure 7A ) . The topp4-1 mutant had almost no inflorescences , but both rga-t2 topp4-1 and gai-t6 topp4-1 double mutants had 2–3 cm inflorescences , relatively longer than those of topp4-1 ( Figure 7A–B ) . And the triple mutant rga-t2 gai-t6 topp4-1 was taller than the double mutants ( Figure 7A–B ) . These results suggested that RGA and GAI are repressors of TOPP4-mediated stem elongation . TOPP4 therefore may genetically associate with RGA and GAI in the GA signaling pathway . We also analyzed the relationship between TOPP4 and three other DELLA proteins , RGL1 , RGL2 , and RGL3 . The rgl mutations failed to rescue the dwarfism of topp4-1 ( Figure S8A ) . Previous studies provided evidence that RGL1 , RGL2 , and RGL3 may not be required for the repression of stem elongation , but may be mainly involved in seed germination and floral development [9] , [12]–[14] . We next examined the flower development and seed germination in those double mutants . The defective flower morphology of topp4-1 was also observed in rgl1-1 topp4-1 , rgl2-1 topp4-1 , and rgl3-1 topp4-1 ( Figure S8B–C ) . The seed germination of rgl1-1 topp4-1 and rgl3-1 topp4-1 was slightly resistant to paclobutrazol ( PAC , a specific inhibitor blocking the kaurene oxidase reaction in GA biosynthetic pathway ) , similar to that of topp4-1 , whereas rgl2-1 topp4-1 had more resistance to PAC , similar to the single mutant rgl2-1 ( Figure S8D ) [9] . Therefore , it seemed that TOPP4 has no genetic interaction with RGL1 , RGL2 , and RGL3 , regarding stem elongation , flower development , or seed germination . From the genetic results , we hypothesized that RGA and GAI could be over-accumulated in topp4-1 . To test this hypothesis , the levels of GFP-RGA and GAI in wild type , topp4-1 , and TOPP4 overexpressing transgenic plants were assessed by immunoblotting . The topp4-1 plants had more RGA and GAI than wild type , whereas these proteins were significantly lower in TOPP4 overexpressing plants than in wild type ( Figure 8A ) . At the same time , the RGA protein level was also significantly increased in three amiRNA lines ( Figure 8B ) . Thus , TOPP4 may positively regulate the degradation of RGA and GAI and the dwarfed phenotype of topp4-1 may be caused by overaccumulation of these two proteins . The degradation of DELLA proteins required dephosphorylation at Ser or Thr [36] , but the phosphatase responsible for this activity had not been identified . We considered that TOPP4 is likely a candidate for this process . This was tested by comparing the degradation of GFP-RGA in wild-type plants and the topp4-1 mutant . First , a previously reported cell-free system [36] was used in which total proteins were solubilized in a degradation buffer and incubated at 22°C for 0 , 15 , 30 , and 60 min followed by determination of the GFP-RGA abundance by immunoblotting . GFP-RGA protein from wild-type plants was rapidly degraded after 15 min of incubation and little was detected after 60 min ( Figure 8C ) . Degradation rate was clearly slower in the topp4-1 mutant ( Figure 8C ) . Addition of TOPP4 protein immunoprecipitated from wild-type plants to total protein extracts of the topp4-1 mutant increased the rate of GFP-RGA degradation , although not to the level of wild type ( Figure 8C ) . Mutated topp4-1 protein from the topp4-1 mutant did not reverse the delayed degradation ( Figure 8C ) . These results demonstrated that TOPP4 regulates the stability of DELLA protein . Carbobenzoxy-Leu-Leu-leucinal ( MG132 ) , a specific 26S proteasome inhibitor , is reported to block the degradation of DELLA proteins [24] . We used this inhibitor in this cell-free system to determine if the TOPP4-mediated degradation of DELLA protein is dependent on the ubiquitin-proteasome pathway . Supplementation with 100 µM MG132 strongly blocked the degradation of GFP-RGA both in wild type and topp4-1 ( Figure 8C ) , suggesting that the 26S proteasome acts downstream of TOPP4 in the degradation of DELLA protein . GA can rapidly induce DELLA protein degradation [19] , so we tested this effect in the topp4-1 background . Seedlings were incubated in Murashige-Skoog ( MS ) liquid medium supplemented with 100 µM GA3 for 0 , 45 , and 90 min . Total proteins were extracted from these seedlings and assessed by immunoblotting analyses . GFP-RGA accumulation was rapidly reduced after 45 min treatment with GA3 in wild-type seedlings ( Figure 8D ) , but this process was apparently delayed in the topp4-1 seedlings ( Figure 8D ) , confirming that TOPP4 is critical for the GA-induced degradation of DELLA proteins . When seedlings were treated with both GA3 and cycloheximide ( CHX , a chemical blocking protein synthesis ) , the degradation of GFP-RGA was still delayed in topp4-1 ( Figure 8D ) , suggesting that this phenomenon was not affected by de novo protein synthesis . Taken together , we concluded that TOPP4 facilitates the GA-induced degradation of DELLA proteins through a 26S proteasome pathway . The nuclear localization of TOPP4 suggests a role for regulating nuclear proteins such as transcription factors . DELLA proteins also function in the nucleus and therefore they may physically interact to each other . To test this possibility , we performed a protein-protein interaction assay . Because only RGA and GAI showed genetic relevance with TOPP4 , we then only examined interactions of TOPP4 with RGA or GAI proteins both in vitro and in vivo using a number of different biochemical approaches . Recombinant Histidine ( HIS ) -RGA , HIS-GAI , and glutathione S-transferase ( GST ) -TOPP4 were purified from E . coli and an in vitro pull-down experiment was carried out . HIS-RGA or HIS-GAI was pulled down together with GST-TOPP4 using a glutathione sepharose 4B resin ( Figure 9A ) . However , this GST bound GAI protein was gradually reduced when the amount of FLAG-topp4-1 was increased in the same reaction system ( Figure 9B ) , indicating that mutated topp4-1 protein and TOPP4 can competitively interact with DELLA proteins . Next , TOPP4 or topp4-1 was expressed as DNA binding domain ( BD ) protein fusions , and RGA and GAI were expressed as transactivation domain ( AD ) protein fusions in yeast strain Y190 . Interactions of TOPP4-BD and RGA-AD or GAI-AD were confirmed by β-galactosidase ( β-gal ) activity ( Figure 9C ) . Mutated topp4-1 seemed to interact with RGA or GAI slightly more than did the wild-type TOPP4 in yeast two-hybrid assay ( Figure 9C ) . Further , to determine the interaction of TOPP4 and RGA or GAI in planta , we performed co-immunoprecipitation ( co-IP ) and bimolecular fluorescence complementation ( BiFC ) assays . We used 35S-TOPP4-GFP plants as materials and Col as a negative control in co-IP assay . TOPP4 protein was immunoprecipitated with anti-GFP antibody and TOPP4-bound proteins were subjected to immunoblotting analysis . Both RGA and GAI were co-immunoprecipitated with TOPP4 in 35S-TOPP4-GFP , but could not be detected in immunoprecipitated complexes of Col ( Figure 9D ) . Finally , when TOPP4-YFPN and RGA-YFPC or GAI-YFPC were transiently co-expressed in leaves of Nicotiana benthamiana , YFP fluorescence was clearly detected in nuclei , which were confirmed by 4 , 6-diamidino-2-phenylindole ( DAPI ) staining ( Figure 9E ) . These results strongly supported the existence of in vitro and in vivo interactions between TOPP4 and the two DELLA proteins RGA and GAI . We next asked whether phosphorylated RGA and GAI could be dephosphorylated by TOPP4 . Immunoblotting analysis using anti-GFP antibody resulted in a single band of RGA ( Figure 10A ) . After treatment with active calf intestinal phosphatase ( CIP ) , the band had greater electrophoretic mobility , which is representative of the dephosphorylated form ( Figure 10A ) . Similar results were obtained for GAI protein ( Figure 10A ) . This finding was consistent with the GAI phosphorylation status reported by Fu et al . [22] . To further demonstrate that phosphorylated RGA and GAI are the substrates of TOPP4 , we incubated total protein extractions from the topp4-1 mutant with GST-TOPP4 or mutated GST-topp4-1 produced by E . coli . Both GFP-RGA and GAI treated with GST-TOPP4 , but not GST-topp4-1 , showed increased electrophoretic mobility ( Figures 10B and S9A ) . These results are consistent with those obtained by CIP treatment , indicating that TOPP4 , but not topp4-1 , can dephosphorylate phosphorylated RGA and GAI directly . Moreover , we analyzed the post-translational modification of GFP-RGA in wild type and the topp4-1 mutant by two-dimensional gel electrophoresis ( 2-DE ) . Protein gel blots showed several spots with different isoelectric point ( pI ) values in wild-type , topp4-1 , and 35S-TOPP4 transgenic plants . In wild type , the basic forms of GFP-RGA , which represent the dephosphorylated status , were dominant ( Figures 10C and S9B ) , while in topp4-1 , the basic forms were decreased , and the acidic forms , which represent the phosphorylated status , were increased compared to wild type ( Figures 10C and S9B ) . In TOPP4 overexpressing plants , the basic forms of GFP-RGA were increased significantly . The spot at the more acidic side ( indicated by open arrowhead ) was decreased while that at the more basic side ( indicated by solid arrowhead ) was increased compared to wild type ( Figure 10C ) . These results provided further confirmation that TOPP4 can dephosphorylate DELLA proteins . The phosphorylated forms of RGA were increased in topp4-1 , along with its slow degradation and high accumulation in the mutant ( Figure 8 ) , suggesting that the dephosphorylated forms of DELLA proteins may facilitate their destruction . In addition , to investigate whether TOPP4 can dephosphorylate DELLA protein in the absence of GA , we transformed 35S-TOPP4 into gai-1 , a mutant in which GAI cannot be degraded through GA-GID1 for the deletion of its DELLA domain [44] . The result showed that overexpression of TOPP4 could not rescue the dwarfed phenotype of gai-1 ( Figure S10 ) , suggesting that TOPP4-mediated DELLA dephosphorylation is dependent on the formation of the GA-GID1-DELLA complex .
Reversible phosphorylation and dephosphorylation , controlled by protein kinases and protein phosphatases , respectively , is one of the most important mechanisms of the post-translational modifications of proteins . Previous studies revealed the crucial role of dephosphorylation in plant development , mediated by a protein phosphatase 2A ( PP2A ) , a protein phosphatase 2C ( PP2C ) , and a protein phosphatase 6 ( PP6 ) [48]–[51] . However , the regulatory functions of PP1s in plant development are poorly understood . In this study , we report the isolation and phenotypic characterization of a topp4-1 mutant identified from EMS-mutagenized Arabidopsis plants . Our results suggested a positive role of TOPP4 in the GA signaling pathway , through regulating the stability of DELLA proteins . Protein kinases and phosphatases play important roles in several phytohormonal signaling pathways in plants . For example , in BR signaling , brassinosteroid-insensitive 2 ( BIN2 ) phosphorylates and inactivates the transcription factor brassinazole-resistant 1 ( BZR1 ) to inhibit plant growth , whereas PP2A dephosphorylates BZR1-P and promotes or inhibits the expression of its downstream response genes [49] . In ABA signaling , PP2C dephosphorylates SNF1-related protein kinase 2s ( SnRK2s ) to block ABA-mediated stress responses [50] . A key enzyme in ethylene biosynthesis pathway , 1-aminocyclopropane-1-carboxylic acid synthase ( ACS ) , is stabilized by phosphorylation by mitogen-activated protein kinase 6 ( MPK6 ) and destabilized by dephosphorylation by PP2A [52] , [53] . In GA signal transduction , however , the function of protein phosphorylation on the stability of DELLA proteins has remained controversial . Fu et al . [54] revealed that both protein kinases and protein phosphatases were required for the GA-induced degradation of the barley DELLA protein SLENDER ( SLN1 ) . Subsequent studies indicated that protein phosphorylation increased the interaction between DELLA proteins and SCF ubiquitin ligase [21] , [22] , [55] . Conversely , Itoh suggested that phosphorylation of SLENDER RICE1 ( SLR1 ) was independent of its degradation in rice [56] . However , recent work showed that Ser/Thr phosphatase inhibitors suppressed the degradation of RGL2 and RGA in Arabidopsis [36] , [57] . More recently , a rice casein kinase I named as early flowering 1 ( EL1 ) , was identified and shown to stabilize the rice DELLA protein SLR1 by phosphorylation [37] . Based on our genetic and biochemical data and previous studies [36] , [57] , we concluded that the stability of RGA and GAI is regulated by protein phosphorylation and dephosphorylation ( Figure 10D ) . The phosphorylated forms of RGA and GAI are stable and active , inhibiting the GA signaling pathway in Arabidopsis , consistent with the action of SLR1 in rice [37] . TOPP4 dephosphorylates the phosphorylated RGA and GAI , targeting them for the GA-induced degradation by the ubiquitin-proteasome pathway to promote stem elongation . Degradation of RGA and GAI relieves their restraint on GA signaling . In this process , GA promotes TOPP4 protein accumulation through GID1 ( Figure 6C , E ) , thereby enhancing the dephosphorylation and degradation of DELLAs . But in the topp4-1 mutant , GA cannot promote the accumulation of mutated topp4-1 protein and topp4-1 cannot dephosphorylate RGA and GAI ( Figures 6C and 10B ) . The phosphorylated RGA and GAI are degraded slowly in response to GA3 ( Figure 8D ) , and their accumulation blocks GA signal transduction , resulting in GA-related mutant phenotypes , especially severe dwarfism . In TOPP4 overexpressing plants , less RGA and GAI accumulate than in wild-type plants due to excessive dephosphorylation , leading to a taller inflorescence ( Figures 2F and S6 ) . Therefore , our data indicated that TOPP4 is a positive regulator of the GA signaling pathway and functions in stimulating stem elongation by destabilizing DELLA proteins . This dephosphorylation process is likely dependent on the formation of GA-GID1-DELLA module , because overexpression of TOPP4 could not rescue the dwarfed phenotype of gai-1 ( Figure S10 ) . However , unlike some other phosphorylated proteins [BR-signaling kinase 1 ( BSK1 ) ] with many phosphorylation sites in plants [58] , RGA protein may have very few phosphorylation sites , which was supported by the results of 2-DE ( Figures 10C and S9B ) . Our dephosphorylation analyses showed that the difference between phosphorylated and dephosphorylated status of DELLA proteins was very little ( Figure 10A , B ) , and their status is therefore difficult to assess . This might be one of the reasons that the protein phosphorylation-dephosphorylation modification of DELLAs has been controversial . An important finding in this work is that the topp4-1 mutant , with a 246Thr to 246Met amino acid substitution attributed to a G-to-A single nucleotide alteration , displays severe growth defects . The 246Thr is not a conservative site in all TOPPs ( Figure S11 ) . Its mutation does not affect the interaction of topp4-1 with DELLA proteins ( Figure 9C ) , and the mutated topp4-1 and TOPP4 can competitively interact with DELLA proteins ( Figure 9B ) . But the mutation impairs the dephosphorylation function of topp4-1 on DELLAs ( Figure 10B ) . This result can explain that the single mutation of TOPP4 causes a dominant-negative effect on plant growth and development , resulting in severe defects in topp4-1 . The dominant-negative effect is further confirmed by genetic data: expressing either 35S-topp4-1 or pTOPP4-topp4-1 in wild type could mimic the topp4-1 mutant phenotypes ( Figure 2A ) . pTOPP4-TOPP4 only very slightly reversed the defects of topp4-1 and even 35S-TOPP4 could not completely recover it ( Figures S3 and 3A ) . And knocking down topp4-1 gene in the topp4-1 mutant could partially rescue the deficient phenotypes ( Figure 2C ) . Therefore , this dominant-negative material is crucial for elucidating the distinct functions of TOPPs in Arabidopsis . There are nine PP1s ( TOPP1–TOPP9 ) in Arabidopsis [40] , [59] , and they share 90 . 9–99 . 7% amino acid similarities ( Figure S11 ) [60] . It seems that there might be a high degree of functional redundancy among them . In this study , we demonstrated that the amiRNA lines of TOPP4 showed dwarfed phenotypes , with overaccumulated RGA protein ( Figures 3C , 3D , 8B ) . In those amiRNA lines , both TOPP4 gene expression and TOPP4 protein level were dramatically reduced ( Figure 3E–F ) . These results confirmed that TOPP4 is a major protein phosphatase in regulating GA-mediated DELLA protein degradation in Arabidopsis . The PP1 catalytic subunit often binds the regulatory subunit to form a functional enzyme . These regulatory subunits determine the catalytic activity , target the catalytic subunit to specific subcellular compartment , and modulate the specificity of substrates [61] . There are about 100 predicted PP1-binding regulatory subunits in animals [62] . However , to date , only inhibitor-3 ( inh3 ) , Arabidopsis I-2 ( AtI-2 ) , PP1 regulatory subunit2-like protein 1 ( PRSL1 ) have been identified in plants [41] , [42] , [63] . In this study , we found that TOPP4 can directly bind RGA and GAI proteins . However , these two proteins have neither PVxF motif nor SILK motif which are present in PP1 regulatory subunits [64] . It is likely that TOPP4 may require another unknown regulatory subunit for controlling the dephosphorylation of RGA and GAI in vivo . TOPP4 is ubiquitously expressed in various organs throughout different growth stages . Therefore , it may be involved in regulation of many developmental processes . The DELLA deficient mutants rga-t2 and gai-t6 only partially rescued topp4-1 phenotypes , suggesting that TOPP4 may promote plant growth also through other signaling pathways . The plasma membrane-localization of TOPP4 also implies that it may participate in many other signal transductions . Identification of the regulatory subunits of TOPP4 in different subcellular locations , tissues , and development stages of plants may provide significant insights into the molecular mechanism of this protein on plant development . To conclude , we have identified a key phosphatase that can directly dephosphorylate DELLA proteins in Arabidopsis; and we elucidated a mechanism of TOPP4 in regulating GA-mediated stem elongation by controlling DELLA protein stability . Future work will focus on identification of a protein kinase involved in phosphorylating Arabidopsis DELLA proteins , the specific phosphorylation sites on DELLA proteins regulating their stability , and the roles of TOPP4 in other developmental processes .
EMS-mutagenized Arabidopsis thaliana ( L . ) Heynh transgenic line E361-1 was screened for mutant topp4-1 . After back-crossing three times with the wild-type Col-0 , topp4-1 plants were used for subsequent research . The T-DNA insertion mutant lines , N466328 and SALK_090980 , were obtained from European Arabidopsis Stock Centre ( NASC ) and Arabidopsis Biological Resource Center ( ABRC ) , respectively . Primers used for identifying homozygous lines are indicated in Table S2 on line . The single , double and triple mutants rga-t2 , gai-t6 , rga-t2 gai-t6 , rga-t2 topp4-1 , gai-t6 topp4-1 , rga-t2 gai-t6 topp4-1 , rgl1-1 topp4-1 , rgl2-1 topp4-1 , and rgl3-1 topp4-1 were generated from the cross of topp4-1 with DELLA penta mutant ( gai-t6 rga-t2 rgl1-1 rgl3-1 SGT625-5-2 , cs16298 , from ABRC ) . Primers used for genotyping are indicated in Table S2 on line . pRGA-GFP-RGA line ( cs6942 ) , ga1-3 ( cs3104 ) , ga4 ( cs62 ) , gai-1 ( cs63 ) , and gid1a-2/gid1b-3/gid1c-1 ( gid1a/b/c , cs16297 ) were ordered from ABRC . pRGA-GFP-RGA topp4-1 was generated from the cross of pRGA-GFP-RGA with topp4-1 . The topp4-1 plants from the F2 population of a cross between the topp4-1 mutant in Col-0 ecotype background and Ler-0 ( cs20 , from ABRC ) were selected for mapping . Simple sequence length polymorphism ( SSLP ) markers were used to define the mutant gene to chromosome 2 [65]–[67] . The markers used in fine mapping are listed in the Table S1 on line , including In/Del and CAPS . All of the eight new markers were developed by our lab . We sequenced the 90-Kb between markers T5I7-29008 and T28M21-47168 to identify the TOPP4 gene finally . For the complementation experiment and overexpressing transgenic line 35S-TOPP4 , a 1696-bp genomic sequence consisting of the entire coding region was PCR-amplified by PCR from the genome of wild-type Col-0 with primer set 5′-GGGGTACCTCTTTGCGCGTAATTTTCT-3′ and 5′-CGAGCTCCTCAAGAAAGACCAAATCCA-3′ . Underlined regions introduce Kpn I and Sac I sites , respectively . The amplified fragment was cloned into pCAMBIA 1300 . Transgenic lines expressing GFP-tagged TOPP4 ( 35S-TOPP4-GFP ) , 35S-topp4-1 , and 35S-TOPP4-RFP were generated by amplifying the cDNA of Col-0 or topp4-1 with a primer set 5′-TCTAGAATGGCGACGACGACGAC-3′ and 5′-GGTACCTCCTCCTCCAATCTTTGTGGACATCATGA -3′ . Underlined regions introduce Xba I and Kpn I sites , respectively . The amplified fragment was cloned into pCAMBIA 1300-GFP or pCAMBIA 1300-RFP . For pTOPP4-TOPP4 , the TOPP4 promoter about 2-Kb upstream of ATG was generated with a primer set 5′-CAAGCTTTTCCGACTTAATCCGGTCCA-3′ and 5′-CTCTAGACCTAATTTTTTCGACCCC-3′ . Underlined regions introduce Hind III and Xba I sites , respectively . The 35S promoter of 35S-TOPP4 was replaced by this amplified fragment . For pTOPP4-topp4-1 , the 35S promoter of 35S-topp4-1 was replaced by the TOPP4 promoter . For promoter analysis ( pTOPP4-GUS ) , the promoter was generated with a primer set 5′-CAAGCTTTTCCGACTTAATCCGGTCCA-3′ and 5′-CGGGATCCCCTAATTTTTTCGACCCC-3′ . Underlined regions introduce Hind III and BamH I sites , respectively . The amplified fragment was cloned into pCAMBIA 1300-GUS that we reconstructed . For transient expression in Arabidopsis protoplasts , TOPP4-YFP were generated by amplifying the cDNA of TOPP4 with a primer set 5′-GTCGACATGGCGACGACGACGAC -3′ and 5′-GGATCCAATCTTTGTGGACATCATGA -3′ . Underlined regions introduce Sal I and BamH I sites , respectively . The amplified fragment was cloned into PA7-YFP . Primers for amiRNA-TOPP4 were designed by Web MicroRNA Designer 3 ( http://wmd3 . weigelworld . org/cgi-bin/webapp . cgi ) oligo design algorithm using the RS300 vector sequence and the amiRNA sequences of TOPP4 gene ( amiR-TOPP4-1: 5′-TACCTAATTTTTTCGACGCCA-3′; amiR-TOPP4-2: 5′-TAAAATTACGCGCAAAGACTA-3′ ) . Detailed information using overlapping PCR and primer sets is available at Web MicroRNA Designer 3 web site . The full amiRNA fold-back fragment was subsequently cloned into pCAMBIA 1300 by Xba I/Kpn I . The alignment of the nucleotide sequence for targets of amiR-TOPP4 with the same regions of other TOPPs in Arabidopsis was presented in Figure S12 . Plant transformation plasmids were introduced into Agrobacterium tumefaciens strain GV3101 , and then were transformed into Arabidopsis plants using the flower-dipping method [46] . T1 transgenic lines were selected on MS [68] plates with 25 mg/L hygromycin ( Solarbio , Beijing , China ) . Genetic and phenotypic analyses were performed mainly in the T2 generation . The T2 transgenic plants carrying the pTOPP4-GUS construct were immersed in GUS staining solution [50 mM Na-Phosphate buffer , pH 7 . 0 , 1 mM EDTA , 0 . 1% Triton X-100 , 100 µg/mL chloramphenicol , 1 mg/mL 5-bromo-4-chloro-3-indolyl-β-D-glucuronic acid ( X-gluc ) , 2 mM ferricyanide , 2 mM ferrocyanide] and incubated overnight at 37°C . Then , they were decolorized in chloral hydrate solution [8 g of chloral hydrate , 1 mL of glycerol , and 2 mL of water] . The stained tissues were observed and photographed by light microscopy ( 80i , Nikon ) . Subcellular localization of TOPP4-GFP was photographed by confocal microscopy ( Olympus FluoView FV1000MPE ) . For plasmolysis studies , roots of 10-day-old 35S-TOPP4-GFP transgenic plants were analyzed as described previously [69] and observed by confocal microscopy . All qRT-PCR measurements were performed using a MX 3000 Real-time PCR system ( Stratagene , La Jolla , CA ) with SYBR Premix Ex Taq ( Takara Bio , Inc . , Shiga , Japan ) . Total RNA ( E . Z . N . A Plant RNA Kit , OMEGA Bio , tek , Norcross , GA ) was extracted from 0 . 05 g of tissue from 2-week-old seedlings grown on MS medium or MS medium containing 10 µM GA3 ( Sigma , St . Louis , MO ) . The cDNAs were synthesized from 1 µg of total RNA using the PrimeScript RT reagent Kit ( Perfect Real Time ) ( Takara Bio , Inc . , Shiga , Japan ) . We used the housekeeping gene GLYCERALDEHYDE-3-PHOSPHATE DEHYDROGENASE C SUBUNIT ( GAPC ) as a normalization control . All experiments were performed with three replicates . The primers used for qRT-PCR are listed in Table S2 on line . For transient expression in Arabidopsis protoplasts , mesophyll protoplast separation and PEG4000-mediated transfection were performed as previously described [70] . TOPP4-YFP was used in this process . YFP signals were observed with a confocal microscopy . For transient expression in Nicotiana benthamiana leaves , the Agrobacterium strain containing 35S-TOPP4-GFP construct was infiltrated into leaves of 4-week-old tobacco plants , and GFP and RFP were observed 2 days after transformation by a confocal microscopy . For inflorescence analyses , 7-day-old plants grown in soil were sprayed with 100 µM GA3 or water for control every 3 days for 3 weeks . For seed germination , seeds were grown on the MS plates contained 0 , 5 , and 10 µM PAC . Seed germination was scored 6 days after vernalization . Thirty to fifty seedlings were measured each time . These experiments were repeated three times independently . The primary antibodies used in this study were anti-RGA ( Agrisera , Vännäs , Sweden ) , anti-GFP ( Invitrogen , Carlsbad , CA ) , anti-GST ( ZSGB-Bio , Beijing , China ) , anti-HIS ( ZSGB-Bio , Beijing , China ) , anti-TOPP4 , and anti-GAI . Anti-TOPP4 and anti-GAI were made in our lab . Anti-GAI was generated in rabbits by immunizations with full-length protein sequence . The anti-TOPP4 antibodies were generated in rabbits using 150 amino acids at the N terminal of TOPP4 . Both anti-GAI and anti-TOPP4 antibodies were affinity purified , and the specificity of them were determined using mutants gai-t6 and N466328 , respectively . Immunoblotting analysis was performed as previously described [71] . For immunoblotting detection of TOPP4 in plasma membrane , plasma membrane extraction was isolated from 2-week-old wild-type seedlings and anti-GFP and anti-PIN1 antibodies ( N782245 , NASC ) were used for detecting TOPP4 and PIN1 protein , respectively . For DELLA protein level assay , 20-day-old wild-type , topp4-1 , 35S-TOPP4 , and three amiRNA transgenic plants were used . For DELLA protein degradation assay , total protein extracts from 20-day-old pRGA-GFP-RGA and pRGA-GFP-RGA topp4-1 plants were prepared as previously described [36] and treated with TOPP4 immunoprecipitated from wild-type plants using anti-TOPP4 antibody , topp4-1 from the topp4-1 mutant , or 100 µM MG132 ( Calbiochem , Darmstadt , Germany ) . For GA3 treatment , 20-day-old pRGA-GFP-RGA and pRGA-GFP-RGA topp4-1 plants were incubated in MS liquid medium containing 100 µM GA3 or 100 µM GA3 together with 50 µM CHX for the indicated time periods . For TOPP4 protein level analysis , 2-week-old wild-type , topp4-1 , gid1a/b/c or 35S-TOPP4-GFP plants were treated with 10 or 100 µM GA3 for 4 h . The coomassie brilliant blue-stained rubisco small subunit ( RbcS ) protein was used as a loading control as indicated . Each experiment was repeated at least three times . Relative band intensities were then calculated for each immunoblot panel by Emage J . The yeast strain Y190 was used in our experiments . Yeast transformations were performed according to the MATCHMAKER two-hybrid system 3 ( Clontech , Shiga , Japan ) . Full-length cDNA of TOPP4 or topp4-1 fused to the DNA-binding domain of GAL4 was used as the bait protein and GAI or RGA fused to the transcriptional activation domain of GAL4 was used as the prey protein . Yeast clones containing the GAL4-BD-TOPP4/GAL4-BD-topp4-1 and GAL4-AD-GAI or GAL4-AD-RGA constructs were plated on synthetic dextrose ( SD ) -His-Trp-Leu medium for 5 days at 30°C to assay for interaction . The results of β-gal filter assay were observed in one hour . β-gal activity were detected according to the manufactures protocol ( Clontech , Shiga , Japan ) . This experiment was repeated at least three times . The GST-TOPP4 , HIS-GAI , HIS-RGA proteins were expressed in E . coli BL21 . The recombinant proteins were co-incubated in the presence of glutathione sepharose 4B resin ( GE , Fairfield , CT ) , which was used to selectively bind the GST fusion proteins with PBS buffer [140 mM NaCl , 2 . 7 mM KCl , 10 mM Na2HPO4 , 1 . 8 mM KH2PO4] . The bound proteins were eluted with 1× sodium dodecyl sulfate ( SDS ) loading buffer and analyzed with anti-GST and anti-HIS antibodies . For competitive pull-down assay , FLAG-topp4-1 was expressed in E . coli BL21 . Pull-down reactions were performed in the presence of 5 µg GST-TOPP4 , 2 . 5 µg HIS-GAI , and 1 . 25 µg or 2 . 5 µg FLAG-topp4-1 . HIS-GAI and FLAG-topp4-1 were mixed , pulled-down with GST-TOPP4 and detected by anti-HIS antibody [72] . These experiments were repeated at least three times . Co-immunoprecipitation studies of TOPP4 and RGA or GAI were performed on 10-day-old seedlings of Col and 35S-TOPP4-GFP . RGA and GAI proteins from these two materials were adjusted to the same amount . Immunoprecipitation of TOPP4 protein used anti-GFP antibody . Protein G agarose ( GE , Fairfield , CT ) was used to precipitate the immunoprotein complexes with IP buffer [150 mM NaCl , 50 mM Tris-HCl , pH 7 . 5 , 1% NP-40 , 1% protease inhibitors] . After immunoprecipitation , beads were washed four times with IP buffer . Proteins were then released and collected by boiling in 2× SDS loading buffer for 5 min . Immunoprecipitation products were detected by immunoblotting with RGA- or GAI-antibody , respectively . This experiment was repeated at least three times . The full-length open reading frame sequences for the Arabidopsis TOPP4 , RGA , and GAI were amplified and cloned into pEearleygate201-YN and pEearleygate202-YC BiFC vectors to generate TOPP4-YFPN , RGA-YFPC , and GAI-YFPC [73] , [74] . Co-infiltration of Agrobacterium strains containing the BiFC constructs and the p19 silencing plasmid was carried out at OD 600 of 0 . 7∶0 . 7∶1 . 0 and infiltrated into leaves of 4-week-old Nicotiana benthamiana plants [24] . The BiFC signal was observed 3 days after infiltration using a fluorescence microscope . Leaves were incubated with 0 . 2 mg/L DAPI for nuclei staining . For CIP treatment , 70 µg of total protein extracts from 20-day-old topp4-1 seedlings ( for GAI assay ) and pRGA-GFP-RGA topp4-1 seedlings ( for RGA assay ) were added with 50 U CIP ( NEB , Ipswich , MA ) or the same amount of denatured CIP and incubated at 37°C for 3 h and then detected by immunoblotting . For TOPP4 treatment , 70 µg of total protein extracts from 20-day-old topp4-1 seedlings ( for GAI assay ) and pRGA-GFP-RGA topp4-1 seedlings ( for RGA assay ) with buffer [50 mM Tris-HCl ( pH 7 . 0 ) , 0 . 1 mM Na2EDTA , 5 mM DTT , 0 . 01% ( w/v ) Brij 35 , 1 mM MnCl2 , 1 µM protease inhibitors] were added with 5 µg GST-TOPP4 or GST-topp4-1 and incubated at 30°C for 1 h , after which they were subjected to immunoblotting . An equal amount of extracts was used as the control for each treatment . The reaction was terminated by adding loading buffer . Each experiment was repeated at least seven times . Total proteins were extracted from 20-day-old pRGA-GFP-RGA , pRGA-GFP-RGA topp4-1 , and pRGA-GFP-RGA 35S-TOPP4 plants and separated by 2-DE using a 13-cm , pH 4–7 immobilized pH gradient gel ( IPG ) strip in Ettan IPGphor 3 Isoelectric Focusing System ( GE , Fairfield , CT ) , and the second-dimensional separation was performed on 8% SDS PAGE gel . The same amount of RGA protein was loaded for each sample . The proteins were detected using anti-GFP antibody . This experiment was repeated three times . Sequence data from this article can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: TOPP4 ( At2g39840 ) , RGA ( At2g01570 ) , GAI ( At1g14920 ) , RGL1 ( At1g66350 ) , RGL2 ( At3g03450 ) , RGL3 ( At5g17490 ) , TOPP1 ( At2g29400 ) , TOPP2 ( At5g59160 ) , TOPP3 ( At1g64040 ) , TOPP5 ( At3g46820 ) , TOPP6 ( At5g43380 ) , TOPP7 ( At4g11240 ) , TOPP8 ( At5g27840 ) , TOPP9 ( At3g05580 ) , TOPP1 ( Vicia faba , AB038648 ) , PP1 ( Oryza sativa , OSU31773 ) , ser/thr PP1 ( Zea mays , ZEAMMB73_175230 ) , PPP1CC ( Homo sapiens , HGNC:9283 ) , PPP1CC ( Mus musculus , MGI:104872 ) , PP2A ( At1g69960 ) .
|
Gibberellins ( GAs ) are essential regulators of plant growth and development . They are tightly related to crop productivity in the first “green revolution . ” GA triggers its responses by targeting DELLA proteins , the important repressors , for degradation . This process is believed to be regulated by protein phosphorylation and dephosphorylation , but there are not any reports describing the identification of phosphatases regulating this critical event . By screening an ethyl methane sulfonate ( EMS ) -mutagenized Arabidopsis thaliana population , we identified a protein phosphatase TOPP4 , a member of protein phosphatase 1 ( PP1 ) , that acts as a positive regulator in the GA signaling pathway . TOPP4 promotes the GA-induced degradation of DELLA proteins by directly dephosphorylating these proteins . This study provides an important insight for the switch role of protein phosphorylation and dephosphorylation in GA signal transduction and sheds light on PP1 protein phosphatases in regulating plant growth and development .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"developmental",
"biology",
"plant",
"science",
"plant",
"hormones",
"plant",
"growth",
"and",
"development",
"biology",
"and",
"life",
"sciences",
"hormones"
] |
2014
|
Arabidopsis DELLA Protein Degradation Is Controlled by a Type-One Protein Phosphatase, TOPP4
|
Four sesquiterpene lactones , mikanolide , deoxymikanolide , dihydromikanolide and scandenolide , were isolated by a bioassay-guided fractionation of Mikania variifolia and Mikania micrantha dichloromethane extracts . Mikanolide and deoxymikanolide were the major compounds in both extracts ( 2 . 2% and 0 . 4% for Mikania variifolia and 21 . 0% and 6 . 4% for Mikania micrantha respectively , calculated on extract dry weight ) . Mikanolide , deoxymikanolide and dihydromikanolide were active against Trypanosoma cruzi epimastigotes ( 50% inhibitory concentrations of 0 . 7 , 0 . 08 and 2 . 5 μg/mL , for each compound respectively ) . These sesquiterpene lactones were also active against the bloodstream trypomastigotes ( 50% inhibitory concentrations for each compound were 2 . 1 , 1 . 5 and 0 . 3 μg/mL , respectively ) and against amastigotes ( 50% inhibitory concentrations for each compound were 4 . 5 , 6 . 3 and 8 . 5 μg/mL , respectively ) . By contrast , scandenolide was not active on Trypanosoma cruzi . Besides , mikanolide and deoxymikanolide were also active on Leishmania braziliensis promastigotes ( 50% inhibitory concentrations of 5 . 1 and 11 . 5 μg/mL , respectively ) . The four sesquiterpene lactones were tested for their cytotoxicity on THP 1 cells . Deoxymikanolide presented the highest selectivity index for trypomastigotes ( SI = 54 ) and amastigotes ( SI = 12 . 5 ) . In an in vivo model of Trypanosoma cruzi infection , deoxymikanolide was able to decrease the parasitemia and the weight loss associated to the acute phase of the parasite infection . More importantly , while 100% of control mice died by day 22 after receiving a lethal T . cruzi infection , 70% of deoxymikanolide-treated mice survived . We also observed that this compound increased TNF-α and IL-12 production by macrophages , which could contribute to control T . cruzi infection .
Chagas’ disease is caused by the protozoan parasite Trypanosoma cruzi , and even though over 100 years have passed since its discovery , this disease still represents a serious problem for global health [1] . Thus , this disease , which was traditionally limited to Latin America , has crossed these boundaries and spread worldwide . Increased travelling and migration of individuals from endemic to non-endemic areas has presented a worrisome scenario for congenital infection , blood transfusions and organ transplantations [2 , 3] . Trypanosoma cruzi infects a wide range of mammalian hosts , including humans . About six to seven million people are known to be affected by this infection [4] . The alarming numbers would be even greater if the underdiagnosed cases were taken in account . Vector control programs have not been completely successful in preventing parasite transmission [5] . Leishmaniasis is another protozoan parasitic disease caused by Leishmania spp . , which presents a worldwide distribution . According to World Health Organization , 300000 cases of visceral leishmaniasis are reported annually , with 200000 deaths One million cases of cutaneous leishmaniasis have been reported in the last five years and 310 million people are at risk of acquiring this parasitosis [6] . Cutaneous leishmaniasis is the most common clinical form in Argentina caused by L . ( V . ) braziliensis and L . ( L . ) amazonensis , with some recent reports about human and canine leishmaniasis [7] . Available anti-parasitic drugs for both Chagas’ disease and leishmaniasis are not sufficiently safe or effective [8 , 9] and no protective vaccines have been developed so far [10] . Natural products are an interesting source of new drugs that might , in the near future , replace current medications , which are known to have severe side effects . Sesquiterpene lactones are a group of natural compounds , mainly found in the Asteraceae family , which show interesting biological activities such as antitumoral , antiinflammatory , antibacterial and antiparasitic , among others [11] . Some of the most representative compounds with antiparasitic activity are artemisinin , parthenolide , costunolide , helenalin , mexicanin , psilostachyins and cynaropicrin [12–14] . The Mikania genus ( Asteraceae ) is found in the tropics of America and Asia and many of its species which are known with the common name of “guaco” are used for treating fever , rheumatism , colds and respiratory diseases , as well as for snake bites and scorpion stings . In recent years , there has been an increasing interest in the study of species from the genus Mikania , since many of its species have been reported to have a wide range of bioactivities [15] . Mikania variifolia Hieron . and M . micrantha Kunth are species native to South America . These species grow in the Northeastern region of Argentina , South of Brazil , Paraguay and Uruguay . In particular , M . micrantha is considered a very invasive weed , growing in cattle fields [16] and has spread throughout Asia . It is used in popular medicine as vulnerary and as antidote [17] . In previous investigations , the isolation of sesquiterpene lactones , flavonoids and caffeoylquinic acids esters have been reported [18–21] . Antimicrobial and antiviral activities have also been described for this species [22] . We have previously reported the antiprotozoal and antiviral activities of extracts from M . micrantha . The organic extract of this species has proved to be active against T . cruzi epimastigotes and Leishmania braziliensis promastigotes . The bioassay-guided fractionation of this extract led to the identification of two fractions with trypanocidal activity which showed the presence of sesquiterpene lactones [23] . In this work we describe the isolation of the active constituents of M . variifolia and M . micrantha and the anti-Trypanosoma cruzi and antileishminial activity of the isolated compounds .
The aerial parts of Mikania variifolia Hieron . ( Asteraceae ) were collected in December 2012 in the province of Entre Rios , Argentina . A voucher specimen ( BAF 788 ) was deposited at the Herbarium of the Museo de Farmacobotánica—Facultad de Farmacia y Bioquímica , Universidad de Buenos Aires . The aerial parts of Mikania micrantha Kunth ( Asteraceae ) were collected in April 2011 in the province of Tucumán , Argentina . A voucher specimen ( LIL609699 ) was deposited at the Herbarium of Instituto Miguel Lillo , Facultad de Ciencias Naturales , Universidad Nacional de Tucumán . Trypanosoma cruzi epimastigotes ( RA strain ) were grown in a biphasic medium . Cultures were routinely maintained by weekly passages at 28°C . T . cruzi bloodstream trypomastigotes ( RA strain ) and the recombinant Tulahuen strain expressing β-galactosidase ( Tul-β-Gal ) were obtained from infected CF1 mice by cardiac puncture at the peak of parasitemia on day 15 post-infection [24] . Trypomastigotes were routinely maintained by infecting 21-day-old CF1 mice . Leishmania braziliensis promastigotes ( MHOM/BR/75/M2903 strain ) were grown in liver infusion tryptose medium LIT , which was prepared as follows: 5 g/L liver infusion ( Sigma 2023-072K1066 ) , 5 g/L tryptose ( Britania ) , 2 g/L glucose ( Sigma ) , 68 mM NaCl , 5 . 4 mM KCl , 22 mM HPO4Na2 , supplemented with 20 mg/L hemin ( Sigma ) and 10% ( vol/vol ) fetal calf serum ( FCS ) ( Internegocios ) . Culture maintenance was performed by weekly passages at 26°C . The extraction of the aerial parts of M . variifolia ( 400 g ) and M . micrantha ( 100 g ) was done by maceration with dichloromethane ( 8L and 1L , respectively ) at room temperature . The organic extracts were filtered and taken to dryness . The extract residue of M . variifolia was suspended in ethanol:water ( 70:30 ) ( 50 mL ) and partitioned successively with hexane ( 3x40 mL ) and dichloromethane ( 3x40 mL ) . The dichloromethane fraction ( 3 . 5 g ) was subjected to open-column chromatography over silica gel 60 with a gradient of dichloromethane/ethyl acetate/methanol , collecting fractions of 70 mL each . Fractions were then combined according to their TLC profile on Silica gel 60 F254 using hexane:ethyl acetate ( 1:1 ) as mobile phase , into six final fractions ( MV1–MV6 ) . These fractions were subsequently tested for trypanocidal activity against T . cruzi epimastigotes . Compound 1 precipitated from fraction MV3 ( eluted with dichloromethane:ethyl acetate ( 3:1 ) . Compound 2 precipitated from MV4 eluted with dichloromethane:ethyl acetate ( 2:1 ) . The supernatant from this fraction was subjected to chromatography on a Silica gel column eluted with hexane:ethyl acetate ( 1:1 ) and compound 3 precipitated from one of the fractions . The fractionation procedure of M . micrantha active extract was described previously [23] . The active fractions MM3 and MM4 were combined ( 1 . 25 g ) and subjected to Silica gel column ( 51 cm x 3 . 5 cm , 110 g ) chromatography eluted with a gradient of hexane , dichloromethane , ethyl acetate and methanol . Fractions of 20 mL were collected . Compound 1 and compound 2 precipitated from fractions eluted with dichloromethane:ethyl acetate ( 95:5 ) and dichloromethane:ethyl acetate ( 90:10 ) , respectively . Compound 3 crystallized from the supernatant of fractions eluted with dichloromethane:ethyl acetate ( 90:10 ) . The resulting mother liquors were subjected to preparative TLC on Silica gel 60 F254 using hexane:ethyl acetate ( 1:1 ) as mobile phase to obtain compound 4 . The isolated compounds were identified by proton and carbon nuclear magnetic resonance ( 1HNMR ) and 13CNMR , ( Bruker Avance 500 ) ( 500 MHz , in CDCl3 ) , heteronuclear single quantum correlation ( HSQC ) , heteronuclear multiple bond correlation ( HMBC ) , correlated spectroscopy ( COSY ) , electron impact mass spectrometry ( EI-MS ) ( Thermo Scientific EM/DSQII ) , infrared spectroscopy ( IRS ) ( Bruker FT-IR IFS 25 ) and by comparison with literature data and reference compounds . The purity of isolated compounds was checked by HPLC/DAD using a Varian Pro Star instrument equipped with a reversed-phase column Phenomenex–KinetexXB-C18 ( 250 mm x 4 . 6 mm and 5 μdp ) , a Rheodyne valve ( 20 μL ) and a photo diode array detector set at 210 nm . The chromatographic system used was that reported by Laurella et al . [23] . For HPLC analysis , samples were eluted with a gradient of water ( A ) and acetonitrile ( B ) from 100% A to 0% A in 60 min and back to initial conditions . A flow rate of 1 mL/min was employed and the separation was done at room temperature ( 18–25°C ) . Data were analysed with Varian Star 5 . 5 ( USA ) . Mikania micranta and M . variifolia organic extracts and pure isolated compounds ( used as reference ) were dissolved in methanol and water ( 9:1 ) at a concentration of 5 mg/mL and 1 mg/mL respectively . Water employed to prepare working solution was of ultrapure quality ( Milliq ) . Methanol and acetonitrile ( HPLC ) J . T . Baker were used . The evaluation of T . cruzi epimastigotes growth inhibition was performed by the [3H] thymidine uptake assay [25] . Briefly , exponentially growing epimastigotes were adjusted to a cell density of 1 . 5 × 106 parasites/mL in fresh medium . Parasites were allowed to grow for 72 h at 28°C in medium with or without different concentrations of each compound ( 0–5 μg/mL ) in triplicate . Percent of inhibition was calculated as {100 − [ ( cpm of treated parasites/ cpm of untreated parasites ) × 100]} . The compound concentration at which the parasite growth was inhibited by 50% ( IC50 ) was determined after 72 h . Benznidazole was used as positive control . The trypanocidal effect of the purified compounds was also tested on bloodstream trypomastigotes . Briefly , mouse blood containing trypomastigotes was diluted in complete RPMI -1640 Medium ( Sigma-Aldrich R6504- Batch 029K83102 ) with 10% ( vol/vol ) fetal calf serum ( FCS ) ( Internegocios ) to adjust the parasite concentration to 3 × 105/mL . Parasites were seeded in duplicate in a 96-well microplate in the presence of each compound ( 0–50 μg/mL ) or controls and incubated at 4 °C for 24 h . The number of remaining living parasites in each sample was determined in 5 μL of cell suspension diluted 1/5 in lysis buffer ( 0 . 75% NH4Cl , 0 . 2% Tris , pH 7 . 2 ) and counted in a Neubauer chamber . Benznidazole was used as positive control . The effect of each compound on intracellular forms of T . cruzi was assayed using β-galactosidase transfected parasites [25] . Briefly , 96-well plates were seeded with Raw 264 . 7 cells at 5x103 per well in 100 mL of culture medium and incubated for 24 h at 37 °C in a 5% CO2 atmosphere . Cells were washed and infected with transfected blood trypomastigotes expressing β-galactosidase at a parasite/cell ratio of 10:1 . After 12 h of co- culture , plates were washed twice with PBS to remove unbound parasites and each pure compound was added at different concentrations ( 0 . 001–50 μg/mL ) in 150 μL of fresh complete RPMI medium without phenol red ( Gibco , Rockville , MD ) . Controls included infected untreated cells ( 100% infection control ) and uninfected cells ( 0% infection control ) . The assay was developed 5 days later by the addition of chlorophenol red-β-D-galactopyranoside ( CPRG ) ( 100 mM ) and 1% Nonidet P40 . Plates were then incubated for 4–6 h at 37 °C and the absorbance was measured at 595 nm in a microplate reader ( Bio-Rad Laboratories , Hercules , CA ) . The percentage inhibition was calculated as 100–{[ ( absorbance of treated infected cells ) / ( absorbance of untreated infected cells ) x 100} and the IC50 value was estimated . Benznidazole was used as positive control . The growth inhibition of Leishmania braziliensis promastigotes was evaluated by the MTT method . Parasites ( 5x106 ) were settled at a final volume of 150 μL in a flat-bottom 96-well microplate and cultured at 37°C in a 5% CO2 atmosphere in the absence or presence of increasing concentrations of the pure compounds . After 72 h , 3- ( 4 , 5- dimethylthiazol-2-yl ) -2 , 5-diphenyltetrazolium bromide ( MTT ) was added at a final concentration of 1 . 5 mg/mL . Plates were incubated for 2 h at 37°C . The purple formazan crystals formed were completely dissolved by adding 150 μL of ethanol and the absorbance was read at 595 nm in a microplate reader . The inhibition percentage was calculated as {100 − [ ( DO595nm of treated parasites/ DO595nm of untreated parasites ) × 100]} . Compounds were tested at 0–50 μg/mL . Amphotericin B was used as positive control . The human monocyte leukemia THP1 ( ATCC TIB202 ) cell line were obtained from the vendor . Cells were thawed , expanded and settled at a concentration 5x105 in a final volume of 150 μL in a flat-bottom 96-well microplate and cultured at 37°C in a 5% CO2 atmosphere in the absence or presence of increasing concentrations of the pure compounds . After 24 h , the cell viability was determined by the trypan blue exclusion method in the absence and presence of increasing concentrations of the compounds [26] . The 50% cytotoxic concentration ( CC50 ) and the selectivity index ( SI = CC50/IC50 ) were determined for each compound for T . cruzi trypomastigotes and amastigotes . RAW 264 . 7 ( ATCC TIB71 ) macrophage cell line were obtained from ATCC and cultured ( 1×105 ) in duplicate in 24-well plates in the presence of 2 . 5 or 25 μg/mL of Compound 3 , LPS ( 10 μg/mL ) plus INF-γ ( 0 . 30 μg/mL ) ( positive control ) , or medium alone ( negative control ) . After 24 h , supernatants were collected and stored at -80 °C until cytokine analysis . IL-12 and TNF-α levels were determined by sandwich ELISA ( BD Pharmingen , and R&D System , Minneapolis , MN , respectively ) according to manufacturer’s instructions . Supplied standards were used to generate a standard curve . Inbred male Balb/c mice were nursed at the Departamento de Microbiología , Facultad de Medicina , Universidad de Buenos Aires . Groups of six Balb/c male mice ( 6 to 8 weeks old ) were infected with 1000 bloodstream T . cruzi trypomastigotes by the intraperitoneal route . Five days after infection , the presence of circulating parasites was confirmed by the microhematocrit method ( Sülsen et al . , 2013 ) . Mice were treated daily with compound 3 or benznidazole ( 1 mg/kg of body weight/day ) for five consecutive days ( days 4 to 8 post-infection ) by the intraperitoneal route . Drugs were resuspended in DMSO and then adjust to concentration in 0 . 1 M phosphate buffered saline ( PBS , pH 7 . 2 ) ; this vehicle was also employed as a negative control . Levels of parasitemia were monitored every 2 days in 5 μL of blood diluted 1:5 in lysis buffer ( 0 . 75% NH4Cl , 0 . 2% Tris , pH 7 . 2 ) by counting parasites in a Neubauer chamber . The number of animal deaths was recorded daily . The weight of each animal was evaluated during the acute phase of infection . Results were expressed as the ratio between the weight on each day and the weight registered the day of the infection multiplied by 100 . Animal experiments were approved by the Review Board of Ethics of Universidad de Buenos Aires , Facultad de Medicina ( Argentina ) , with the No 2943/2013 and conducted in accordance with the Guide for the Care and Use of Laboratory Animals of the National Research Council [27] . Results are presented as means ± SEM . GraphPad Prism 5 . 0 software ( GraphPad Software Inc . , San Diego , CA ) was employed to carry out calculations . To calculate the IC50 values , the percentages of inhibition were plotted against the drug concentration and fitted with a straight line determined by a linear regression ( Sigma Plot 10 software ) . Results presented are representative of three to four independent experiments . Parasitemia and weight loss were analyzed using a non parametric test: Mann-Whitney test . The survival curves were analyzed with a log rank test . The statistical significance was determined by one-way analysis of variance ( ANOVA ) performed with the GraphPad Prism 5 . 0 software ( GraphPad Software Inc . , San Diego , CA ) . Comparisons were referred to the control group . P values <0 . 05 were considered significant .
The Mikania variifolia dichloromethane extract was evaluated for its trypanocidal activity on T . cruzi epimastigotes . At 100 μg/mL , the extract caused a 97 . 2% of inhibition of parasite growth . Two active fractions were obtained by bioassay-guided fractionation which caused inhibitions of 99 . 3% and 96 . 1% , respectively at 10 μg/mL . Three compounds ( 1–3 ) were isolated from these fractions . From the active fractions ( MM3 and MM4 ) of the organic extract of Mikania micrantha , four compounds ( 1–4 ) were isolated . Compounds 1–4 were identified by spectroscopic methods and by comparison with literature data [18] . Compound 1 was identified as mikanolide; compound 2 as dihydromikanolide , compound 3 as deoxymikanolide and compound 4 as scandenolide ( Fig 1 ) . Two major peaks ( rt = 21 . 4 and 19 . 9 min ) appeared in the HPLC chromatogram of M . micrantha and M . variifolia dichloromethane extracts , corresponding to mikanolide and deoxymikanolide , respectively . The content of these sesquiterpene lactones is shown in Table 1 . The trypanocidal activity of the isolated compounds was analyzed on T . cruzi epimastigotes . All the compounds were active , with 50% inhibitory concentration ( IC50 ) values of 0 . 08 , 0 . 7 and 2 . 5 μg/mL , for deoxymikanolide , mikanolide and dihydromikanolide , respectively . Conversely , scandenolide was not active against epimastigotes ( IC50 = 137 . 2 μg/mL ) ( Fig 2 ) . For benznidazole , an IC50 of 1 . 7 μg/mL was registered ( S1 Fig ) . The trypanocidal activity against bloodstream trypomastigotes was then analyzed for each drug . As shown in Fig 3 , an important decrease in the remaining live parasite count was observed for dihydromikanolide , deoxymikanolide and mikanolide , in comparison with controls ( IC50 = 0 . 3 , 1 . 5 and 2 . 1 μg/mL , respectively ) . The reference drug ( benznidazole ) showed an IC50 of 10 . 4 μg/mL ( S1 Fig ) . Finally , the inhibition of amastigotes replication was tested on transgenic T . cruzi expressing the β-galactosidase . Mikanolide , deoxymikanolide and dihydromikanolide were active against this replicative form of the parasite with IC50 values of 4 . 5 , 6 . 3 and 8 . 5 μg/mL , respectively ( Fig 4 ) . The IC50 value for benznidazole was 1 . 1μg/mL . Scandenolide showed IC50 values of 68 . 4 and 60 . 6 μg/mL on trypomastigotes and amastigotes , respectively , therefore it was not active against this replicative form . A moderate activity against L . braziliensis promastigotes was observed with mikanolide , deoxymikanolide and dihydromikanolide ( IC50 values of 5 . 1 , 11 . 5 and 57 . 1 μg/mL , respectively ) . By contrast , no activity was registered for scandenolide ( IC50 = 252 . 0 μg/mL ) ( Fig 5 ) . The reference drug , amphotericin B , presented an IC50 value of 0 . 26 μg/mL ( S2 Fig ) . The in vitro cytotoxic effect of each compound was evaluated on a human cell line by the trypan blue exclusion method . Viable human cells were incubated in the absence and presence of increasing concentrations of the compounds . Deoxymikanolide presented a CC50 value of 79 . 37 μg/mL , being the SI for T . cruzi trypomastigotes and amastigotes 54 and 12 . 5 , respectively . These SI values were higher than those obtained for mammalian cells . CC50 values for dihydromikanolide and mikanolide were 12 . 98 and 22 . 3 μg/mL , respectively . Selectivity indexes for these STLs were 50 and 1 . 52 ( dihydromikanolide ) , and 10 . 66 and 4 . 96 ( mikanolide ) , for trypomastigotes and amastigotes , respectively . Selectivity indexes of the compounds for L . braziliensis were lower than 10 ( Table 2 ) . Since deoxymikanolide presented good selectivity on differents T . cruzi stages , this compound was selected for an in vivo study . Balb/c mice were infected with a lethal dose of T . cruzi ( RA strain ) and treated for 5 consecutive days with either deoxymikanolide or the vehicle . As shown in Fig 6A , infected mice that received deoxymikanolide presented a lower blood parasitemia , as compared to controls . Moreover , in terms of area under the parasitemia curve , a 1 . 5-fold reduction was observed in deoxymikanolide-treated mice with respect to controls; whereas a 2 . 3-fold was observed for benznidazole-treated mice ( S1 Fig ) . Interestingly , the treatment with deoxymikanolide was able to reduce the weight loss observed during the acute phase of infection , as compared to control mice ( p = 0 . 028 , Mann-Whitney test ) at 20 days post infection ( dpi ) ( Fig 6B ) . More importantly , in deoxymikanolide-treated mice , a significant decrease in the mortality caused by T . cruzi infection was observed ( p = 0 . 02 , Log-rank test ) . While nearly 70% of deoxymikanolide-treated mice survived the acute phase of infection , 100% mortality was observed in control mice by day 22 post-infection ( Fig 6C ) . These results highlight the in vivo efficacy of deoxymikanolide to induce the killing of circulating trypomastigotes during the acute infection . We further studied in vitro the ability of deoxymikanolide to modulate the host’s immune response , particularly , macrophage cells , which are the first line of defense against T . cruzi infection . Upon macrophage stimulation with 25 μg/mL deoxymikanolide , a significant increase in the secretion of TNF-α was observed as compared with the negative control ( p<0 . 01 ) . Moreover , a slight increase in the IL-12 production was observed when deoxymikanolide was used at 2 μg/mL and a higher production rate was observed when this STL was tested at 25 μg/mL ( Fig 7 ) .
Four sesquiterpene lactones of the germacranolide type have been isolated and identified from M . variifolia and M . micrantha organic extracts . Mikanolide , dihydromikanolide and deoxymikanolide were identified in M . variifolia , while the same compounds together with scandenolide were found to be present in M . micrantha . All the compounds , with the exception of scandenolide , showed high activity on T . cruzi epimastigotes . Mikanolide , deoxymikanolide and dihydromikanolide were also active against trypomastigote and amastigote forms . Due to the existence of coinfections with both T . cruzi and Leishmania spp . , the isolated compounds were also tested in vitro against L . braziliensis , the main agent of tegumentary leishmaniasis in Argentina [28] . Mikanolide and deoxymikanolide showed significant activity against L . braziliensis promastigotes , while dihydromikanolide displayed moderate activity . On the other hand , scandenolide was not active against this parasite stage . The major determinant for the antiprotozoal activity of STLs would be the presence of an α , β- unsaturated lactone group in the sesquiterpene lactone structure [29] . Nevertheless , some of α-santonin derivatives , lacking the exomethylene group , have shown trypanocidal activity [30] . Steric and electronic factors could also influence the anti-T . cruzi activity of STLs [31] . The four isolated compounds differ from each other in the number of epoxy groups , in the presence or absence of the exocyclic double bond and in the presence of an OAc group at C-3 . In comparing the results of activity obtained for mikanolide and dihydromikanolide on the different T . cruzi stages , the absence of the exocyclic double bond in dihydromikanolide with respect to mikanolide , does not seem to affect its activity against this parasite . Despite the different number of epoxy groups present in mikanolide and deoxymikanolide structures , both compounds showed similar activity on this parasite . On the other hand , the presence of the exocyclic double bond seems to be important for the leishmanicidal activity , since mikanolide and deoxymikanolide were active , while dihydromikanolide was moderately active . Scandenolide did not show any antiprotozoal activity , suggesting that the presence of an OAc at C-3 could be responsible for the decrease of such activity . A strong correlation between biological activity and cytotoxicity of STLs has been reported [29] . Although many of them display a considerable cytotoxic activity against mammalian cells , some of these compounds show more selectivity against the parasites [31] . Deoxymikanolide presented a CC50 that was approximately four times higher than mikanolide , thus proving to be more selective for the parasite . There are previous reports describing the presence of mikanolide , dihydromikanolide , deoxymikanolide and scandenolide in M . micrantha [18] . Although these sesquiterpene lactones have been evaluated for other activities [20 , 22 , 32] , this is the first study that describes their trypanocidal and leishmanicidal effects . Herein , the STLs of M . variifolia have been isolated for the first time . An optimal response to treatment of trypanosomatid diseases is strongly linked to an efficient immune response . This response is mediated by the activation of macrophages with consequent production of nitrogen and oxygen intermediates that are toxic to the parasite , as well as the activation of T helper and cytotoxic cells producing INF-γ [10 , 33–36] . The biological activities of lactones , including sesquiterpenoids are largely due to their inhibitory effects on the activity of a plethora of enzymes of prokaryotic organisms and eukaryotic cells [37] . Sesquiterpene lactones have proved to be potent inhibitors of different transcription factors [38 , 39] . The best recognized interference of this type of compounds is the inhibition of multiple factors within the nuclear factor-κB ( NF-κB ) signaling system [40 , 41] . These effects are presumed to be related to molecular mechanisms determining the immune activity of sesquiterpene lactones . In that sense , it has been previously reported that a guaianolide type sesquiterpene lactone with a lactone-diol moiety stimulates production of interleukin-8 ( IL-8 ) [42] and up-regulates the lipopolysaccharide ( LPS ) -induced production of pro-inflammatory cytokines such as IL-6 and tumor necrosis factor-α ( TNF-α ) . Harmatha et al . [43] , have demonstrated an increase in NO production and in cytokines such as IFN-γ , IL-1β , IL-6 , VEGF and GM-CSF in a range of low concentrations of 10 nM-10 mM by a guaianolide type sesquiterpene lactone . Based on its activity and selectivity , we selected deoxymikanolide for the evaluation of its immunomodulatory activity on macrophage cells , which are the first line of defense against T . cruzi and Leishmania infection . The results obtained suggest that , besides the antiprotozoal activity that deoxymikanolide has per se , this compound could stimulate the host´s immune system through the secretion of cytokines , thus exerting a protective effect against parasite infection .
Four sesquiterpene lactones , mikanolide , deoxymikanolide , dihydromikanolide and scandenolide were isolated from M . variifolia and M . micrantha . With the exception of scandenolide , all the sesquiterpene lactones showed trypanocidal and leishmanicidal activities . Deoxymikanolide was the most promising compound based on its activity , selectivity and its capacity to stimulate cytokine production . Besides , this compound was also active in a murine model of T . cruzi infection . This finding makes it an interesting lead molecule which may be useful for the development of new drugs , alone or in combination with actual therapy , for the treatment of trypanosomatid diseases , in order to reduce side effects .
|
Chagas' disease and Leishmaniasis are parasitic diseases that affect especially poor people in developing countries . They are caused by the protozoan parasites Trypanosoma cruzi and different Leishmania species , respectively . According to the World Health Organization they are considered , among others , Neglected Tropical Diseases . The drugs currently in use for the treatment of these parasitoses are not at all effective and have severe drawbacks . Nature has proved to be a rich source of bioactive compounds . Among antiparasitic drugs , the sesquiterpene lactone artemisinin and the alkaloid quinine and their derivatives are used nowadays for human malaria treatment . Sequiterpenes lactones , present mainly in species from Asteraceae family , are interesting compounds due their pharmacological properties . This group of compounds has shown trypanocidal and leishmanicidal activity and are considered as promising leads for antiparasitic drug discovery .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion",
"Conclusion"
] |
[
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"cytokines",
"chemical",
"compounds",
"immunology",
"lactones",
"microbiology",
"parasitic",
"diseases",
"protozoan",
"life",
"cycles",
"parasitic",
"protozoans",
"esters",
"developmental",
"biology",
"trypomastigotes",
"protozoans",
"leishmania",
"molecular",
"development",
"epimastigotes",
"life",
"cycles",
"chemistry",
"amastigotes",
"immune",
"system",
"trypanosoma",
"cruzi",
"trypanosoma",
"eukaryota",
"physiology",
"biology",
"and",
"life",
"sciences",
"protozoology",
"physical",
"sciences",
"organisms"
] |
2017
|
Assessment of sesquiterpene lactones isolated from Mikania plants species for their potential efficacy against Trypanosoma cruzi and Leishmania sp.
|
Leukocyte telomere length ( LTL ) is a complex genetic trait . It shortens with age and is associated with a host of aging-related disorders . Recent studies have observed that offspring of older fathers have longer LTLs . We explored the relation between paternal age and offspring's LTLs in 4 different cohorts . Moreover , we examined the potential cause of the paternal age on offspring's LTL by delineating telomere parameters in sperm donors . We measured LTL by Southern blots in Caucasian men and women ( n=3365 ) , aged 18–94 years , from the Offspring of the Framingham Heart Study ( Framingham Offspring ) , the NHLBI Family Heart Study ( NHLBI-Heart ) , the Longitudinal Study of Aging Danish Twins ( Danish Twins ) , and the UK Adult Twin Registry ( UK Twins ) . Using Southern blots , Q-FISH , and flow-FISH , we also measured telomere parameters in sperm from 46 young ( <30 years ) and older ( >50 years ) donors . Paternal age had an independent effect , expressed by a longer LTL in males of the Framingham Offspring and Danish Twins , males and females of the NHLBI-Heart , and females of UK Twins . For every additional year of paternal age , LTL in offspring increased at a magnitude ranging from half to more than twice of the annual attrition in LTL with age . Moreover , sperm telomere length analyses were compatible with the emergence in older men of a subset of sperm with elongated telomeres . Paternal age exerts a considerable effect on the offspring's LTL , a phenomenon which might relate to telomere elongation in sperm from older men . The implications of this effect deserve detailed study .
Leukocyte telomere length ( LTL ) is a complex genetic trait . Though highly variable , it is heritable [1–6] and longer in women than men [1 , 3 , 4 , 6–9] . Environmental factors , including smoking [3 , 10] , obesity [9–11] , psychological stress [12] and low socio-economic status ( SES ) [13] are ostensibly associated with shortened LTL , underscoring the roles of not only genetic factors but also the environment in fashioning leukocyte telomere dynamics ( length and attrition rate ) . Shortened LTL is also observed in individuals with aging-related diseases , including hypertension [1 , 7] , insulin resistance [11 , 14 , 15] , atherosclerosis [16 , 17] , myocardial infarction [16 , 18 , 19] , stroke [9] and dementia [20 , 21] . Further , aging itself may modify the relationship between LTL and some of these variables [reviewed 22] . Although the mechanisms that account for variations among humans in LTL are not fully understood , increased oxidative stress and inflammation – two potential determinants of life span and aging-related diseases [23 , 24] – are likely to heighten age-dependent telomere attrition in leukocytes [9 , 12 , 14 , 25 reviewed in 26] . As aging-related diseases and environmental factors may cause premature mortality , and as men's LTL [1 , 7 , 9 , 14 , 25] and life expectancy [27] are shorter than women's , the potential relation between LTL and human longevity has been explored in several studies , which yielded conflicting results [18 , 28–31] . In light of these diverse observations , we need to further understand the biological determinants of LTL and their role in aging-related diseases . Two studies , comprising 125 [32] and 2 , 433 [33] participants observed a positive correlation between the LTL in adult offspring and paternal age at their birth ( paternal age ) . Here we report a two-phase exploration of the dependency of LTL in the offspring on paternal age . First , we describe the details of this phenomenon in 4 samples of wide age range distributions: the Offspring of the Framingham Heart Study ( Framingham Offspring ) , the National Heart Lung and Blood Institute Family Heart Study ( NHLBI-Heart ) , the Longitudinal Study of Aging Danish Twins ( Danish Twins ) , and the UK adult twin registry ( UK Twins ) . Second , we present data suggesting the emergence of a subset of sperm with longer telomeres in older men . Such a process might explain the dependency of the offspring's LTL on paternal age .
Unryn et al . [32] and more recently De Meyer et al . [33] observed that offspring's LTL was positively associated with paternal age . We confirmed this finding in leukocytes of 4 cohorts of wide age ranges from 3 countries . It is not clear why the offspring's LTL was positively correlated with paternal age in only males of the Framingham Offspring and in the Danish Twins . We suspect that in females a host of factors , including menopause [15 , 22] , might confound associations between LTL and indices related to aging . Thus , much larger numbers of subjects might be needed to elucidate links between women's LTL and aging-related parameters . Indeed , in the present study , the effect of paternal age on the daughter's LTL was clearly observed in the NHLBI-Heart where sample size was more than 2 times and 5 times larger than those of the Framingham Offspring and the Danish Twins , respectively , and in the UK Twins , where sample size was about 10 and 20 times larger than those of the Framingham Offspring and Danish Twins , respectively . When observed , the dependency of offspring's LTL on paternal age was considerable . This was particularly pronounced in the elderly males of the Danish Twins . Given that the elderly represent a selected group [22] , it is possible that the dependency of offspring's LTL on paternal age is more expressed in those individuals who survived or resisted diseases of aging to reach old age . We also note that elderly males of the Danish Twins displayed no age-dependent LTL attrition . However , this was unlikely to be a feature of their old age . Rather , it probably related to the small cohort and its narrow age range ( 74–84 years; Table 5 ) , which contrasted with the wide age ranges of the Framingham Offspring , the NHLBI-Heart and the UK Twins cohorts [22] . What are the potential factors that explain association between offspring's LTL and paternal age ? The key might be age-dependent telomere dynamics in sperm . Mean telomere length in sperm does not decrease , and , in fact , increases with the donor's age [34 , 35] , a phenomenon we confirmed . The donor's age effect on sperm telomere length was not due to the change in the sperm profile of MMP , a parameter serving as an indicator of sperm motility and ‘fertilizing' capacity [36] . Based on the flow-FISH analysis , the increase in sperm telomere length in older men appeared to arise from a subset of their sperm with longer telomeres . However , further studies in larger cohorts over a wide age span of donors are needed to confirm our conclusions and explore their meaning . While a female is born with all the eggs she will have , spermatogenesis is an ongoing process throughout most of the male's life . Consequently , the number of germ-line divisions is much higher in males than females and this sex gap widens with age [reviewed in 37 , 38] . More replications augment the chance for spontaneous germ-line mutations that arise during the mitotic phase of spermatogenesis . Transmitted to their offspring by older fathers , these germ-line mutations might cause very rare diseases such as achondroplasia and craniosynostosis ( Apert's , Crouzon's and Pfeiffer's syndromes ) , as well as other disorders reviewed in [37–40] . However , with advancing age , the numerous replications of the male germ-line also may exert a powerful selection pressure that yields stem cells , which survive the impact of aging . This has been shown in aging male Drosophila melanogaster in which the population of germ-line stem cells is only a subset of the original population at a younger age [41] . Oxidative stress inflicts considerable damage on hematopoietic stem cells [42] , and , in principle , this may apply to other stem cells . As oxidative stress heightens DNA damage and telomere loss [43 , 44] , it may be a key determinant in aging-related diseases [45] . Thus , sperm with longer telomeres might arise from a subset of germ-line stem cells that either sustained less aging-related oxidative stress—perhaps because of increased resistance to its action—or underwent fewer replications prior to meiosis . Elongation of telomere length might also arise from epigenetic processes that take place in stem cells of the germ-line as men get older [46] . The genetic makeup of these sperm ( with or without DNA mutations ) would then be transmitted across generations , endowing some offspring with longer LTL . Telomere length is highly variable among chromosomes [47] . Analyses of single chromosomes in somatic cells of parents and their offspring and in twins indicate that the inheritance of telomere length is allele specific [48–50] . This suggests that telomere length is determined in the zygote , with no apparent role of telomerase to match the lengths of homologous telomeres [reviewed in 51] . As the paternal age effect on LTL has not been examined in newborns , we do not know whether it is mediated by endowing the offspring with longer telomeres at birth or by attenuating the rate of telomere shortening afterwards . In conclusion , offspring's LTL positively correlates with paternal age . A possible explanation for this phenomenon is the presence in older men of a subset of sperm with longer telomeres . The potential impact of paternal age on aging-related diseases and longevity in offspring is of particular relevance because offspring of older fathers comprise an increasing proportion of society [52 , 53] and as the dependency of offspring's LTL on paternal age does not display a paternal age threshold . Finally , the paternal age effect joins a growing list of factors that impact leukocyte telomere biology in humans , underscoring the complex nature of LTL . Deciphering the genetic , epigenetic and environmental factors that account for inter-individual variation in LTL may provide considerable insight into the biology of human aging and aging-related diseases .
To be included in the first phase of the study , each individual ( offspring ) had to meet two criteria: a ) the offspring's LTL was available from ongoing investigations , and b ) the offspring's age ( at blood collection ) and parents' ages at the time of the offspring's birth were available . For the first phase , we studied 235 males and 197 females from the Framingham Offspring , 355 males and 492 females from the NHLBI-Heart , 44 male and 88 female twins from the Danish Twins , and 1 , 954 females from UK Twins . Data on individuals and their partner's occupation were available from UK Twins to ascertain SES using the UK registrar general's classification [54] . No comparable SES information was available for the other cohorts . The Framingham Heart Study includes the original cohort and a second cohort , the Framingham Offspring [55] . LTLs of the Framingham Offspring were from blood collected at exam 6 ( 1995–1998 ) . LTLs from the NHLBI-Heart were from blood collected at the follow-up exams of the original pedigrees from 2002–2003 [56] . The Longitudinal Study of Aging Danish Twins started in 1995 and the LTLs of the Danish Twins were from blood collected during the 1997 survey [57] . LTLs from UK Twins [58] were from blood collected in 2001–2004 . For the second phase , we recruited sperm donors from the staff , student body , and faculty of the UMDNJ , NJMS . These were all healthy Caucasian men . For the Southern blot analysis , we studied 8 young ( 18–19 years ) and 8 older ( 50–59 years ) donors . For the Q-FISH analysis , we measured telomere signal intensity in X and Y sperm from 8 individuals ( age 21–60 years ) , and in sperm sorted based on MMP from 5 young ( 20–30 years ) , and 5 older ( 56–65 years ) donors . For the flow-FISH analysis , we studied sperm from 11 young ( 22–28 years ) and 9 older ( 51–65 years ) donors . Older and young donors provided their sperm on the same day , so that sperm from the two groups were processed in parallel . For the first phase , LTL was determined from Southern blots of the TRFL , as described before [7] . For the second phase , we measured sperm telomere parameters by Southern blots of TRFL ( Figure 2 ) , Q-FISH ( Figure 3 ) , and flow-FISH ( Figure 4 ) . We used chicken red cell nuclei ( CRBCN ) from a single stock as a constant reference for telomere signal for the Q-FISH ( Figure 3 ) and flow-FISH ( Figure 4 ) analyses . Sperm samples were processed within two hours of ejaculation . They were maintained at room temperature for 30 min . Samples were processed with Viscolytic System at 35 °C for 20 min and filtered through SpermPrep II Sephadex columns ( ZDL ) for DNA isolation in preparation for TRFL analysis and Q-FISH analysis of telomeres in sperm subsets based on MMP . Samples were filtered through double layers of Miracloth ( Calbiochem ) , sonicated for 10 sec × 3 on ice , and collected by centrifugation ( 10 min , 600 g ) for flow-FISH analysis and Q-FISH analysis of X and Y sperm . For the Q-FISH of telomere signals in X and Y sperm , the isolated sperm were re-suspended in PBS , treated with Fix & Perm Reagent A and Fix & Perm Reagent B ( Caltag Lab ) plus 10 mmol/L DTT and 0 . 05 μg/μL heparin for 15 min at room temp . Sperm were cytospinned onto slide glasses , air dried and kept at −20 °C until use . Q-FISH: Slides were treated with 1 mg/mL pepsin plus 10 mmol/L DTT and 0 . 05 μg/μL heparin for 10 min at 37 °C and dehydrated by a series of ethanol incubation ( 70 , 90 , and 100% , 1 min each ) and air dried . Slides were hybridized in 70% formamide , 1% blocking reagent ( Roche , Cat#1096176 ) , 10 mmol/L Tris; pH 7 . 2 , 2 ng/μL cy-5 labeled ( CCCTAA ) 3 PNA probe , FITC labeled PNA probe for the Y chromosome ( Dako ) and rhodamine labeled PNA probe for the X chromosome ( a gift from PerSeptive Biosystems ) . Preparations were denatured at 75 °C for 10 min . Slides were kept at room temp for 2 hrs , washed with 70% formamide , 10 mmol/L Tris; pH 7 . 2 for 15 min at room temp 2 times and washed with PBS + 0 . 025 % tween for 5 min at room temp 3 times and counterstained with 200 mg/mL DAPI . Telomere signals were captured and quantified by fluorescence microscope ( Zeiss ) , Metafar ( Metasystem ) from 100–200 X sperm and similar numbers of Y sperm on each slide . For the Q-FISH of telomere signals in unsorted sperm and sperm sorted based on DiOC6 ( 3 ) stain , sperm were suspended in PBS and stained with DiOC6 ( 3 ) at 37 °C for 20 min and 2 . 5 μg/mL propidium iodide ( PI ) added . They were sorted by FACS Vantage SE with FACS DiVa option ( BD Biosciences ) into 3 fractions ( PI positive , PI negative with low DiOC6 ( 3 ) staining , and PI negative with high DiOC6 ( 3 ) staining [36] . Slides were then prepared as above for telomere analysis . A reference consisting of telomere signals from CRBCN was smeared onto the slide glass next to sperm cells and air dried . Telomere signals from sperm and CRBCN were captured and quantified as above . Telomere length in relative units was expressed by the intensity ratio of telomere signal from sperm/ CRBCN . For flow-FISH , isolated sperm were hybridized in 70% formamide , 1% blocking reagent , 10 mmol/L Tris; pH 7 . 2 , 2 ng/μL cy-5 labeled ( CCCTAA ) 3 PNA probe and denatured at 75 °C for 10 min . CRBCN were hybridized in 0 . 2 ng/μL cy-5 labeled ( CCCTAA ) 3 PNA probe plus 1 . 8 ng/μL unlabeled ( CCCTAA ) 3 PNA probe . Sperm were kept at room temp over night and washed with 70% formamide , 10 mmol/L Tris; pH 7 . 2 , 0 . 1% BSA , 0 . 1% tween 20 for 15 min at room temp ( x 4 ) and 0 . 15 mol/L NaCl , 50 mmol/L Tris pH 7 . 5 , 0 . 1% BSA , 0 . 1% tween 20 for 5 min ( x 4 ) . Sperm and CRBCN were re-suspended in PBS for FACS analysis . The acquisition of light scatter and fluorescence signals were done on a FACSCalibur flow cytometer ( BD Biosciences ) and the analysis performed with CellQuest software ( BD Biosciences ) . To correct for daily shifts in the linearity of the flow cytometer and fluctuations in laser intensity and alignment , SPHERO Rainbow Calibration particles ( Spherotech ) were acquired at the beginning of each experiment . A total of 10 , 000 events ( sperm ) were acquired for each specimen analyzed . The fluorescence of gated events was typically analyzed on a linear scale . CRBCN and sperm were distinguished based on their forward and side angle scatter properties . All telomere measurements for both phases of the study were performed without knowledge about the donors of blood or sperm . For the first phase , means and standard deviation ( SD ) for continuous variables or proportions for categorical variables were computed for all study variables . To evaluate the relationships between LTL , age , parent's age at birth of an offspring , and covariates , we performed correlation and linear regression analyses . In the regression analyses we used LTL as a dependent and parents' ages as primary predictor variables with age , BMI , cigarette smoking and SES , if available , as covariates . For the NHLBI-Heart , UK Twins and Danish Twins data , non-independence of family members or twins and the fact that they shared the same father was adjusted for by regression clustering by family number . All analyses were performed for males and females , separately . SAS ( SAS Institute ) was used to analyze the Framingham Offspring ( version 9 . 1 ) and NHLBI – Heart ( version 8 . 2 ) data . Stata software was used for UK Twins ( version 9 ) and Danish Twins ( version 9 . 2 ) data . We note that beta coefficients are presented in the tables and correlation coefficients are presented in Figure 1 . For the second phase , we used Student t test for data displayed in Figures 2A and 4A and repeated measures ANOVA for data displayed in Figure 3B . For the flow-Fish analysis ( Figure 4B ) , we also utilized normal probability plots and the Fisher exact test comparing two proportions to examine regions over which there was a greater proportion of longer telomeres in older subjects . Vertical bars in figures denote SE . The Boston Medical Center and the New Jersey Medical School Institutional Review Boards , the review boards at each NHLBI FHS center , the Scientific-Ethical Committee for Vejle and Funen Counties , and the St . Thomas' Hospital Research Ethics Committee approved the studies . All participants gave written informed consent .
|
Leukocyte telomere length becomes shorter with age and is apparently a biomarker of aging and a forecaster of longevity in humans . Leukocyte telomere length is heritable , longer in women than in men , and is relatively shorter in persons who suffer from aging-related diseases , cardiovascular diseases in particular . This study found in four different populations that leukocyte telomere length in adult offspring was positively correlated with paternal age at the time of birth of the offspring . Analysis of telomeres in sperm of young ( <30 years ) and older ( >50 years ) donors revealed the emergence in the older donors of a subset of sperm with elongated telomeres . The mechanisms behind this enigmatic , age-dependent elongation in telomere length of sperm are unknown but may relate to epigenetic factors or the survival of a subset of germ-line stem cells , resilient against aging . It is also unknown how older fathers endow their offspring with longer telomeres in their leukocytes . The potential impact of paternal age on leukocyte telomere length and , conceivably , aging-related diseases and longevity in the offspring is of relevance because offspring of older fathers comprise an increasing proportion of society .
|
[
"Abstract",
"Introduction",
"Discussion",
"Methods"
] |
[
"homo",
"(human)",
"genetics",
"and",
"genomics",
"public",
"health",
"and",
"epidemiology"
] |
2008
|
Offspring's Leukocyte Telomere Length, Paternal Age, and Telomere Elongation in Sperm
|
Poxviruses have evolved multiple strategies to subvert signaling by Nuclear Factor κB ( NF-κB ) , a crucial regulator of host innate immune responses . Here , we describe an orf virus ( ORFV ) virion-associated protein , ORFV119 , which inhibits NF-κB signaling very early in infection ( ≤ 30 min post infection ) . ORFV119 NF-κB inhibitory activity was found unimpaired upon translation inhibition , suggesting that virion ORFV119 alone is responsible for early interference in signaling . A C-terminal LxCxE motif in ORFV119 enabled the protein to interact with the retinoblastoma protein ( pRb ) a multifunctional protein best known for its tumor suppressor activity . Notably , experiments using a recombinant virus containing an ORFV119 mutation which abrogates its interaction with pRb together with experiments performed in cells lacking or with reduced pRb levels indicate that ORFV119 mediated inhibition of NF-κB signaling is largely pRb dependent . ORFV119 was shown to inhibit IKK complex activation early in infection . Consistent with IKK inhibition , ORFV119 also interacted with TNF receptor associated factor 2 ( TRAF2 ) , an adaptor protein recruited to signaling complexes upstream of IKK in infected cells . ORFV119-TRAF2 interaction was enhanced in the presence of pRb , suggesting that ORFV119-pRb complex is required for efficient interaction with TRAF2 . Additionally , transient expression of ORFV119 in uninfected cells was sufficient to inhibit TNFα-induced IKK activation and NF-κB signaling , indicating that no other viral proteins are required for the effect . Infection of sheep with ORFV lacking the ORFV119 gene led to attenuated disease phenotype , indicating that ORFV119 contributes to virulence in the natural host . ORFV119 represents the first poxviral protein to interfere with NF-κB signaling through interaction with pRb .
Orf virus ( ORFV ) , the type member of the genus Parapoxvirus of the Poxviridae , is the causative agent of orf or contagious ecthyma , a ubiquitous disease of sheep and goats characterized by proliferative lesions affecting muco-cutaneous tissues of the mouth and muzzle [1 , 2] . Orf is a zoonotic disease affecting humans in close contact with infected animals [3–5] . Like other parapoxviruses ( PPV ) , ORFV is highly epitheliotropic , and keratinocytes and their ontogenetically related counterparts in the oral mucosa are the most important if not the only cell type to support ORFV replication in vivo [6] . In addition to producing the essential protective stratum corneum of the epidermis , keratinocytes function as immune sentinels and instigators of inflammatory responses in the skin , representing a specialized branch of the innate immune system . Keratinocytes are well equipped for pathogen sensing as they express a broad spectrum of pattern recognition receptors ( PRRs ) , including surface and endosomal toll-like receptors ( TLRs ) , NOD-like receptors ( NLRs ) , and retinoic acid-inducible gene ( RIG-I ) -like receptors , and rapidly respond to cell injury and infection by releasing critical pro-inflammatory chemokines and cytokines such as tumor necrosis factor α ( TNFα ) and interleukin 1 ( IL-1 ) [7 , 8] . Engagement of these receptors initiates downstream pro-inflammatory cascades , including the NF-κB signaling pathway , which mediates innate immune responses and contributes to skin homeostasis by regulating keratinocyte proliferation and differentiation [9] . The NF-κB family of transcription factors consists of five members , NF-κB-p65 ( RelA ) , RelB , c-Rel , NF-κB-p50/p105 , and NF-κB-p52/p100 , which contain an N-terminal Rel homology domain ( RHD ) responsible for homo- and heterodimerization and for sequence specific DNA binding [10] . In unstimulated cells , NF-κB dimers are sequestered in the cytoplasm through binding to the inhibitor kappa-B alpha ( IκBα ) . Following cell stimulation , IKK complex-mediated phosphorylation of IκBα results in proteasomal degradation of IκB and nuclear translocation of p65/p50 dimers , which bind κB-responsive DNA elements , interact with transcription co-regulators , and activate or repress gene expression [11 , 12] . The critical IKK complex consists of two kinases , IKKα and IKKβ , and the regulatory subunit IKKγ/NF-κB essential modulator ( NEMO ) [13 , 14] . Various stimuli , including those initiated by proinflammatory cytokines TNFα and IL-1 , lead to IKK activation . Engagement of the TNF receptor 1 ( TNF-R1 ) results in sequential recruitment of TRADD ( TNF-R1-associated death domain ) , TRAF2 ( TNF receptor-associated factor 2 ) and RIP1 ( Receptor-interacting protein 1 ) [15] . Multiple ubiquitination events on RIP1 and NEMO bring the TAK1 ( TGF-β activated kinase 1 ) complex close to the IKK complex . TAK1-mediated IKKβ phosphorylation and IKKβ auto-phosphorylation activate IKKβ , which then phosphorylates IκBα [16] . Engagement of the IL-1 receptor , on the other hand , results in recruitment of IRAK1 ( IL-1 receptor–associated kinase ) and activation of TRAF6 ( TNF receptor-associated factor 6 ) , which then ubiquitinates and activates TAK1 resulting in IKK activation [17 , 18] . Many viruses with dissimilar life styles are known to interfere with the NF-κB pathway . In particular , poxviruses have evolved multiple strategies to counteract NF-κB function , indicating that inhibition of NF-κB-mediated transcription is important for successful infection of the host . This is not surprising as poxvirus infections are sensed by NF-κB-activating PRRs such as endosomal TLRs , RIG-I-like receptors , and the inflammasome [19] . General features of poxviral NF-κB inhibitors include , 1- individual viruses encode for multiple inhibitors , with vaccinia virus ( VACV ) encoding at least twelve [20] . While orthologs of some NF-κB inhibitors are found in viruses belonging to multiple poxvirus genera ( e . g . VACV A52R , VACV E3L ) , others are restricted to a particular genus ( e . g . VACV A46R and VACV B14R in Orthopoxvirus ) or even selected viruses within a genus ( PPV ORFV002 ) [21–29]; 2- in contrast to other classes of poxviral immunomodulators , poxviral NF-κB inhibitors have no or little resemblance to host proteins; 3- although inhibitors target extracellular , membrane , cytosolic , or nuclear events of NF-κB regulation , most inhibitors directly target NF-κB subunits or the proximal IKKs; 4- despite the multiplicity of inhibitors , there seems to be low or no redundancy as judged by the effect of individual gene deletions on viral pathogenesis; 5- with a few exceptions ( myxoma virus MYXV150 , cowpoxvirus CPXV006 ) no single gene-deletion rendered complete virus attenuation [30–34]; 6- most inhibitors are expressed early after virus entry into cells . Apart from VACV E3L ( ORFV020 ) , PPV lack homologues of poxviral NF-κB inhibitors present in other poxviral genera , suggesting that PPV have evolved novel mechanisms to counteract the NF-κB signaling pathway . Recently , we have described four NF-κB inhibitors encoded by ORFV , ORFV024 , ORFV002 , ORFV121 and ORFV073 [35–38] . ORFV024 was shown to inhibit phosphorylation of IKK kinases , thus preventing activation of IKK complex . ORFV121 , a virulence determinant , was shown to bind to- and inhibit phosphorylation and nuclear translocation of NF-κB-p65 [37] . ORFV002 binds NF-κB-p65 and reduces its acetylation by co-activator p300 thus inhibiting transactivation [36] . Decreased NF-κB-p65 acetylation is a consequence of ORFV002 interfering with NF-κB-p65 phosphorylation by mitogen- and stress-activated protein kinase 1 ( MSK1 ) [39] . ORFV073 inhibits NF-κB signaling by preventing activation of IKK complex through interaction with NEMO , the regulatory subunit of the IKK complex [38] . The retinoblastoma tumor suppressor , pRb , is a multifunctional , predominantly nuclear protein encoded by the RB1 gene . pRb affects cell cycle , differentiation and metabolism , genome stability and apoptosis , mostly but not exclusively through transcription regulation [40] . Central to pRb function is its ability to nucleate complexes containing multiple interacting partners , thus participating in various regulatory circuits . Viruses have evolved functions to modulate those pathways by targeting pRb to their advantage . For example , adenovirus ( Ad ) protein E1A interaction with pRb and other factors represses select host genes to promote productive virus infection [41] . The human cytomegalovirus pp71 tegument protein and the human papillomavirus E7 oncoprotein bind to pRb and induce its degradation thus driving cells to a cell cycle stage that potentially favors efficient replication of viral genomes [42 , 43] . pRb has been shown to affect the regulation of the NF-κB pathway following TNFα signaling or vesicular stomatitis virus infection , although mechanisms involved are not yet defined [44 , 45] . Many viral and cellular pRb-interacting proteins contain the motif LxCxE ( where x means any amino acid ) that binds to the LxCxE cleft of the pRb pocket domain [46] . Select pRb-binding proteins that interact through LxCxE with pRb also bind p107 and p130 , the other two members of the retinoblastoma family of proteins . Here we show that ORFV119 , a LxCxE motif-containing virion protein unique to PPV , interferes with NF-κB signaling in a pRb-dependent manner early in infection by inhibiting IKK complex activation .
Primary ovine fetal turbinate ( OFTu ) cells were obtained from Howard D . Lehmkuhl ( USDA ) and were maintained in minimal essential medium ( MEM ) supplemented with 10% fetal bovine serum ( FBS ) ( Atlanta Biologicals , Flowery Branch , GA ) , 2 mM L-glutamine , gentamicin ( 50 μg/ml ) , penicillin ( 100 IU/ml ) , and streptomycin ( 100 μg/ml ) . Human osteosarcoma cells ( Saos-2 , provided by Timothy M . Fan , UIUC ) , Human embryonic kidney ( HEK 293T ) and cervical cancer ( HeLa ) cells ( obtained from American Type Culture Collection ) , were cultured in Dulbecco's modified essential medium ( DMEM ) supplemented as above . Cells were incubated at 37°C with 5% CO2 . ORFV strain OV-IA82 [47] was used as a parental virus to construct ORFV119 gene deletion virus OV-IA82-Δ119 and for experiments involving wild type virus infection . OV-IA82-Δ119 was used as parental virus to construct revertant virus OV-IA82-RV119Flag , and OV-IA82-RV119LxGxE-Flag , a revertant virus carrying a CxG substitution in ORFV119 LxCxE motif . To knockdown pRb expression in OFTu cells , siRNA directed against ovine RB1 was used . A pool of three RB1 sense ( S ) and anti-sense ( AS ) siRNA with following sequences 1-RB1- ( S ) : CTACCTATAGCAGAAGTAT , RB1- ( AS ) : ATACTTCTGCTATAGGTAG; 2-RB1- ( S ) : GCTTATATATTTGACACAA , RB1- ( AS ) : TTGTGTCAAATATATAAGC; 3-RB1- ( S ) : CTCAGATTCACCTTTATTT , RB1- ( AS ) : AAATAAAGGTGAATCTGAG ( Custom Oligos: Sigma Aldrich ) were used . Pooled RB1 siRNA ( 15 nM of each siRNA ) were transfected in OFTu cells ( 30 , 000–70 , 000/well ) using MISSION transfection reagent ( Sigma Aldrich , Cat # S1452 ) following the manufactures’ protocol . At 48 h post transfection , pRb knockdown was examined by SDS-PAGE , using antibody against pRb ( abcam , Cat # ab85607 ) . One MISSION siRNA Universal Negative Control ( SIC001 , sigma aldrich ) was included in all experiments . To obtain OV-IA82Rb- or OV-IA82-RV119LxGxE-Flag-Rb- virus stocks ( OV-IA82 and OV-IA82-RV119LxGxE-Flag viruses propagated in cells with reduced pRb levels , respectively ) , OFTu cells with reduced pRb levels ( OFTuRb- cells ) were infected with OV-IA82 or OV-IA82-RV119LxGxE-Flag virus and supernatants from infected cultures were collected at 24 h p . i and used for NF-κB-p65 nuclear translocation assays . To construct expression plasmids pORFV119Flag and pORFV119LxGxE-Flag , the ORFV119 or ORFV119LxGxE coding sequences were PCR-amplified from the OV-IA82 genome and a plasmid containing synthetic ORFV119LxGxE sequence ( Genscript ) , respectively with primers 119-3xFlag-FW ( EcoRI ) : 5’-TAAGGCCTCTGAATTCAATGGACTCTCGTAGGCTC GCTCTT-3’; 119-3xFlag-RV ( BamHI ) : 5’-CAGAATACGTGGATCCTCAATCGCTGTCG CTGTCGCCGAG-3’ and cloned into p3xFlag-CMV-10 vector ( pFlag ) ( Clontech , Mountain View , CA ) . Similarly , to obtain a ORFV119-green fluorescence protein ( ORFV119GFP ) expression vector , ORFV119 sequence was PCR amplified from OV-IA82 genome with primers 119-GFP-FW ( EcoRI ) : 5’- TAAGGCCTCTGAATTCATGGACTCTCGTAGGCTCGCTCTT-3’ and 119-GFP-RV ( BamHI ) : 5’- CAGAATACGTGGATCCAGATCGCTGTCGCTGTCGCCGA GCG-3’ , and cloned into the vector pEGFP-N1 ( Clontech , Mountain View , CA ) . DNA sequencing of plasmids confirmed the identity and integrity of the constructs . To generate gene deletion mutant virus OV-IA82-Δ119 , a recombination cassette ( pΔ119-RFP ) containing Vaccinia virus 7 . 5 promoter ( VV7 . 5 ) -driven Red Fluorescent Protein ( RFP ) gene flanked by 528 bp sequences representing ORFV119 left and right flanking regions was synthesized and cloned in pUC57 vector ( Genscript , Piscataway , NJ ) . Similarly , to generate revertant viruses OV-IA82-RV119Flag and OV-IA82-RV119LxGxE-Flag , recombination cassettes pRV119Flag-GFP and pRV119LxGxE-Flag-GFP were synthesized containing N-terminally tagged ORFV119 or ORFV119LxGxE sequences , respectively , and a GFP reporter gene , all flanked by approximately 600 bp of homologous sequence on either side to mediate recombination ( Genscript , Piscataway , NJ ) . To obtain OV-IA82-Δ119 , OFTu cells were infected with OV-IA82 ( MOI , 1 ) and transfected with recombination plasmid pΔ119-RFP . To obtain OV-IA82-RV119Flag and OV-IA82-RV119LxGxE-Flag , OFTu cells were infected with OV-IA82-Δ119 ( MOI , 1 ) and transfected with recombination plasmids pRV119Flag-GFP or pRV119LxGxE-Flag-GFP . Recombinant viruses were isolated by limiting dilution and plaque assay using fluorescence microscopy as previously described [36] . Identity and integrity of DNA sequences in purified viruses was confirmed by PCR and DNA sequencing . Semi-purified viruses were used in infection experiments . OFTu cells infected with OV-IA82 , OV-IA82-RV119Flag , OV-IA82-Δ119 , or OV-IA82-RV119LxGxE-Flag ( MOI , 0 . 1 ) were disrupted by three cycles of freeze and thaw at three days p . i . Cellular debris were removed by centrifugation at 1500 rpm for 5 min , and supernatants pelleted by ultracentrifugation at 25 , 000 rpm for 1 h . Pellets were resuspended in MEM and aliquots frozen at -80°C . Viral titers were obtained by the Spearman-Karber’s 50% tissue culture infectious dose ( TCID50 ) method . Extracellular enveloped virus ( EEV ) and intracellular mature virus ( IMV ) were purified using double sucrose gradient purification protocol with modifications [48] . OFTu cells infected with OV-IA82 , OV-IA82-Δ119 , OV-IA82-RV119LxGxE-Flag or OV-IA82-RV119Flag ( MOI , 0 . 1 ) were harvested at three days p . i . and supernatant ( EEV ) and cellular ( IMV ) fractions collected following centrifugation at 1 , 500 rpm for 5 min . Virus in supernatants was pelleted by ultra-centrifugation ( 25 , 000 rpm for 1 h ) , while cellular fractions were disrupted by three cycles of freeze and thaw and centrifuged to remove cell debris . Further purification steps were identical for EEV and IMV . Both preparations were sonicated , pelleted through a 36% sucrose cushion , and purified using double sucrose gradient centrifugation . Virus-containing bands were collected , and virus was recovered by centrifugation and resuspended in 10 mM TrisHcl . Whole cell extracts ( 60 μg ) from OV-IA82-RV119Flag infected OFTu cells , and purified EEV and IMV virion proteins ( 10 μg ) were resolved by SDS-PAGE , blotted to nitrocellulose membranes and probed with primary antibodies against Flag ( Genscript , Cat # A00187 ) , ORFV structural protein ORFV086 or Glyceraldehyde-3-Phosphate Dehydrogenase ( GAPDH ) ( sc-25778; Santa Cruz ) overnight at 4°C [49] . Blots were incubated with appropriate HRP-labeled secondary antibodies ( 1:15 , 000 ) ( anti-mouse , sc-2031 and anti-rabbit , sc-2004; Santa Cruz ) for 1 h at room temperature ( RT ) and membranes were developed using chemiluminescent reagents ( SuperSignal West Femto , Thermo Fischer Scientific ) . To examine the effect of protein synthesis inhibitor cycloheximide ( CHX ) on expression of ORFV119 and control cellular protein p53 . OFTu cells mock treated or treated with CHX ( 50 μg/ml ) ( Sigma-Aldrich , St . Louis , MO ) for 30 min were infected with OV-IA82RV119Flag in presence or absence of CHX and harvested at 30 min , 1 h , 1 . 5 h and 2 h p . i . Total protein extracts ( 50 μg ) were resolved by SDS-PAGE , blotted and transferred to nitrocellulose membranes , probed with antibody against Flag , p53 ( sc-6243; Santa Cruz ) and GAPDH ( sc-25778; Santa Cruz ) as described above . To evaluate ORFV119 expression , OFTu cells were infected with OV-IA82 or OV-IA82-RV119Flag ( MOI , 10 ) for 2 h , 4 h , 6 h , 8 h , 12 h or 24 h p . i . Total protein extracts ( 50 μg ) were resolved by SDS-PAGE , blotted and transferred to nitrocellulose membranes , and probed with primary antibody against Flag and GAPDH . Blots were then incubated with appropriate HRP-labeled secondary antibodies and developed using chemiluminescent reagents . To assess the subcellular localization of ORFV119 , OFTu cells ( 1 . 5 x 105 cells/well ) were cultured in four-well chamber slides ( ibidi , Martinsried , Germany ) for 16 h and mock infected or infected with OV-IA82-Δ119 or OV-IA82-RV119Flag ( MOI , 10 ) . Cells were fixed at 3 h , 6 h , 12 h , 16 h or 24 h p . i with 4% formaldehyde for 20 min , permeabilized with 0 . 2% Triton X 100 for 10 min , blocked with 1% bovine serum albumin ( Sigma-Aldrich , St . Louis , MO ) for 1 h , and incubated with primary antibody against Flag ( Cat # A00187-200; Genscript ) overnight at 4°C . Cells were then stained with Alexa fluor 594 labelled secondary antibody ( Thermo Fisher Scientific , Cat # A-11005 ) , counterstained with DAPI ( 2 μg/ml ) for 10 min and visualized by confocal microscopy ( A1 , Nikon ) . To assess the effect of ORFV119 in virus replication , one-step growth curves were determined in OFTu cells infected with OV-IA82 , OV-IA82-Δ119 or OV-IA82-RV119Flag ( MOI , 10 ) . Virus yields were quantitated at 6 h , 12 h , 24 h , 36 h , and 48 h p . i . as described above . To investigate the effect of OV-IA82-Δ119 infection on expression of NF-κB regulated genes , OFTu cells were mock infected or infected with OV-IA82 , OV-IA82-Δ119 or OV-IA82-RV119Flag ( MOI , 10 ) and harvested at 2 h p . i . RNA was extracted using RNeasy Mini Kit ( QIAGEN , Cat # 74104 ) and reverse transcribed as previously described [35] . mRNA for select NF-κB regulated genes was quantified using ABI Real time PCR system ( Applied Biosystems , Foster city , CA ) , Power SYBR Green PCR Master Mix ( Cat # 4368708 , Applied Bio ) and the following primers , TNFα ( Fwd: 5’-CCTTCAACAGGCCTCTGGTT-3’; Rev: 5’-GTGGGCTACCGGCTTGTTAT-3’ ) IL1β ( Fwd 5’-AAATCCCTGGTGCTGG ATAG-3’; Rev: 5’-GTTGTCTCTTTCCTCTCCTTGT-3’ ) NF-κB1 ( Fwd: 5’- CAGAGAGGA TTTCGTTTCCGT-3’; Rev: 5’-TGCAGATTTTGACCTGAGGGT-3’ ) IL36α ( Fwd: 5’-ATGTCTTCACACCTTGGCAGT-3’; Rev: 5’-ATCGGGTGTACCCTGGATAA-3’ ) TLR2 ( Fwd: 5’-TTGCTCCTGTGACTTCCTGTC-3’; Rev: 5’- GAGCGTCACAGCGGTAGC-3’ ) . Data analysis was performed as previously described [35] . Experiments were conducted with biological triplicates and at least three technical replicates . Statistical analysis was performed by using Student’s t test . The effect of ORFV119 in TNFα-induced activation of NF-κB was assessed by NF-κB-p65 nuclear translocation assay . HeLa cells transfected with control plasmid ( pEGFP-N1 ) or a plasmid encoding ORFV119-GFP fusion protein ( pORFV119GFP ) were treated with TNFα ( 20ng/ml ) at 12 h post transfection for 30 min , fixed , permeabilized and blocked as described above , incubated with primary antibody against NF-κB-p65 ( Cell Signaling , Cat # 8242s ) for 2 h , stained with Alexa fluor 594-labeled secondary antibody ( Thermo Fisher Scientific , Cat # A-11005 ) for 1 h , counterstained with DAPI , and examined by confocal microscopy . Numbers of GFP expressing cells ( approximately 300 cells/sample ) exhibiting nuclear NF-κB-p65 staining were determined in randomly selected fields and results were shown as mean percentage of GFP/ORFV119GFP expressing cells containing nuclear NF-κB-p65 over three independent experiments . Statistical analysis of data was performed by using the Student’s t test . To examine the role of ORFV119 on NF-κB-p65 nuclear translocation during ORFV infection , OFTu cells were mock infected or infected with OV-IA82 , OV-IA82-RV119Flag , OV-IA82-Δ119 or OV-IA82-RV119LxGxE-Flag ( MOI , 10 ) . Cells were fixed at 30 min , 1 h , 2 h , 4 h , and 6 h p . i . and processed for NF-κB-p65 staining as described above . Cells ( approximately 300/sample ) were randomly selected and scored as mean percentage of cells containing nuclear NF-κB-p65 over three independent experiments . Statistical analysis of data was performed by using Student’s t test . To investigate the effects of CHX on NF-κB-p65 nuclear translocation during ORFV infection , OFTu cells were pre-treated with CHX for 30 min , infected with OV-IA82 or OV-IA82-Δ119 in presence or absence of CHX , and fixed at 30 min or 1 h p . i . NF-κB-p65 nuclear translocation assay , scoring , quantification and analysis were performed as described above . To investigate the effect of OV-IA82 , OV-IA82Rb- , OV-IA82-RV119LxGxE-Flag and OV-IA82-RV119LxGxE-Flag-Rb- virus infection on NF-κB-p65 nuclear translocation , OFTu or OFTuRb- cells were infected with OV-IA82 , OV-IA82Rb- , OV-IA82-RV119LxGxE-Flag or OV-IA82-RV119LxGxE-Flag-Rb- virus ( MOI , 10 ) . Cells were fixed at 1 h p . i . and NF-κB-p65 nuclear translocation assay was performed as described above . Cells ( approximately 300/sample ) were randomly selected and scored for mean percentage of cells containing nuclear NF-κB-p65 over two independent experiments . Statistical analysis of data was performed by using Student’s t test . Additionally , to assess the effect of pRb on NF-κB-p65 nuclear translocation , Saos-2 cells , a pRb negative cell line , were mock infected or infected with OV-IA82 or OV-IA82-Δ119 ( MOI , 50 ) and fixed at 1 h , 1 . 5 h , 2 h , 4 h , and 6 h p . i . NF-κB-p65 nuclear translocation assay , quantification and analysis were performed as described above . HeLa cells transfected with control plasmid ( pFlag ) or a plasmid encoding ORFV119-3xFlag fusion protein ( pORFV119Flag ) were treated with TNFα ( 20ng/ml ) , and harvested 10 and 20 min post treatment . OFTu cells mock infected or infected with OV-IA82 , OV-IA82-RV119Flag , OV-IA82-Δ119 or OV-IA82-RV119LxGxE-Flag ( MOI , 10 ) were harvested at 30 min and 1 h p . i . Total protein extracts ( 50 μg ) were resolved by SDS-PAGE , blotted and transferred to nitrocellulose membranes and probed with specific antibody against phospho-IKKα/β ( Ser176/180 ) ( Cat # 2697; Cell Signaling ) , phospho-IκBα ( Ser32/36 ) ( Cat # 9246; Cell Signaling ) , phospho-NF-κB-p65 ( Ser536 ) ( Cat # 3033; Cell Signaling ) , IKKα/β ( sc-7607; Santa Cruz ) , IκBα ( sc-371; Santa Cruz ) , NF-κB-p65 ( sc-7151; Santa Cruz ) and GAPDH ( sc-25778; Santa Cruz ) or Flag . Blots were processed as described above . Protein bands were quantified for densitometric analysis using ImageJ software , Version 1 . 6 . 0 ( National Institute of Health , Bethesda , MD ) and fold changes calculated . Statistical analysis was performed by using Student’s t test . To investigate the effect of ORFV infection on NF-κB-p65 activation , OFTu cells were infected with OV-IA82 or OV-IA82-Δ119 ( MOI , 10 ) and harvested at 30 min , 1 h , 2 h , 4 h , and 6 h p . i . Total protein extract ( 50 μg ) were resolved by SDS-PAGE , blotted and transferred to nitrocellulose membranes , probed with phospho-NF-κB-p65 and NF-κB-p65 antibodies , and developed as described above . The effect of ORFV119 on poly ( I:C ) , poly ( A:T ) or ORFV DNA induced NF-κB-mediated transcriptional activity was investigated using a NF-κB promoter luciferase assay . Firefly luciferase gene under the control of NF-κB promoter ( pNF-κB-Luc ) and with a plasmid encoding sea pansy ( Renilla reniformis ) luciferase under the control of herpesvirus TK promoter ( pRL-TK ) were used in studies . HeLa cells cultured in 12-well plates were co-transfected with the vectors pNF-κB-Luc ( 450 ng; Clontech , Mountain View , CA ) , and pRL-TK ( 50 ng; Promega , Madison , WI ) and pFlag or pORFV119Flag . At 24 h after transfection , cells were induced with poly ( I:C ) ( Amersham , Pittsburgh , PA ) ( 500ng ) , poly ( A:T ) ( Inviogen , San Diego , CA ) ( 750ng ) or ORFV DNA ( 1μg ) . ORFV DNA was extracted from OV-IA82 virus stock using QIAamp DNA Blood Mini Kit ( Qiagen , Germantown , MD ) . Luciferase activities were determined at 20 h post- induction using the Dual Luciferase Reporter Assay ( Promega ) and a luminometer . Data were analyzed as previously described [35] . Statistical analysis of the data was performed by using Student's t test . To investigate the effect of ORFV119 on E2F transcriptional activity in ORFV infected cells , OFTu cells were co-transfected with a pE2F-Luc ( 450 ng; Signosis , Santa Clara , CA ) and pRL-TK ( 50 ng ) plasmids . At 24 h post transfection , cells were mock infected or infected with OV-IA82 , OV-IA82Δ119 , or OV-IA82-RV119LxGxE-Flag ( MOI = 10 ) . Firefly and sea pansy luciferase activities were measured at 1 , 2 , 4 and 6 h p . i . and expressed as relative fold changes in luciferase activity as described above . The interaction of ORFV119 or ORFV119LxGxE with pRb and ORFV119 with TRAF2 was assessed in virus-infected cells by co-immunoprecipitation . OFTu cells infected with OV-IA82-RV119Flag or OV-IA82-RV119LxGxE-Flag ( MOI , 10 ) or mock infected were harvested at 12 h p . i . Total protein extraction and co-immunoprecipitation was performed using nuclear complex Co-IP Kit ( Active Motif , Carlsbad , CA ) following the manufacturer’s instructions . For co-immunoprecipitation , total protein extracts were incubated with antibodies against Flag , TRAF2 and pRb in the high stringency buffer ( IP High buffer [Cat # 101676] , 300 mM NaCl [Cat # 101684] , detergent [Cat # 101683] , 1M DTT [Cat # 3483-12-3; sigma] , Protease inhibitor cocktail [Cat # P8340; sigma] ) overnight at 4°C , and then incubated with pre-washed 50 μl slurry of protein G agarose beads ( Cat # 16–266; Millipore ) at 4°C for 2 h . Beads were washed four times with high stringency buffer ( described above ) and bound proteins were eluted in Laemmli buffer . For immunoblot analysis , eluted proteins and control total protein cell lysates were resolved by SDS-PAGE , blotted and transferred to nitrocellulose membranes , probed with antibodies against Flag , TRAF2 and pRb and developed as described above . Light chain specific antibody against Rabbit IgG ( Cat # ab99697; Abcam ) was used for TRAF2 blots . A flag-expressing protein ORFV113Flag was used as a control for specificity of ORFV119 and TRAF2 interaction . The interaction of pRb with TRAF2 in the presence of pORFV119Flag also was assessed by co-immunoprecipitation . Antibodies for co-immunoprecipitation included anti-pRb ( Cat # 9309s , Cell Signaling ) , anti-TRAF2 ( Cat # sc-876 , Santa cruz ) and anti-Flag . Co-immunoprecipitations of total proteins extracts were performed as described above . The interaction of ORFV119 or ORFV119LxGxE with putative cellular binding partners was assessed by co-immunoprecipitation in pORFV119Flag or pORFV119LxGxE-Flag transfected HEK 293T or HeLa cells at 12 h post transfection . Antibodies for co-immunoprecipitation included pRb ( Fig 3A , B-Cat # ab85607 , abcam ) ( Fig 3E , F-Cat # 9309s , Cell Signaling ) , TRAF2 ( Cat # sc-876 , Santa cruz ) , TAK1 ( Cat # sc-7126 , Santa cruz ) , RIP1 ( Cat # 3493 , cell signaling ) , TRAF6 ( sc-7221 ) or NEMO ( sc-8330 ) . Co-immunoprecipitations of total proteins extracts were performed as described above . The interaction of ORFV119 with TRAF2 also was assessed in Saos-2 cells , which do not express pRb . Saos-2 cells transfected with control plasmid ( pFlag ) and pORFV119Flag or co-transfected with pRb ( Origene , Rockville , MD ) were harvested at 12h post transfection . Co-immunoprecipitations of total proteins extracts were performed as described above . Antibodies for co-immunoprecipitation included anti-TRAF2 , anti-Flag and anti-pRb . Co-immunoprecipitation efficiency was calculated by normalizing the band intensities of co-immunoprecipitated proteins to those corresponding immunoprecipitated proteins and to the expression of corresponding input lysates as previously described [50] . Five-month-old lambs randomly allocated to three experimental groups were inoculated with either OV-IA82-Δ119 ( n = 4 ) or OV-IA82-RV119Flag ( n = 4 ) , or mock-infected ( n = 3 ) . Following anesthesia , the mucocutaneous junction of the right lower lip near the labial commissure was scarified along a two-centimeter line , and virus inoculum ( 0 . 5 ml ) containing 107 . 5 TCID50/ml was applied topically using cotton swabs . In addition , the inner sides of hind limbs were scarified in a five-centimeter line and inoculated as above . Animals were monitored for 21 days for the presence of characteristic orf lesions . Pictures were taken of the labial inoculation sites at days 3 , 5 , 9 , 12 , 16 and 21 p . i . and the lesion sizes measured with a ruler . Skin biopsy specimens from hind limb inoculation sites were collected at days 2 , 5 , 8 , 12 and 21 p . i . fixed in 10% buffered formalin , embedded in paraffin , sectioned , and stained with hematoxylin and eosin using standard methods . All animal procedures were approved by University of Nebraska-Lincoln Institutional Animal Care and Use Committee ( IACUC; protocol 1318 ) and were performed in accordance with the Guide for the Care and Use of Agricultural Animals in Agricultural Research and Teaching . PPV ORFV119 amino acid sequences were aligned using Clustal Omega ( EMBL-EBI ) . ORFV GeneBank accession numbers are ( virus strains in parentheses ) AY386263 ( OV-IA82 ) , AAP89015 ( Orf11 ) , ABA00637 ( NZ2 ) , 9AHH34303 ( B029 ) , ADY76823 ( D1701 ) , NP957896 ( OV-SA00 ) , AKU76741 ( OV-GO ) , AKU76609 ( OV-YX ) , AHZ33817 ( NA1/11 ) and KP010356 ( OV-SJ1 ) . PCPV and BPSV GeneBank accession numbers are AEL20654 ( PCPV F00-120R ) , AEO18268 ( PCPV It1303/05 ) , NP958027 ( BPSV BV-AR02 ) and KM875471 ( BPSV BV-TX09c5 ) .
ORFV119 encodes for proteins of 170 to 206 amino acids , with predicted molecular weights of 18 . 6 to 22 . 2 kDa and percentage amino acid identities ranging from 77% to 100% ( Fig 1 ) . Homologs in Pseudocowpox virus ( PCPV ) and Bovine papular stomatitis virus ( BPSV ) are 89% and 54–56% identical to OV-IA82 ORFV119 , respectively , whereas no ORFV119 homologue was found in the HL953 strain of parapoxvirus of red deer in New Zealand [51] . ORFV119 lacks homology to known proteins outside the PPV genus , and domains suggestive of protein function . However , a LxCxE motif ( OV-IA82 ORFV119 positions 192–196 ) and a downstream stretch of acidic amino acids usually found in LxCxE motif-containing proteins are located at the C-terminus of the protein and highly conserved in PPV119 proteins ( Fig 1 ) . The LxCxE motif is required by several cellular and viral proteins to interact with members of the retinoblastoma family of proteins , which controls important aspects of cell physiology , including cell cycle progression , differentiation , and apoptosis [52] . The ORFV119 LxCxE motif and surrounding sequences follow the pattern XLXCXEXXX , where X should not be a positively charged amino acid ( e . g . Lys or Arg ) and X should preferably be a hydrophobic residue , which is predicted to bind pRb with high affinity [46] . A predicted mitochondrial localization sequence is found in the N-terminus of most ORFV strains and in PCPV ( underlined in Fig 1 ) , but not in BPSV119 . The kinetics of ORFV119 was assessed during ORFV replication in OFTu cells by Western blot using a recombinant virus expressing N-terminally Flag-tagged ORFV119 ( OV-IA82-RV119Flag ) . A protein of approximately 32 kDa was detected at 2 h p . i . with increasing protein levels observed at later time points . At 24 h p . i . , the last time point investigated , an additional slightly higher molecular weight species of ORFV119 ( approximately 35 kDa ) was detected ( Fig 2A ) . The observed protein molecular weight was approximately 10 kDa higher than predicted , suggesting that the protein is post-translationally modified in some manner . Previous reports on ORFV replication have shown that early genes were expressed as early as 1 h p . i . , with late genes expressed between 6 to 12 h p . i [35–38] . Newly replicated viral DNA was first detectable at 4 to 6 h p . i . accumulating rapidly between 8 and 16 h p . i [53] . Infectious virus was detected between 16 and 18 h p . i . , with continuous virus production until 40 h p . i . [53] . Thus , ORFV119 is a early viral protein . OV-IA82-Δ119 , a virus lacking the ORFV119 gene , exhibited growth kinetics and virus yields comparable to those of wild type OV-IA82 and revertant OV-IA82-RV119Flag viruses in single step growth curves in OFTu cells , indicating that the gene is non-essential for growth in these cells ( S1 Fig ) . To examine the intracellular localization of ORFV119 , OFTu cells were infected with OV-IA82-RV119Flag and examined by confocal microscopy at various times post-infection . ORFV119 staining was not evident at 3 and 6 h time points . At 12 h p . i . , weak punctate ORFV119 staining was observed in the cytoplasm and adjacent to the plasma membrane . By 16 h p . i . , enhanced ORFV119 staining was evident , and fluorescent circular to ovoid structures ( 389±30nm ) were observed in the cytoplasm and especially in close proximity to the cell membrane . Remarkably , at late times p . i . ( 24 h ) , OV-IA82-RV119Flag infected cells exhibited abundant ORFV119 nuclear staining ( Fig 2B ) . Fluorescence was specific for Flag-tagged ORFV119 as no signal was observed in cells infected with OV-IA82-Δ119 at all examined times p . i . ( Fig 2B ) . To investigate whether ORFV119 interacts with retinoblastoma protein pRb , 293T and HeLa cells were transfected with plasmids pORFV119Flag ( ORFV119Flag ) , pORFV119LxGxE-Flag ( ORFV119LxGxE ) or control plasmid ( pFlag ) , and protein extracts were prepared at 12 h post-transfection . Reciprocal co-immunoprecipitation assays with either anti-Flag or anti-pRb antibodies demonstrated that ORFV119 co-immunoprecipitates with pRb in both cell types . Co-immunoprecipitation of ORFV119 and pRb , however , was not observed with pORFV119LxGxE-Flag , a plasmid encoding ORFV119 in which C in the LxCxE motif was replaced by G , a change shown to abrogate interaction with pRb ( Figs 3A and 3B , S2 ) [54] . To confirm the interaction in the context of the virus infection , OFTu cells were mock-infected or infected with OV-IA82-RV119Flag or OV-IA82-RV119LxGxE-Flag ( MOI , 10 ) , and cell lysates prepared at 12 h p . i . Reciprocal co-immunoprecipitation with either anti-Flag or anti pRb antibodies showed that ORFV119 but not ORFV119LxGxE-Flag co-immunoprecipitates with pRb ( Fig 3C–3F ) . Together , these results indicate that ORFV119 directly or indirectly interacts with pRb . The observation that the integrity of the LxCxE motif is required for the interaction further suggests that ORFV119 might directly bind pRb . Preliminary RNA-Seq experiments indicated increased transcription of multiple NF-κB regulated genes in cells infected with OV-IA82-Δ119 compared to cells infected with OV-IA82 virus , suggesting that ORFV119 inhibits NF-κB signaling . To rule out any confounding effect from cytokines that potentially might be present in the virus inocula , viruses used in these studies were semi-purified as described in Materials and Methods . Real-time PCR analysis of gene expression showed increased levels of NF-κB-regulated genes TNFα ( 6 . 68-fold ) , TLR2 ( 6 . 48-fold ) , NF-κB1 ( 3 . 26-fold ) and IL36α ( 3 . 7-fold ) in cells infected with OV-IA82-Δ119 compared to OV-IA82 at 2 h p . i ( Fig 4A ) . To assess the effect of ORFV119 on NF-κB-p65 nuclear translocation , OFTu cells were mock infected or infected with OV-IA82 , OV-IA82-Δ119 , OV-IA82-RV119LxGxE-Flag or OV-IA82-RV119Flag and NF-κB-p65 localization was examined by immunofluorescence . Infection with OV-IA82-Δ119 and OV-IA82-RV119LxGxE-Flag but not OV-IA82 or OV-IA82-RV119Flag led to rapid nuclear translocation of NF-κB-p65 as early as 30 min p . i . ( Fig 4B and 4C ) . The effect was transient as the percentage of cells expressing nuclear NF-κB-p65 returned to those in wild type and revertant virus-infected cells between 1 and 2 h p . i . ( Fig 4C , P<0 . 05 ) . Notably , levels of NF-κB-p65 nuclear translocation observed for OV-IA82-RV119LxGxE-Flag were significantly reduced compared to those observed with OV-IA82-Δ119 ( Fig 4C ) . Consistent with the nuclear translocation kinetics , levels of phosphorylated NF-κB-p65 ( Ser536 ) , which accumulates in the cytoplasm prior to nuclear translocation , were increased at early times p . i . ( 30 min and 1 h ) with OV-IA82-Δ119 ( S3 Fig ) . Together , these data show that ORFV119 is a poxviral NF-κB inhibitor acting transiently very early in infection and that ORFV119 LxCxE motif is important for the full inhibitory activity of the protein . To explore the possibility that pRb transcriptional activity is involved in ORFV119 inhibition of NF-κB signaling in virus infected cells , we examined E2F-mediated gene transcription early in infection . OFTu cells were transfected with a plasmid encoding for a firefly luciferase reporter gene under the control of a E2F-responsive promoter and at 24 h post transfection cells were mock infected or infected with OV-IA82 , OV-IA82Δ119 or OV-IA82-RV119LxGxE-Flag . Luciferase activities were measured at 1 , 2 , 4 and 6 h p . i . Similar low luciferase activity was observed in mock and virus-infected cells at 1 , 2 and 4 h p . i . Significantly higher luciferase activity was observed in virus infected cells at 6 h p . i . ; however , no significant difference was observed among the three viruses ( S4 Fig ) . Thus , data suggest that ORFV119 mediated NF-κB inhibition does not involve E2F-mediated gene transcription early in infection . To further investigate the mechanism of ORFV119 in NF-κB inhibition , OFTu cells were mock-infected or infected with OV-IA82 , OV-IA82-Δ119 , OV-IA82-RV119LxGxE-Flag or OV-IA82-RV119Flag for 30 min or 1 h p . i . and phosphorylation of IKKα/β , IκBα and NF-κB-p65 was assessed by Western blot . Infection by OV-IA82Δ119 and OV-IA82-RV119LxGxE-Flag led to marked and early phosphorylation of IKKα/β ( Ser176/180 ) , IκBα ( Ser32/36 ) and NF-κB-p65 ( Ser536 ) compared to OV-IA82 infected cells ( Fig 5A ) . In OV-IA82Δ119-infected cells , relative fold increases of phosphorylated IKKα/β ( 14 . 5 and 19 . 7 fold ) , IκBα ( 21 . 5 and 11 . 23 fold ) and NF-κB-p65 ( 9 . 24 and 15 . 35 fold ) were observed at 30 min and 1 h p . i . , respectively ( Fig 5B and 5C ) . Similarly , in OV-IA82-RV119LxGxE-Flag-infected cells , relative fold increases of phosphorylated IKKα/β ( 27 . 35 and 21 . 42 fold ) , IκBα ( 19 . 35 and 21 . 24 fold ) and NF-κB-p65 ( 13 and 13 . 24 fold ) were observed at 30 min and 1 h p . i . , respectively ( Fig 5B and 5C ) . These results indicate that ORFV119 inhibits phosphorylation of the IKK complex , a NF-κB activating event . The observation of ORFV119 staining structures of approximate virion size in infected cells at 16 h p . i . ( Fig 2B ) together with the early inhibitory effect of ORFV119 on NF-κB signaling suggested ORFV119 may be a virion component available during and/or immediately after virus entry . To examine this possibility , extracellular enveloped virus ( EEV ) and intracellular mature virus ( IMV ) were purified from OFTu cells infected with an OV-IA82 , OV-IA82-Δ119 , OV-IA82-RV119LxGxE-Flag and OV-IA82-RV119Flag virus . As in infected cell extracts at 24 h p . i . ( Fig 2A ) , western blot analysis showed a protein doublet ( ~32 and 35 kDa ) corresponding to ORFV119Flag or ORFV119LxGxE-Flag in both virion fractions . As expected , the ORFV119 protein was not detected in OV-IA82 ( ORFV119 lacks Flag tag ) and OV-IA82-Δ119 virions ( Fig 6A ) . A control virion core protein ORFV086 was detected as a predominant 21 kDa band in OV-IA82 , OV-IA82-Δ119 , OV-IA82-RV119LxGxE-Flag and OV-IA82-RV119Flag EEV and IMV virions [49] ( Fig 6A ) . As a control for potential contamination of purified virions with cellular proteins , a GAPDH control was used . GAPDH protein was not detected in OV-IA82 , OV-IA82-Δ119 , OV-IA82-RV119LxGxE-Flag and OV-IA82-RV119Flag purified EEV and IMV virions ( Fig 6A ) . To assess whether early inhibition of NF-κB-p65 nuclear translocation by ORFV119 involves de novo viral protein synthesis in the infected cells , OFTu cells were pre-treated with the protein synthesis inhibitor cycloheximide ( CHX ) for 30 min followed by infection with OV-IA82 or OV-IA82-Δ119 for 30 min and 1 h in presence of the drug . If de novo synthesis of ORFV119 is required for inhibiting NF-κB signaling , then increased levels of NF-κB activation should be observed in OV-IA82 infected CHX treated cells . Low levels of ORFV119 proteins were detected at 30 min and 1 h p . i . likely representing input virion-associated protein; however , ORFV119 protein levels declined in CHX-treated OV-IA82-RV119Flag infected OFTu cells at subsequent times ( Fig 6B ) . Under these treatment conditions , expression of control host protein p53 also was inhibited ( Fig 6C ) . Immunofluorescence analysis showed that inhibition of NF-κB-p65 nuclear translocation was unaltered in OV-IA82 infected cells regardless of CHX treatment ( Fig 6D and 6E ) . Together , these results indicate that ORFV119 is a virion component , and suggest that virion-associated ORFV119 alone is responsible for early inhibition of NF-κB signaling . To determine if ORFV119 alone is sufficient for inhibiting TNFα-induced nuclear translocation of NF-κB-p65 , immunofluorescence assays were performed in HeLa cells transiently expressing GFP or ORFV119-GFP fusion protein ( 119GFP ) . Following TNFα induction ( 30 min ) ORFV119-GFP-expressing cells exhibited significantly reduced nuclear translocation of NF-κB-p65 ( 12 . 7% ) compared to control cells expressing GFP alone ( 68% ) ( Fig 7A and 7B , P<0 . 05 ) . ORFV119 effect on TNFα induced activation of NF-κB-p65 was further investigated by examining phosphorylation of IKKα/β ( Ser176/180 ) , IκBα ( Ser32/36 ) and NF-κB-p65 ( Ser536 ) in HeLa cells transfected with pFlag or pORFV119Flag plasmids . ORFV119 expression markedly reduced the TNFα induced phosphorylation of IKKα/β ( 20 and 50% ) , IκBα ( 25 and 68% ) and NF-κB-p65 ( 40 and 60% ) in cells expressing ORFV119Flag compared to control pFlag expressing cells at 10 and 20 min after TNFα induction . ( Fig 8A–8D , P<0 . 05 and P<0 . 01 ) . Together , results indicate that ORFV119 inhibits TNFα-induced NF-κB signaling by preventing activation of the IKK complex in the absence of any other viral protein . To examine if ORFV119 also affected poly ( I:C ) , poly ( A:T ) or ORFV DNA-induced NF-κB transcriptional activity , HeLa cells were co-transfected with pFlag or pORFV119Flag together with a plasmid encoding for a firefly luciferase reporter gene under the control of a NF-κB-responsive promoter . Cells were induced with poly ( I:C ) , poly ( A:T ) or ORFV DNA at 24 h post-transfection and luciferase activities were determined at 20 h post-induction . No significant effect of ORFV119 expression on poly ( I:C ) , poly ( A:T ) or ORFV DNA-induced NF-κB-mediated transcription was observed ( S5 Fig ) . Thus , data suggest that ORFV119 functions primarily through TNFα-induced NF-κB signaling . Given that: 1 ) ORFV119 interacts with pRb ( Fig 3 ) , 2 ) the ORFV119 LxCxE motif is required for that interaction ( Fig 3 ) and 3 ) OV-IA82-RV119LxGxE-Flag—a virus containing a mutation in the ORFV119 LxCxE motif that abrogates pRb binding—was unable to efficiently inhibit NF-κB signaling ( Fig 5 ) , we examined the involvement of pRb in ORFV119-mediated inhibition of NF-κB signaling and NF-κB-p65 nuclear translocation in cells either lacking or with reduced levels of pRb . Cells with reduced levels of pRb ( OFTuRb- ) were prepared from OFTu cells using siRNAs targeting ovine RB1 ( see Materials and Methods ) . pRb protein knockdown of approximately 60% was routinely obtained for RB1 siRNA-transfected cells at 48 h post-transfection ( Fig 9A , lanes 2 and 3 ) . Given the possibility that pRb may be associated with the virion due to its interaction with ORFV119 , virus stocks were prepared in either OFTu or OFTuRb- cells ( see Materials and Methods ) . To evaluate the effect of reduced pRb levels on ORFV119 ability to inhibit NF-κB signaling , NF-κB-p65 nuclear translocation assays were performed in OFTu or OFTuRb- cells using OV-IA82 , OV-IA82Rb- ( OV-IA82 virus propagated in cells with reduced pRb levels ) , OV-IA82-RV119LxGxE-Flag or OV-IA82-RV119LxGxE-Flag-Rb- virus ( OV-IA82-RV119LxGxE-Flag virus propagated in cells with reduced pRb levels ) virus stocks . As expected from data described above for OV-IA82 ( Figs 4 and 5 ) , levels of NF-κB-p65 nuclear translocation at 1 h p . i . following infection of OFTu cells using OV-IA82 virus were low ( 1 . 5% positive nuclei ) . However , significantly increased NF-κB-p65 nuclear translocation was observed for treatment conditions where either OFTuRb- cells ( OFTuRb- cells/ OV-IA82 virus ) or OV-IA82Rb- virus ( OFTu cells/ OV-IA82Rb- virus ) were used ( 6 . 1 and 6 . 5% NF-κB-p65 positive nuclei , respectively ) . Notably , using both OFTuRb- cells and OV-IA82Rb- virus resulted in significantly increased levels of NF-κB-p65 nuclear translocation in infected cells ( 17 . 2% NF-κB-p65 positive nuclei ) ( Fig 9B and 9C ) . As expected for a ORFV119 protein lacking the pRb binding motif ( LxCxE ) , no significant difference was observed in NF-κB-p65 nuclear translocation for treatment conditions where OV-IA82-RV119LxGxE-Flag or OV-IA82-RV119LxGxE-Flag-Rb- viruses were used to infect either OFTu or OFTuRb- cells ( 44 . 25% and 49 . 2% NF-κB-p65 positive nuclei , respectively ) ( S6 Fig ) . Thus , pRb contributes to the ORFV119-mediated inhibition of NF-κB signaling in infected OFTu cells . Additional experiments examining involvement of pRb in ORFV119-mediated inhibition of NF-κB signaling were conducted in human osteosarcoma Saos-2 cells , a pRb-deficient cell line [55] . pRb was not detected in Saos-2 cell extracts by western blot . ( Fig 9A , Lane1 ) . If pRb is mediating ORFV119 inhibition of NF-κB-p65 nuclear translocation , increased NF-κB-p65 nuclear translocation would be expected in OV-IA82 virus infected Saos-2 cells . Saos-2 cells were mock infected or infected with OV-IA82 or OV-IA82-Δ119 and NF-κB-p65 localization was examined by immunofluorescence at indicated times p . i . No significant differences in NF-κB-p65 positive nuclei in OV-IA82-or OV-IA82-Δ119-infected cells were observed at any time post-infection ( Fig 10A and 10B ) . Thus , in the absence of pRb , the early NF-κB inhibitory phenotype observed for OV-IA82 is lost . Together , data using cells with reduced pRb levels or lacking pRb altogether indicate that ORFV119-mediated inhibition of NF-κB-signaling is largely pRb-dependent . To assess potential interactions of ORFV119 with components of the TNFα-induced NF-κB signaling pathway , co-immunoprecipitation assays were conducted with antibodies against TRAF2 , TAK1 , RIP1 , TRAF6 and NEMO . OFTu cells were mock-infected or infected with OV-IA82-RV119Flag ( MOI , 10 ) and harvested at 12 h p . i . Total cell protein extracts were immunoprecipitated with anti-Flag or anti-target cellular protein antibodies . Reciprocal co-immunoprecipitation demonstrated that ORFV119 co-immunoprecipitates with TRAF2 ( Fig 11A and 11B ) . Similar results were obtained in 293T cells transiently expressing pORFV119Flag ( Fig 11E and 11F ) ; however , no interaction was observed between TRAF2 and transiently expressed ORFV119LxGxE ( S7 Fig ) . As a control for specificity of ORFV119 and TRAF2 interaction , a Flag tagged ORFV113Flag was used . No interaction was observed between ORFV113Flag and TRAF2 in OFTu cells infected with OV-IA82-RV113Flag ( Fig 11C and 11D ) nor in 293T cells transfected with pORFV113Flag ( Fig 11G and 11H ) . In the context of viral infection , reciprocal co-immunoprecipitation of ORFV119 with TAK1 , RIP1 , TRAF6 and NEMO was not observed . Together , these results indicate that ORFV119 directly or indirectly interacts with the scaffold protein TRAF2 . Further , dependence on a LxCxE motif for interaction suggested that pRb might play a role in ORFV119-TRAF2 complex formation . As ORFV119-mediated inhibition of the NF-κB signaling is pRb dependent ( Figs 4 , 5 , 9 and 10 ) and ORFV119 interacted with TRAF2 in a LxCxE motif-dependent manner ( S7 Fig ) , we hypothesized that a ORFV119-pRb complex may be required to efficiently interact with TRAF2 . To examine this possibility , 293T cells were transfected with pFlag or pORFV119Flag and harvested at 12h p . i . Reciprocal co-immunoprecipitation assays using total cellular protein extracts with either anti-TRAF2 or anti-pRb antibodies demonstrate that TRAF2 and pRb co-immunoprecipitation , while weak or absent in pFlag transfected cells , is enhanced by the presence of ORFV119 in pORFV119Flag transfected cells ( Fig 12A and 12B ) . To further evaluate a pRb requirement for efficient ORFV119-TRAF2 interaction , co-immunoprecipitation experiments were performed in Saos-2 , a pRb-deficient cell line . In contrast to results described above ( Fig 11 ) , where ORFV119-TRAF2 interaction was observed in pRb expressing OFTu and 293T cells , no interaction was detected in Saos-2 cells ( Fig 12D and 12E ) . Notably , transfection of Saos-2 cells with both pORFV119 and pRb plasmids restored the interaction as reciprocal co-immunoprecipitation of ORFV119 and TRAF2 were observed ( Fig 12F and 12G ) . Taken together , these data indicate that pRb is important for ORFV119-TRAF2 interaction and , further , they suggest that a ORFV119-pRb complex may be required for efficient interaction with TRAF2 . The contribution of ORFV119 to virus virulence was investigated in sheep , a natural ORFV host . Animals were inoculated with OV-IA82-Δ119 ( n = 4 ) , OV-IA82-RV119Flag ( n = 4 ) or PBS ( control group , n = 3 ) in the right labial commissure and the inner side of the thighs , and disease course was monitored for 21 days . All virus-inoculated animals developed clinical orf as evidenced by erythema , papules , pustules , and scabby tissue deposition on the lips ( S8A Fig ) . However , beginning on day 3 p . i . , lesions in sheep inoculated with OV-IA82-RV119Flag were 25% to 90% larger and exhibited more extensive pustules and scabby tissue deposition than those inoculated with deletion mutant virus ( S8A and S8B Fig ) . By day 16 p . i . , lesions in all sheep inoculated with OV-IA82-Δ119 were resolved . In contrast , three of four sheep inoculated with OV-IA82-RV119Flag ( sheep # 79 , #101 , and #639 ) still exhibited well-defined lesions at day 16 p . i . and regressing lesions still were present in two of the animals ( #79 and #101 ) at day 21 p . i . ( end point of the experiment ) ( S8A Fig ) . No lesions were observed in PBS-inoculated controls . Histological examination of punch biopsies collected from the thighs at various times post-infection showed no significant differences in time to onset and type of skin changes between groups of virus-inoculated animals . Overall , results indicate that lesion development is more restricted and time to resolution is reduced in sheep inoculated with OV-IA82-Δ119 compared to revertant virus . Thus , ORFV119 contributes to ORFV virulence in the natural host .
The NF-κB signaling pathway plays key roles in the skin by regulating innate immune responses , inflammation , cell survival and cell proliferation [56–58] . Crucial cytoplasmic and nuclear events in the signaling pathway are targeted by various proteins encoded by the highly epitheliotropic ORFV [35–37] . Here , we describe an ORFV protein , ORFV119 , that interacts with pRb and prevents activation of NF-κB signaling very early during infection . Inhibition of NF-κB-signaling by ORFV119 is largely dependent on its ability to interact with pRb , which prevents activation of the IKK complex and downstream NF-κB signaling . ORFV119 was shown to interact with pRb in infected and uninfected cells , indicating that other ORFV proteins are not required for the interaction . Similar to oncoproteins of small DNA viruses , binding to pRb was dependent on the ORFV119 LxCxE motif since a CxG substitution in the motif abrogated the interaction completely ( Fig 3 ) . This suggests that ORFV119 , like small DNA virus oncoproteins , might directly bind pRb . Interestingly , another highly epitheliotropic poxvirus , molluscum contagiosum virus ( MCV ) , encodes a protein ( MC007L ) unrelated to ORFV119 that localizes to mitochondria and interacts with pRb through an LxCxE motif [59] . The function of MC007L in MCV infection is unknown . pRb has not been previously implicated in antiviral responses against poxviruses; however , it has been associated with NF-κB modulation by other viruses . For example , transiently expressed adenovirus E1A , a pRb-binding oncoprotein , was shown to inhibit NF-κB-dependent transcription induced by TNFα , the effect being dependent on E1A/pRb interaction [44] . And pRb was shown to be required for the activation of the NF-κB pathway in response to vesicular stomatitis virus infection , although viral mechanisms involved were not described [45] . Through association with pRb , ORFV119 inhibits NF-κB signaling by preventing IKK complex activation ( Fig 5 ) , with ORFV119-TRAF2 interaction likely underlying the inhibition ( Fig 11 ) . Results suggest that a ORFV119-pRb complex may be required for efficient interaction of ORFV119 with TRAF2 , leading to inhibition of NF-κB signaling ( Fig 12 and S7 Fig ) . TRAF2 is a RING finger protein recruited to TNF receptors to regulate NF-κB signaling both positively and negatively [60] . Viral proteins that interact with TRAF2 activate or inhibit NF-κB signaling . For example , the MCV protein MC159 was shown to interact with both TRAF2 and NEMO and to inhibit NF-κB activation [61] , while Kaposi’s sarcoma associated-herpes virus ( KSHV ) vFLIP and rotavirus VP4 proteins interact with TRAF2 activating NF-κB signaling [62 , 63] . The effect of these interactions on TRAF2 function remains unknown . Conceivably , viral proteins might interfere with NF-κB signaling by modulating TRAF2 E3 ubiquitin ligase activity and/or by affecting TRAF2 scaffold functions . ORFV119 is a virion protein ( Fig 6 ) functioning very early in infection ( ≤ 30 min ) to inhibit NF-κB signaling . Observed early IKK complex activation on infection of cells with virus lacking ORFV119 ( OV-IA82Δ119 ) indicates ORFV119 inhibits NF-κB signaling induced by an early infection event . Early inhibition of IKK complex activation during infection , ORFV119-TRAF2 interaction in infected cells , and a proposed role for TRAF2 in poxvirus entry [64] suggest that ORFV119 may be inhibiting ORFV entry mediated activation of NF-κB signaling . The importance of viral inhibition of NF-κB activation very early in ORFV infection is further supported by the presence of a second virion-associated NF-κB inhibitor ORFV073 ( Fig 13 ) . ORFV073 inhibits NF-κB signaling by preventing activation of IKK complex through interaction with NEMO , the regulatory subunit of the IKK complex [38] . As early infection events , including those related to cell entry , are likely conserved among poxviruses [65] , early inhibition of NF-κB signaling in poxvirus infected cells may be of greater biological significance than currently appreciated . Notably , at late times post-infection ( ≥24 h p . i ) ORFV119 also is observed in the nucleus of infected cells ( Fig 2 ) , suggesting that in addition to the early virion-associated NF-κB inhibitory function described here , the protein may perform additional functions in the infected cell . The fact that most of pRb localizes to the cell nucleus raises the question as to whether ORFV119 interacts with nuclear pRb at late times post-infection affecting pRb functions such as transcriptional control and cell cycle regulation . Other ORFV NF-κB inhibitors have been detected in the nucleus . ORFV002 localized to the nucleus and inhibited nuclear phosphorylation of NF-κB-p65 by interacting with mitogen stimulated stress kinase 1 ( MSK1 ) [36 , 39] . And , ORFV073 , a virion-associated NF-κB inhibitor , also was shown to localize to the nucleus yet a nuclear function for the protein has not been described [38] . Among other poxviruses , vaccinia virus protein K1 localized to the nucleus where it prevented acetylation of NF-κB-p65 [66] . Examination of published parapoxviral genomes shows that a genomic region encompassing ORFV119 is affected sporadically by gaps or deletions of as much as 5 . 7 kbp [67–69] . Notably , while the extent of the deletion varied considerably , ORFV119 was always affected . The genomic sequence of the PCPV F00-120R strain , for example , was found to contain a 5 . 5 kbp DNA deletion that removed genes 116–121 [67] . Retrospective PCR analysis , however , strongly suggested that loss of the 5 . 5 kbp region occurred during virus passage in tissue culture [68] . Deletions affecting this region were also found in ORFV isolates OV-NP ( 5 . 7 kbp deletion ) and OV-SJ1 ( 1 . 5 kbp deletion ) , which were isolated from goat lesion material [69] . Although retrospective PCR analysis was not conducted , these virus isolates ( cell culture-passage 6 ) were fully attenuated following inoculation of goats , further supporting the notion that the genomic loss occurred during early passage in cells , and indicating that important virulence functions are encoded within this genomic region . Genomic deletions resulting in complex fragmentation of the BPSV ORFV119 homolog have also been reported [70] . An explanation for instability of this genomic region upon virus passage in cell culture is lacking; however , it is possible that loss of these genes may be advantageous for virus growth in cell culture . Here , infection with a deletion mutant virus lacking ORFV119 resulted in restricted lesion development and reduced time to resolution compared with revertant virus infection in the natural host . The attenuated phenotype of infection likely indicates improved host control of ORFV infection in the absence of ORFV119 ( S8 Fig ) . Our results are at variance with a previous report where no differences in virulence and pathogenesis were observed between wild type and ORFV119 deletion viruses [71] . The discrepancy likely reflects differences in viral strains ( IA82 vs SHZ1 ) , extent of genomic deletion , and/or in vitro conditions for virus propagation . Overall , ORFV pathogenesis studies in the natural host using viruses lacking single NF-κB inhibitor genes have shown a remarkable spectrum of phenotypes , ranging from wild type disease ( ORFV002 , ORFV024 ) to moderate ( ORFV073 , ORFV119 ) , or marked attenuation ( ORFV121 ) [35–38] . The multiple NF-κB inhibitors encoded by a poxvirus together with the possibility of overlapping or complementing functions may explain these observations . Alternatively , specific poxviral NF-κB inhibitors may exert only subtle and perhaps transient host range effects on specific infected cells or the infected tissue microenvironment . Regardless , the impact of these subtle changes on viral fitness in nature may be difficult to fully ascertain under experimental conditions .
|
Poxviruses have evolved multiple strategies to subvert signaling by NF-κB , a crucial regulator of host innate immune responses . Viruses often encode multiple inhibitory proteins , which largely target cytoplasmic activation events of NF-κB signaling . The retinoblastoma protein ( pRb ) , a multifunctional protein best known for its tumor suppressor activity , has been suggested to affect NF-κB signaling during virus infection however , viral effectors and mechanisms of actions are unknown . Here , we identified a virion-associated orf virus NF-κB inhibitory protein , ORFV119 , which interacts with pRb . ORFV119 was shown to inhibit IKK complex activation in a pRb-dependent manner early in infection . Results show that ORFV119 interacted with both pRb and TRAF2 and that a ORFV119-pRb complex likely is required for efficient interaction with TRAF2 and inhibition of NF-κB signaling . ORFV119 represents the first poxviral protein to interfere with NF-κB signaling through interaction with pRb .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"phosphorylation",
"transfection",
"medicine",
"and",
"health",
"sciences",
"luciferase",
"poxviruses",
"pathology",
"and",
"laboratory",
"medicine",
"enzymes",
"pathogens",
"microbiology",
"enzymology",
"viral",
"structure",
"viruses",
"signal",
"inhibition",
"immunoprecipitation",
"dna",
"viruses",
"co-immunoprecipitation",
"molecular",
"biology",
"techniques",
"extraction",
"techniques",
"research",
"and",
"analysis",
"methods",
"protein",
"extraction",
"proteins",
"medical",
"microbiology",
"oxidoreductases",
"microbial",
"pathogens",
"vaccinia",
"virus",
"molecular",
"biology",
"virions",
"precipitation",
"techniques",
"biochemistry",
"signal",
"transduction",
"cell",
"biology",
"post-translational",
"modification",
"virology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"cell",
"signaling",
"organisms"
] |
2017
|
A parapoxviral virion protein targets the retinoblastoma protein to inhibit NF-κB signaling
|
Strongyloidiasis is a neglected tropical disease caused by the human infective nematodes Strongyloides stercoralis , Strongyloides fuelleborni fuelleborni and Strongyloides fuelleborni kellyi . Previous large-scale studies exploring the genetic diversity of this important genus have focused on Southeast Asia , with a small number of isolates from the USA , Switzerland , Australia and several African countries having been genotyped . Consequently , little is known about the global distribution of geographic sub-variants of these nematodes and the genetic diversity that exists within the genus Strongyloides generally . We extracted DNA from human , dog and primate feces containing Strongyloides , collected from several countries representing all inhabited continents . Using a genotyping assay adapted for deep amplicon sequencing on the Illumina MiSeq platform , we sequenced the hyper-variable I and hyper-variable IV regions of the Strongyloides 18S rRNA gene and a fragment of the mitochondrial cytochrome c oxidase subunit 1 ( cox1 ) gene from these specimens . We report several novel findings including unique S . stercoralis and S . fuelleborni genotypes , and the first identifications of a previously unknown S . fuelleborni infecting humans within Australia . We expand on an existing Strongyloides genotyping scheme to accommodate S . fuelleborni and these novel genotypes . In doing so , we compare our data to all 18S and cox1 sequences of S . fuelleborni and S . stercoralis available in GenBank ( to our knowledge ) , that overlap with the sequences generated using our approach . As this analysis represents more than 1 , 000 sequences collected from diverse hosts and locations , representing all inhabited continents , it allows a truly global understanding of the population genetic structure of the Strongyloides species infecting humans , non-human primates , and domestic dogs .
The genus Strongyloides ( nematoda: Rhabditida ) contains at least fifty different species of parasitic nematodes [1] , often showing remarkable host specificity [2] . Humans are most commonly infected with Strongyloides stercoralis , which is a soil-transmitted helminth also infecting dogs and non-human primates [1] . Strongyloidiasis is estimated to affect at least 370 million people in over 70 countries [3] , mostly within tropical and sub-tropical regions [1] . Under suitable conditions , transmission may also occur in temperate regions [4–6] . Human infection with Strongyloides fuelleborni fuelleborni which also infects non-human primates , has also been reported from Southeast Asia [7] and some African countries [8–11] . Another sub-species , S . fuelleborni kellyi at this time is only known to parasitize humans and is thought to occur exclusively in Papua New Guinea [9 , 12 , 13] . The majority of these infections are clinically innocuous , even in heavy infections [12] . However , in a small number of infants infected with S . f . kellyi , a syndrome consistent with protein-losing enteropathy occurs . This condition , referred to as “swollen belly” syndrome , has a high fatality rate [12] . Based on examination of a fragment of the 18S ( syn . SSU ) rRNA gene of each subspecies of S . fuelleborni , Dorris et al . [14] suggested that S . f . kellyi should be transferred to the species name Strongyloides kellyi . Later , Hasegawa et al . [15] examined the 18S rRNA gene of multiple species of Strongyloides , including S . stercoralis ( five from dogs , two from humans and one from a chimpanzee ) and S . f . fuelleborni ( six of non-human primate origin , and one from a human in Africa ) . This study also introduced a genotyping scheme based on nucleotide variants at the hyper-variable regions ( HVRs ) of the 18S rRNA gene , describing variations in HVRs I-IV . The same group later described HVR-IV and cytochrome c oxidase subunit 1 ( cox1 ) haplotypes of multiple Strongyloides spp . , including S . f . fuelleborni and S . stercoralis , from humans and non-human primates , predominantly from Japan and Africa [16] . This work was later expanded upon by several other investigators who continued to utilize cox1 in conjunction with the 18S HVR-I and 18S HVR-IV loci as the preferred targets for genotyping S . stercoralis [17–22] . These previous efforts to genotype S . stercoralis and S . fuelleborni involved the time consuming , costly , and difficult process of culture and DNA extraction of multiple individual larvae from each host to be tested [15 , 17–20 , 22] . This approach was required because in mixed genotype infections , chain-termination ( Sanger ) sequencing might only detect the most prevalent genotype , or generate mixed sequence chromatograms due to the presence of multiple genotypes , resulting in dual peaks that can be difficult to interpret . Furthermore , in cases where indels of differing lengths occur in the same amplicon , the chromatograms may be uninterpretable . Earlier studies have demonstrated that a single host may be infected with multiple Strongyloides spp . genotypes [17 , 20] . We considered that next-generation sequencing could be used to address some of these challenges , allowing investigators to genotype infections directly from DNA extracted from stool . In this paper , we describe a novel next-generation sequencing-based method for genotyping nematodes of the Strongyloididae family that employs three PCR assays targeting the informative 18S HVR-I , 18S HVR-IV and cox1 loci . We use this assay to test the hypothesis that additional , previously undetected genotypes of Strongyloides spp . are present globally , and provide some preliminary information on the geographic diversity of known and novel genotypes . To assist in clarity and standardization of Strongyloides spp . genotype nomenclature into the future , we also expand an existing Strongyloides spp . genotyping scheme to accommodate these novel types .
Human , domestic dog and non-human primate fecal samples found positive for Strongyloides species and preserved in ethanol , or fecal DNA extracts from frozen stool were collected from various global locations . All human and dog samples were anonymized post-diagnostic specimens , with the exception of S . stercoralis strain PV001 strain , which is maintained at the University of Pennsylvania and was kindly provided courtesy of Dr . Thomas J . Nolan , University of Pennsylvania . Human samples were collected directly into specimen containers by patients for diagnostic purposes , domestic dog samples were obtained from the ground after defecation . The Gambian baboon sample was from feces collected from the ground in the vicinity of the infected host . Multiple methods were used to confirm Strongyloides infection: real-time PCR [23] , Koga agar plate culture [24] , direct microscopy and formalin-ethyl acetate microscopy . Samples were transported to the Centers for Disease Control and Prevention for analysis either at room temperature ( ethanol preserved feces ) or on dry ice ( DNA extracts and frozen samples ) . Ethanol preserved and fresh-frozen samples were extracted upon receipt and all DNA extracts were stored at -20°C until analysis . Genomic DNA was extracted from stool samples collected from Africa , Europe and Asia ( excluding Cambodia ) as previously described [25] . Briefly , this involved the use of the MagNA Pure LC 2 . 0 Instrument ( Roche Diagnostics ) , following the DNA I Blood Cells High performance II protocol , utilizing DNA isolation kit I ( Roche Diagnostics ) . The samples from Cambodia and samples from all other regions were centrifuged at 1500 g for one minute , followed by resuspension in sterile saline solution and storage at 4°C overnight to remove excess ethanol and inhibitors . Following this , the samples were washed by centrifugation at 1500 g for one minute and resuspended in sterile saline . This suspension was immediately centrifuged a third time at 1500 g for one minute and the pellet used for isolation of DNA . DNA was extracted using a Qiagen DNeasy Power Soil DNA isolation kit ( Qiagen , Germantown , MD ) following the manufactures instructions , but with only a sixty second bead beating stage . Extracted DNA was stored at -20°C . The nuclear 18S rRNA ( HVR-I and HVR-IV ) and mitochondrial cox1 assays were designed specifically for adaptation to the Illumina sequencing platform , utilizing paired-end reads with a read length of 250 base pairs . To simplify data analysis and increase assay sensitivity , the amplicons were kept relatively short ( a maximum of ~450 base pairs ) . This would facilitate the generation of paired reads that overlap and span the entire length of the amplicon after quality/adapter trimming and merging . For the cox1 region , priming sites were selected that; ( 1 ) resolve the two known S . stercoralis lineages ( groups A and B ) previously described into different clusters [19 , 20] , ( 2 ) fulfill the criteria for amplicon length ( <450 base pairs ) , and ( 3 ) potentially detect and differentiate strongylid nematodes ( i . e . hookworms ) as well as various Strongyloides species . All primers were designed manually using Geneious Primer design software ( version 11 ) ; their sequences are shown in Table 1 . Each of the three reactions was optimized by testing amplification performance across a range of annealing temperatures , guided by the NEB Tm calculator ( https://tmcalculator . neb . com/# ! /main ) , and by testing different reaction volumes and additives ( e . g . High GC Enhancer ) . Following agarose gel electrophoresis , reaction conditions that yielded a clean , bright band reflecting high target DNA yield in the absence of spurious bands , were considered optimal . The optimal conditions for each assay are described below . For the cox1 locus , reactions were performed in a volume of 50 μL using reagents provided with the Q5 High-Fidelity DNA Polymerase ( New England Biolabs , Ipswich , MA ) , including 10 μL NEB 5X Q5 Buffer ( New England BioLabs , USA ) , 10 μL NEB 5X Q5 High GC Enhancer ( New England BioLabs , USA ) , 4 μL NEB Deoxynucleotide Solution Mix ( 10 mM each nt ) ( New England BioLabs , USA ) , 1 μL Q5 High-Fidelity DNA Polymerase ( New England BioLabs , USA ) , 2 . 5 μL forward primer ( SSP_COX1_F ) , 2 . 5 μL reverse primer ( SSP_COX1_R ) , 18 μL deionized H2O , and 2 μL DNA template . For the HVR-I and HVR-IV regions , PCRs were performed in a 25 μL reaction also using reagents provided with the NEBNext Q5 Hot Start kit , including 12 . 5 μL of HiFI PCR Mastermix ( New England BioLabs , USA ) , 1 . 5 μL forward primer ( NEW_HVR_I_F or NEW_HVR_IV_F ) , 1 . 5 μL reverse primer ( NEW_HVR_I_R or NEW_HVR_IV_R ) , 7 . 5 μL deionized H2O , and 2 μL DNA template . Each PCR run was accompanied by a positive control consisting of genomic DNA from S . stercoralis PV001 strain , a negative feces DNA extract control and a PCR grade water negative control . The S . stercoralis PV001 strain positive control further served as an internal control for sequencing analysis of samples ( i . e . sequenced multiple times to ensure lack of errors ) , as did a Strongyloides ratti control which served as a control for cross contamination at any step [21] . The reaction was performed on a GeneAmp 9700 thermocycler ( Applied Biosystems , Beverly , MA ) , using the temperature cycling conditions provided in in Table 1 . PCR products of the correct fragment size on agarose gel electrophoresis were prepared for deep-sequencing . Amplicons were purified and normalized using the SequalPrep Normalization Plate Kit ( Thermo Fisher Scientific , Waltham , MA ) and DNA library preparation performed using the NEBNext Ultra DNA Library Prep Kit for Illumina ( New England BioLabs , Ipswich , MA ) . Library indices were added using the NEBNext Mupltiplex Oligos for Illumina Index kit ( New England BioLabs , Ipswich , MA ) . Sequencing was performed using the Illumina MiSeq platform with MiSeq reagent Nano Kit v2 ( PE250bp ) reagent kits ( Illumina , San Diego , CA ) . Bioinformatic analysis of all sequence data was undertaken using a custom workflow designed in Geneious ( Geneious Prime , version 11: www . geneious . com ) . This workflow performed read quality control , assembly of contigs and 18S HVR-I and HVR-IV haplotype assignment after Jaleta et al . [19] , and included the adjustments to that typing scheme that we previously introduced [21] . As the sequence of cox1 is extremely variable , with hundreds of haplotypes , and because our cox1 amplicon is substantially shorter than those previously described , we did not assign haplotypes to our cox1 sequences . This was done to avoid confusion with other studies . Instead , cox1 sequences were assigned to clusters that were visualized by construction of cluster dendrograms with various clusters assigned to colors . Additionally , each cox1 sequence can be uniquely identified by the GenBank ( GB ) accession numbers assigned to them . For generation of cluster dendrograms , a . fasta sequence file containing all cox1 sequences was exported from Geneious and aligned using the ‘msa’ package in R ( https://www . r-project . org/ ) . Using the ‘seqinr’ package ‘dist . alignment’ function , a pairwise identity matrix was constructed , considering gaps in the identity measure . Clustering was then performed using the agglomerative nested clustering approach in the ‘agnes’ R package , using euclidean distances and the average clustering method . From this , the ‘ggtree’ R package was used to generate cluster dendrograms . To aid dendrogram annotation , images of relevant hosts were obtained from PhyloPic ( http://phylopic . org ) or prepared in house at the Centers for Disease Control and Prevention ( CDC ) . This activity was approved as research not involving human subjects , by the Office of the Associate Director for Science , Center for Global Health , at CDC . Because the study did not involve direct interaction with animals , Animal Care and Use committee approval was not required ( Protocol number 2017–535 ) .
We employed this assay previously to screen feces from Australian dogs and humans [21] and identified multiple cryptic Strongyloides spp . genotypes in some dogs , leading us to propose several adjustments to the Strongyloides spp . genotyping scheme established by previous investigators [16 , 17 , 19] . We justified these adjustments [21] by highlighting two important points: ( 1 ) the cryptic Strongyloides spp . sequences we described were identified within feces of the same host as previously described S . stercoralis genotypes; domestic dogs , and ( 2 ) , the cryptic Strongyloides spp . sequences added to the scheme were more similar to each other than to S . ratti [21] . It was noted that some of these cryptic Strongyloides spp . genotypes may have been present in the dogs due to coprophagy , though given that this is speculative these types were added to the scheme nonetheless [21] . This was considered a straightforward solution to the problem of assigning sequences an identity , thereby facilitating ease of comparison to known haplotypes and subsequent discussion . In line with these principles , we adjusted the typing scheme further in this study to include all 18S HVR-I and HVR-IV haplotypes shown in Fig 1 , and we list published examples of these sequences ( Table 2 and Table 3 ) . Sequence data was generated for 60 specimens in this study; sequencing success rates for each marker provided in Table 1 . One third of these specimens ( n = 20 ) had all markers successfully sequenced , 13 had one marker sequenced and 27 had any two markers sequenced ( 144 sequences generated in total ) . Typed specimens were predominantly from humans and domestic dogs ( Canis familiaris ) , and were obtained from several countries representing all inhabited continents ( Table 4 ) . One typed specimen was from a Guinea baboon ( Papio papio ) from The Gambia and contained S . f . fuelleborni . One human specimen from India contained S . f . fuelleborni and three human specimens from Australia also contained a S . fuelleborni . Whether this Australian sub-species of S . fuelleborni is novel or is S . f . kellyi could not be determined using the data available . The remaining Strongyloides spp . genotypes detected were attributed to S . stercoralis ( Table 4 ) . We describe 18S HVR-IV haplotype J and assign it to S . stercoralis ( Fig 1 ) . This haplotype was found in a human specimen from the USA in association with HVR-I haplotypes II and VI , HVR-IV haplotype A , and a cox1 sequence belonging to the dog/human infecting lineage ( lineage A ) of S . stercoralis ( Table 2 , Figs 2 and 3 , red cluster ) . We also describe HVR-IV haplotype T , detected in a human specimen from Australia and attributed to an undetermined sub-species of S . fuelleborni . This sequence was found in conjunction with HVR-I haplotype XII and HVR-IV haplotype M , though sequencing of cox1 from this specimen was not successful ( Table 4 ) . The novel 18S HVR-I haplotype XI is described here for the first time , and is assigned to S . stercoralis on the basis that its sequence was found twice in association with cox1 sequences belonging to S . stercoralis from the dog/human infecting lineage ( Fig 2 , red cluster ) and both times in association with 18S HVR-IV haplotype A ( Table 4 , Human 378_Au and Human 877_IvCo ) . Thirty-five cox1 sequences belonging to Strongyloides spp . were generated ( plus four from Necator americanus and two from Oesophagostomum sp . ) . Clustering of Strongyloides spp . cox1 sequences revealed several trends relating to geography , host preference and associations between certain cox1 clusters and specific 18S genotypes ( Fig 2 and Fig 3 ) . No cox1 sequences were obtained from dogs possessing HVR-I haplotypes IV and/or V , or HVR-IV haplotype B , which represent exclusively dog-infecting types that correspond to S . stercoralis lineage/type B as per Nagayasu et al . ( 2017 ) [20 , 21] . Of the 35 Strongyloides spp . cox1 sequences generated , 33 were obtained from specimens collected from humans and dogs possessing HVR-IV haplotype A . Each of these 33 cox1 sequences were assigned to the cluster highlighted in red ( Figs 2 and 3 ) containing S . stercoralis that infect both dogs and humans . This cluster corresponds to S . stercoralis lineage/type A described by Nagayasu et al . [20] . The remaining 2 Strongyloides spp . sequences of these 35 were attributed to S . fuelleborni . When all published S . fuelleborni and S . stercoralis cox1 sequences were analyzed alongside our data , several geographic trends emerged relating to the distribution of various S . fuelleborni types . We note that S . fuelleborni cox1 sequences from Malaysia , mainland SE Asia , Japan , East Africa and Central Africa each form their own distinct clusters ( Fig 2; dark blue , magenta , light blue , light pink and gray branch colors , respectively ) , suggesting that each is a potential geographic and/or host-adapted sub-variant of S . f . fuelleborni ( Figs 2 and 3 ) . We identified seven fecal specimens possessing mixed Strongyloides spp . genotypes . One specimen ( Human 14WC_US_LA ) possessed the genotype II and VI + A and J , attributed to S . stercoralis . Another from Australia ( Human 368_16_Au ) possessed the genotype XII + T and M , attributed to S . fuelleborni . A specimen from the USA ( Dog US_PA ) contained two haplotypes of S . stercoralis HVR-I ( I and VI ) though a HVR-IV genotype was not successfully ascertained . A human specimen from Ethiopia possessed the genotype I and II + A ( Human 169_Et ) , while another from Australia ( Human 378_Au ) possessed the genotype III and XI + A . A human specimen from Italy ( Human 5333_It ) and Brazil ( Human A4_Br ) each possessed the genotype I and III + A ( Table 4 ) . The cox1 assay identified infections caused by multiple helminth species , including members of the genus Strongyloides and various strongylids ( Fig 4 ) . Also note that we previously reported amplification of cox1 DNA from Ancylostoma sp . , as well as Metastrongylus sp . , possibly some free-living nematodes , and a rotifer ( Fig 4 ) [21] , using this assay . A cox1 amplicon from an Australian specimen ( Human 333_Au ) contained reads from S . fuelleborni and Necator americanus while another specimen ( Human 378_Au ) contained DNA from S . stercoralis and N . americanus . For other specimens , 18S data were generated for Strongyloides spp . , though a cox1 sequence was generated for a strongylid only . In one of these instances where Strongyloides 18S data was generated , a cox1 sequence was detected for N . americanus only ( Human 1_La ) . Similarly , for another specimen with Strongyloides 18S data ( Human 434_Au ) , two cox1 sequences putatively belonging to an Oesophogostomum sp . were obtained ( Fig 4 ) . For a third specimen ( Human 2_Ca ) , a sequence was generated for S . stercoralis 18S HVR-I and while a cox1 amplicon was generated containing DNA from N . americanus and S . stercoralis , the S . stercoralis reads were too few to generate a contig of sufficient quality .
The Strongyloides spp . genotyping assay described provides important advantages over previously described methods . For instance , we demonstrated its capacity to detect and differentiate DNA from different Strongyloides spp . genotypes and also multiple strongylids in DNA extracted directly from stool . We also show that the data generated using this assay remains compatible with data generated using earlier approaches based on Sanger sequencing , facilitating direct comparison of new data generated using this assay with published data . As this analysis represents more than 1 , 000 sequences collected from a diverse range of hosts and world locations , it provides a truly global understanding of the population genetic structure of the Strongyloides spp . infecting humans , non-human primates and domestic dogs .
|
Strongyloidiasis is a neglected tropical disease caused by the human infective worms ( nematodes ) Strongyloides stercoralis , Strongyloides fuelleborni fuelleborni and Strongyloides fuelleborni kellyi . Little is known about the genetic diversity of these nematodes and the possibility of geographically isolated genetic types is a particularly interesting research question , given the unique life cycle of these worms which includes both sexual and asexual stages . We extracted DNA from human , dog and primate feces containing Strongyloides , collected from several countries representing all inhabited continents . Using next-generation sequencing , we analyzed three Strongyloides DNA fragments from these specimens; two fragments of the 18S gene and one from the cytochrome c oxidase subunit 1 ( cox1 ) gene . Using this approach , we discovered some unique S . stercoralis and S . fuelleborni genotypes , and identified a previously unknown S . fuelleborni infecting humans within Australia . We compared our data to all 18S and cox1 sequences of S . fuelleborni and S . stercoralis available in public databases and identified several patterns relating to the global distribution of certain genotypes . This knowledge could allow us to infer the origin of human Strongyloides infections in the future , and assess the role certain animals ( non-human primates and dogs ) might play in the transmission of Strongyloides to humans .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"vertebrates",
"parasitic",
"diseases",
"dogs",
"animals",
"genetic",
"mapping",
"mammals",
"nematode",
"infections",
"primates",
"molecular",
"biology",
"techniques",
"strongyloides",
"stercoralis",
"genotyping",
"old",
"world",
"monkeys",
"research",
"and",
"analysis",
"methods",
"strongyloides",
"monkeys",
"molecular",
"biology",
"macaque",
"haplotypes",
"eukaryota",
"heredity",
"genetics",
"nematoda",
"biology",
"and",
"life",
"sciences",
"apes",
"chimpanzees",
"amniotes",
"organisms"
] |
2019
|
A global genotyping survey of Strongyloides stercoralis and Strongyloides fuelleborni using deep amplicon sequencing
|
Drosophila Little imaginal discs ( Lid ) is a recently described member of the JmjC domain class of histone demethylases that specifically targets trimethylated histone H3 lysine 4 ( H3K4me3 ) . To understand its biological function , we have utilized a series of Lid deletions and point mutations to assess the role that each domain plays in histone demethylation , in animal viability , and in cell growth mediated by the transcription factor dMyc . Strikingly , we find that lid mutants are rescued to adulthood by either wildtype or enzymatically inactive Lid expressed under the control of its endogenous promoter , demonstrating that Lid's demethylase activity is not essential for development . In contrast , ubiquitous expression of UAS-Lid transgenes lacking its JmjN , C-terminal PHD domain , and C5HC2 zinc finger were unable to rescue lid homozygous mutants , indicating that these domains carry out Lid's essential developmental functions . Although Lid-dependent demethylase activity is not essential , dynamic removal of H3K4me3 may still be an important component of development , as we have observed a genetic interaction between lid and another H3K4me3 demethylase , dKDM2 . We also show that Lid's essential C-terminal PHD finger binds specifically to di- and trimethylated H3K4 and that this activity is required for Lid to function in dMyc-induced cell growth . Taken together , our findings highlight the importance of Lid function in the regulated removal and recognition of H3K4me3 during development .
The Drosophila lid gene is essential for development and encodes a protein with multiple domains implicated in chromatin-mediated regulation of transcription , including the recently described lysine demethylase domain , Jumonji C ( JmjC ) . Six lysine residues of histones H3 and H4 can be mono , di or trimethylated , and each modification is found in a stereotypical pattern with respect to the coding region of a gene and correlates with a different transcriptional outcome [1]–[4] . As a general rule , methylation of H3K4 , K36 or K79 is found at active genes whereas H3K9 , K27 and H4K20 methylation is associated with those that are repressed . We and others have shown that overexpression of Lid reduces H3K4me3 levels and that this chromatin mark is elevated in lid mutants , establishing Lid as a JmjC domain-dependent H3K4me3 demethylase [5]–[7] . The four conserved mammalian orthologs of Lid , KDM5a-d , also demethylate H3K4me3 although these proteins show broader substrate specificity than their Drosophila counterpart , also targeting H3K4me2 [8]–[10] . While there is limited data regarding the biological role of the KDM5 family of proteins in mammals , the findings that KDM5b is overexpressed in breast , bladder and prostate cancers [11]–[13] and that mutations in KDM5c are found in patients with X-linked mental retardation [14] suggest that they play important developmental roles . However , a confounding factor to the analysis of the four mammalian KDM5 paralogs is their functional redundancy , as the mouse KDM5a knock out is viable , fertile and displays no change in global H3K4me2/3 levels . In contrast , Lid is the sole KDM5 protein in Drosophila and it is essential for viability [15] , providing an ideal system in which to investigate the function of this family of proteins . Although the KDM5 family of proteins are named based on the function of their catalytic JmjC domain , metazoan KDM5 proteins have several other conserved motifs: a JmjN domain of unknown function that is present in a subset of JmjC proteins , an ARID ( A/T rich interaction domain [16] ) implicated in binding both A/T and G/C rich DNA sequences [17] , [18] , a single C5HC2 zinc finger , and two or three PHD fingers ( plant homeobox domain [19] ) involved in mediating protein-protein interactions [20] . Importantly , while the JmjC domain-dependent demethylase function is well defined in vitro for KDM5 family proteins , the in vivo relevance of this and other domains remains unclear . We have previously shown that Lid is rate-limiting for cell growth induced by the Drosophila homolog of the c-Myc oncoprotein , dMyc [7] . Specifically , Lid binds directly to dMyc and is required for dMyc-dependent activation of one of its growth regulatory target genes , Nop60B . While we have demonstrated that this occurs independently of Lid's lysine demethylase activity , the molecular mechanism by which Lid functions in Myc-mediated growth is yet to be determined . Here we present an investigation of the function of Lid's domains and demonstrate that its demethylase activity is dispensable for development , however its JmjN , PHD3 and C5HC2 domains are all essential . While our observation that Lid's demethylase activity is not essential suggests that regulated removal of H3K4me3 serves primarily to modulate gene expression levels , a genetic interaction between lid and the JmjC domain-containing protein dKDM2 is consistent with these two demethylases acting redundantly on H3K4me3 . We also show that the essential C-terminal PHD finger of Lid binds di- and trimethylated H3K4 and that this domain is required for lid to genetically interact with dMyc . Based on these data , we propose that Lid-dependent recognition of H3K4me2/3 facilitates dMyc binding to promoters rich in this active chromatin mark .
To assess the contribution of each individual domain of Lid to its demethylase activity and animal development , we generated a series of deletions and point mutations that disrupt each domain of Lid to complement our previously characterized demethylase inactive version of Lid ( Lid-JmjC* ) that harbors two point mutations in the JmjC domain and prevents Fe2+ binding ( H637 and E639 to Alanine ) [7] . To enable these analyses , flies carrying UAS transgenes that specifically delete Lid's JmjN , ARID , C5HC2 zinc finger and three PHD fingers were generated to allow conditional Gal4-mediated expression in vivo ( see Materials and Methods for details ) . To assess the ability of our Lid mutants to demethylate , we generated clones of cells overexpressing each protein in larval fat body and examined the levels of Lid and H3K4me3 ( Figure 1 ) . Based on the intensity of the immunofluorescence signal and Western analysis , all of our transgenes expressed Lid at similar levels ( data not shown; Figure 1 ) , with the exception of LidΔPHD2 , for which we were unable to detect Lid overexpression even after combining multiple transgenes ( data not shown ) . To examine the role of Lid's second PHD finger , we created a point mutant in the first cysteine of this C4HC3 zinc finger ( UAS-LidC1296A ) and found that overexpression could be detected after combining two transgenes ( Figure 1K ) . Mutating the second or deleting the third PHD domain of Lid did not affect its ability to demethylate H3K4me3 ( Figure 1K–1N ) . In contrast , Lid's JmjN , PHD1 or C5HC2 domains were essential for enzymatic activity as overexpression of these deletion mutants resulted in no change in global levels of H3K4me3 ( Figure 1D , 1H , 1J ) . While the role of Lid's PHD1 and C5HC2 domains in demethylation remains to be investigated , our finding that Lid's JmjN domain is required for demethylase activity is not surprising based on structural analysis of the demethylase KDM4a which shows its JmjN domain making extensive contacts within the catalytic core of its immediately adjacent JmjC domain [21] . Unlike other deletions that prevented Lid's enzymatic function , expression of LidΔARID resulted in a variable increase in H3K4me3 levels , indicating that this mutant protein can behave as a dominant negative in fat body cells ( Figure 1E , 1F ) . We do not yet understand the mechanism by which LidΔARID increases H3K4me3 levels , but have observed a similar effect upon overexpression of Lid-JmjC* [7] . The ARID of KDM5a , b and Lid are required for demethylase activity in transient transfection assays , however a dominant interfering effect has not been reported [17] , [22] . Our finding that deletion of Lid's ARID can increase H3K4 trimethylation raises the possibility that in addition or concomitant with its ability to bind DNA , this domain may cooperate with Lid's JmjC domain . To determine the importance of Lid's conserved domains in vivo , we ubiquitously expressed our UAS-Lid transgenes in animals lacking endogenous zygotic lid expression . Because Lid is normally expressed ubiquitously throughout development [7] ( data not shown ) , we expressed our UAS-Lid ( and Lid mutant ) transgenes at low uniform levels in lid10424 homozygous mutants using actin-Gal4 ( Figure 2E ) . This approximately two-fold overexpression of Lid is not sufficient to cause any change to global levels of H3K4me3 in wing discs ( Figure 2E ) . As a control , we crossed our Lid transgenes to actin-Gal4 in a wildtype background to ensure that expression of these Lid mutants did not have any deleterious effects . In all cases , expression of our UAS transgenes in a wildtype background gave viable adults , however UAS-LidΔARID , Lid-JmjC* or LidΔPHD1 expressing adult females failed to lay eggs , so were sterile . Surprisingly , ovaries from females expressing these three mutant forms of Lid were phenotypically normal as assessed by dapi and phalloidin staining ( data not shown ) , so the basis for their dominant interference with oviposition is not clear . This effect on egg laying is likely to be due to expression these Lid mutant transgenes in somatic cells of the ovary since germline specific expression using Nanos-Gal4 does not result in sterility ( data not shown ) . As expected , actin-Gal4 driven expression of wildtype Lid rescued lid10424 mutant animals at the expected Mendelian frequency ( Table 1 ) . In contrast , expression of UAS-Lid harboring deletions of its JmjN , ARID , or PHD3 domains fail to rescue lid mutants , suggesting that these domains are essential for development . Expression of LidΔC5HC2 resulted in a small percentage ( 29% ) of lid mutant flies eclosing , all of which died within several days indicating that this domain is essential in adults . In contrast to the third PHD , we found that the first and second PHD fingers of Lid are dispensable for development . While both sexes rescued by LidC1296A were fertile , LidΔPHD1-rescued females were sterile and , like overexpression of this transgene in a wildtype background , LidΔPHD1-rescued flies had phenotypically normal ovaries . We also tested our previously generated JmjC domain point mutant that abolishes demethylase activity for rescue of lid-associated lethality . Actin-Gal4-mediated expression of Lid-JmjC* failed to rescue lid mutants ( Table 1 ) , initially suggesting that Lid's demethylase activity is essential for development . However , since overexpression of Lid-JmjC* behaves as a dominant negative in a tissue specific manner , most notably in larval fat body cells [7] , it may interfere with maternally deposited wildtype Lid in the rescue experiments described above . To address the function of Lid's demethylase activity during development , we therefore generated genomic rescue transgenes that fused 3 . 9 kb of Lid's upstream regulatory region to either a wildtype or JmjC* mutant form of the lid coding region ( gLid-WT and gLid-JmjC* respectively; Figure 2A ) . gLid-WT and gLid-JmjC* transgenes were then crossed into lid10424 and lidk6801 mutant backgrounds , the levels of transgene expression confirmed , and the number of homozygous lid mutant flies scored ( Figure 2B , 2C; data not shown ) . Strikingly , lid mutant animals carrying one or two copies of a gLid-WT or gLid-JmjC* transgene produced phenotypically normal and fertile adult flies at the predicted frequency ( Figure 2C; data not shown ) . Lid's demethylase activity is therefore not essential for Drosophila development . Based on the rescue of lid mutants by enzymatically inactive Lid expressed at endogenous levels , it is likely that this mutant form of Lid failed to rescue in our actin-Gal4 based rescue experiments because its overexpression interferes with maternally deposited wildtype Lid . It is therefore possible that LidΔARID also fails to rescue lid mutants due to its dominant interference with endogenous Lid , thus further examination of this domain will require generation of transgenes using lid's endogenous promoter . A majority of homozygous lid mutant animals die during pupal development and have increased global levels of H3K4me3 [5] , [7] , [22] , [23] . While Lid can remove di and trimethylated histone H3K4 peptides in vitro , we and others have shown that it only targets H3K4me3 in vivo as only this methyl mark is altered upon Lid overexpression , in lid mutants , and in response to Lid RNAi [5] , [7] , [22] . Because expression of the demethylase inactive form of Lid is able to rescue lid mutants , we asked whether these animals also have increased global levels of H3K4me3 . To examine this , we dissected wing discs from wildtype and lid10424 homozygous mutant larvae and compared the levels of H3K4me3 to lid mutants carrying two copies of gLid-WT or gLid-JmjC* by Western blot . As seen in Figure 2B and 2D , gLid-JmjC* animals show increased H3K4me3 indistinguishable from that observed in lid mutants , demonstrating that the increased level of H3K4me3 observed in lid mutant animals is not the cause of their lethality . Mutants and RNAi-mediated knock-down of the C . elegans Lid ortholog RBR-2 result in elevated levels of H3K4me3 and a 15–25% reduction in lifespan [24] . To determine whether this is a conserved , demethylase-specific phenotype , we assessed the lifespan of our demethylase inactive Lid flies . We find that lid mutant males rescued by gLid-JmjC* have a significantly shorter lifespan ( mean of 37 days ) than their wildtype ( 45 days ) or gLid-rescued ( 46 days ) flies ( Figure 3A; data not shown ) . Interestingly , this effect is not observed in females , with the average lifespan of gLid-JmjC*-rescued flies not being significantly different to wildtype ( 48 . 1 and 45 . 1 , respectively; Figure 3B ) . Lacking H3K4me3 demethylase activity therefore has adverse effects on processes required during male adulthood and suggests lifespan phenotypes observed in C . elegans hermaphrodites are likely to be specific to modulation of H3K4me3 levels rather than other RBR-2-dependent processes . Animals with reduced life expectancies also often show sensitivity to oxidative stressed induced by paraquat . We therefore treated wildtype and demethylase inactive flies with paraquat and found a sex-specific effect of this inducer of oxidative damage . In a similar manner to our lifespan studies in which males were more dramatically affected than females , we find that males are sensitive to paraquat whereas females are not ( Figure 3C , 3D ) . Male Drosophila are therefore more sensitive to the loss of Lid-dependent H3K4me3 demethylation than females , although the molecular basis for this remains unclear . One explanation for our finding that the loss of Lid's enzymatic activity does not adversely affect development is that its H3K4me3 demethylase activity is compensated for by another demethylase . To date , the JmjC domain-containing protein dKDM2 is the only other Drosophila protein shown to target H3K4me3 , although it has also been reported to remove H3K36me2 [25] , [26] . To address whether dKDM2 and Lid act in a redundant manner , we tested whether hypomorphic mutations in these two genes genetically interact . lidK6801 homozygotes survive until adulthood at a very low frequency ( 0 . 5% ) , but reach pupal development at 71% of the expected frequency ( Table 2 ) . The strongest dKDM2 allele , dKDM2DG18120 , is semi-lethal with homozygous adults eclosing at 62% of the expected frequency ( Table 2 ) and these adults are phenotypically normal and fertile . By combining these two mutations , we have found that the phenotype of lid , dKDM2 double mutants is significantly stronger than either single mutant ( Table 2 ) , with animals dying during the 1st and 2nd larval instar stages . To demonstrate that Lid's demethylase activity is required for this genetic interaction , we tested whether lid;dKDM2 double mutants could be rescued by our gLid-WT or gLid-JmjC* genomic rescue transgenes . As shown in Table 2 , gLid-WT , but not gLid-JmjC* rescued lid;dKDM2 animals , suggesting that Lid and dKDM2 act redundantly in the regulation of H3K4me3 . We originally isolated lid in a genetic screen for regulators and mediators of dMyc-dependent cell growth based on an adult eye phenotype generated by dMyc expression in post-mitotic cells of the developing eye using GMR-Gal4 [7] . Furthermore , we showed that Lid's demethylase activity was not required for its dMyc-dependent functions . To pursue the mechanism by which Lid functions in cell growth induced by dMyc , we crossed our UAS-Lid mutant transgenes to the dMyc overexpressing fly strain and compared their ability to enhance this eye phenotype to that observed in response to expression of wildtype Lid ( Figure 4A ) . Expression of Lid lacking its JmjN , C5HC2 or PHD3 domains failed to enhance the dMyc overexpression eye phenotype while not altering the levels of overexpressed dMyc ( Figure 4B–4F; data not shown ) . As controls , we expressed the Lid deletion transgenes in a wildtype background and found that they resulted in no adult eye phenotype and , unlike fat body cells , Lid-JmjC* and LidΔARID do not have a dominant negative effects in post mitotic cells of the developing eye . We have previously shown that dMyc binds to two regions of Lid: its JmjC domain and its C5HC2 zinc finger [7] . To verify that all Lid deletion proteins retain their ability to bind dMyc , we carried out in vitro binding assays and found that they all bind equivalently ( Figure 4G ) , suggesting that the JmjN , C5HC2 and PHD3 domains of Lid are likely to be required for its Myc-dependent functions in cell growth . The primary characterized function of Lid is its histone H3 lysine 4 demethylase activity . However , since we have demonstrated that this activity is not Lid's essential function , we chose to further characterize Lid's third PHD finger as this domain is required for it to function with dMyc and is essential for development . Moreover , PHD domains have recently emerged as important interpreters the histone code that act by binding to histone tails that are unmodified , mono- , di- or tri- methylated at specific lysine residues [20] . To address whether Lid's third PHD finger is able to bind methylated histones , we incubated bacterially expressed and purified GST-PHD finger proteins with biotinylated histone peptides mono- , di- or trimethylated at K4 , K9 or K27 in vitro and compared this to the binding of Lid's other two PHD fingers and the known H3K4me2/3 binding protein hING2 [27] , [28] ( Figure 5; data not shown ) . As seen in Figure 5 , Lid's PHD1 finger binds to amino acids 1–21 of histone H3 , but not amino acids 21–40 or to histone H4 . Lid's PHD1 specifically recognizes unmethylated histone H3 ( H3K4me0 ) , as binding is abrogated by mono- , di- or trimethylation of lysine 4 , but not methylation of lysine 9 . We were unable to detect any in vitro histone binding for Lid's PHD2 finger , however PHD3 bound to both H3K4me2 and H3K4me3 , showing a consistent preference for the trimethylated form . The function of both of these PHD fingers is likely to be a highly conserved function of KDM5 proteins , as identical binding specificities have recently been reported for KDM5a [29] . Consistent with the binding of PHD3 to H3K4me2/3 being physiologically relevant , a correlation between KDM5a binding and the presence of this activating chromatin mark has been observed previously using genome-wide arrays , although its physiological relevance has remained elusive [30] , [31] . Significantly , the binding of c-Myc also correlates with regions rich in H3K4me2/3 [32] . Based on our findings that Lid's third PHD finger binds H3K4me2/3 and that deleting this domain abolishes its ability to genetically interact with dMyc , we propose that Lid functions to recruit dMyc to regions with high levels of H3K4me2/3 by specifically recognizing this local chromatin context .
Our finding that Lid's lysine demethylase activity is dispensable for development demonstrates that globally increasing the levels of H3K4me3 is not generally detrimental to development . Similarly , elevating H3K4me1/2 levels by mutating the Drosophila demethylase Lsd1 does not adversely affect development , although these animals show some adult phenotypes and subtle changes to expression of the homeobox genes Ubx and Abd-A [33] , [34] . Likewise , Lid's enzymatic activity may serve to fine-tune some gene expression patterns . To date , three genes , E ( spl ) m4 , m7 and m8 , have been described as direct Lid targets in Drosophila cultured S2 cells , and these show a 4-fold derepression in response to lid RNAi and a concomitant increase in promoter-proximal H3K4me3 levels [35] . Furthermore , a genetic interaction has been observed during wing development between lid and the E ( spl ) gene upstream regulator Notch , suggesting that this regulation is biologically important [35] , [36] . In mammalian cells , MFN2 and Deltex expression are repressed upon KDM5a overexpresion , derepressed when KDM5 is knocked-down , and show changes in H3K4me3 levels in their promoters [31] , [36] . We examined the levels of E ( spl ) m4 , m7 and m8 , Marf1 ( the Drosophila ortholog of MFN2 ) and Deltex , but found that their levels were unaltered in RNA extracts from whole larvae or dissected wing imaginal discs from wildtype , lid mutant or lid mutants rescued by gLid or gLid-JmjC* ( JS , unpublished ) . These genes may therefore be regulated by Lid in a small subset of cells in vivo , so cannot be detected using whole wing disc extracts . Effects on gene expression may also be sex-specific since male flies lacking Lid-dependent demethylase activity have a shortened lifespan and are sensitive to paraquat , whereas females are not . While removing Lid's demethylase activity does not result in lethality , removing this function in combination with another JmjC domain-containing protein , dKDM2 , does . This suggests that in the absence of Lid's demethylase activity , dKDM2 can carry out its essential functions and vice versa . RNAi-mediated knock down of dKDM2 has been found to increase H3K36me2 levels in S2 Drosophila tissue culture cells and H3K4me3 levels in adult flies [25] , [26] . Surprisingly , we find that global levels of H3K4me3 and H3K36me2 are both unchanged in dKDM2DG12810 , dKDM2KG04325 or dKDM2EY01336 homozygous mutant wing discs ( CG and JS , unpublished ) . The reason for the disparity between our results obtained with dKDM2 mutants and previously published data are not clear , but may be due to the difference between the acute loss of dKDM2 mediated by RNAi and the chronic loss in dKDM2 mutants , or to off target effects of the RNAi . The most characterized function of dKDM2 and its mammalian orthologs ( KDM2A , KDM2B ) is its regulation of rRNA expression [25] , [37] . Interestingly , repression of rRNA transcription by KDM5A correlates with changes to H3K63me2 levels , whereas H3K4me3 is unaltered [37] . Based on our genetic interaction between lid and dKDM2 , this may be because Lid/KDM5a compensates for the loss of dKDM2's H3K4me3 demethylase activity . Conversely , it is likely that dKDM2 also functions outside the nucleolus and that H3K4me3 regulation by Lid and dKDM2 is essential for development . It is important to note , however , that while Lid's demethylase activity is required for the genetic interaction between lid and dKDM2 , we cannot rule out the possibility that dKDM2 requires its H3K36me2 demethylase enzymatic activity not its H3K4me3 activity . Both H3K4me3 and H3K36me2 are chromatin marks associated with active transcription , and it is possible that Lid's H3K4me3 demethylase activity is functionally linked to dKDM2-mediated H3K36me2 demethylation . Among the JmjC domain-containing proteins , Lid is most structurally similar to the founding member of this class of demethylases , Jumonji ( JARID2 ) , having a JmjN , ARID and C5HC2 zinc finger in addition to a JmjC domain . In both mammals and Drosophila , Jumonji is enzymatically inactive because it lacks key residues within its JmjC domain required for Fe2+ and α-ketoglutarate binding [38] . Indeed , while Jumonji has been implicated as a regulator of transcription [30] , [38]–[40] , the molecular function of the JmjC domain has remained elusive . Taken in conjunction with our finding that Lid's demethylase activity is not essential for development , this raises the exciting possibility that the JmjC domain has important demethylase-independent functions . Consistent with this hypothesis , we find that a genomic rescue transgene with a deletion of the JmjC domain fails to rescue lid mutants ( CG and JS , unpublished ) . Because more than half the known JmjC domain-containing proteins in mammals and Drosophila do not have an ascribed enzymatic activity , a demethylase-independent functions of this domain may be a common feature of this class of protein . Lid has three PHD fingers and we have demonstrated that its N- and C-terminal PHDs bind specific methylated forms of the histone H3 tail . While Lid's N-terminal H3K4me0-binding PHD finger was not required for development , its third PHD finger , which binds to H3K4me2/3 , is essential for viability and is required for Lid to function in dMyc-mediated cell growth . One long-standing question regarding many transcription factors is the mechanism by which they find their appropriate binding site within the genome , as many transcription factors recognize short DNA sequences that are similar or identical . This suggests that binding site specificity may additionally involve the recognition of non-DNA elements such as local chromatin environments . In mammalian cells , c-Myc shows a clear binding preference for E boxes located within a chromatin context containing highly di- and trimethylated nucleosomal histone H3K4 [32] . However , the mechanism by which Myc recognizes this chromatin landscape is unclear . We propose that Lid utilizes its H3K4me2/3 binding C-terminal PHD finger to tether Myc to its preferred chromatin context , thereby permitting selection of biologically important E boxes . Further experiments to more precisely define the role of Lid's PHD finger in Myc-mediated cell growth are ongoing . In summary , we have demonstrated that Lid's JmjC domain-encoded demethylase activity , its histone H3K4me0-binding N-terminal PHD finger and its PHD2 of unknown function , are dispensable for development . In contrast , all other domains of Lid tested were required to rescue lid homozygous mutants , including its C-terminal , H3K4me2/3 binding , PHD finger that functions in dMyc-mediated cell growth . These findings highlight the importance of characterizing the function of individual domains of transcriptional regulators such as Lid in order to understand the mechanisms by which they regulate gene expression in a developmental context .
UAS-lid and UAS-lidJmjC* have been described previously [7] . All other Drosophila strains were obtained from the Bloomington stock center . Deletions within Lid were made in the pUASp vector by site directed mutagenesis and delete the following amino acids: LidΔJmjN ( AA160–206 ) , LidΔARID ( AA223–314 ) , LidΔPHD1 ( AA450–499 ) , LidΔC5HC2 ( AA830–883 ) , LidΔPHD2 ( AA1296–1354 ) , LidΔPHD3 ( 1749–1838 by introducing a stop codon ) . LidC1296A mutates the first cysteine of Lid's second PHD finger . lid genomic rescue transgenes were generated by fusing a 4 . 5 kb PCR-generated Xho I fragment containing the lid upstream region and a 4 . 8 kb Xho I/Not I fragment containing the remainder of the lid coding sequence ( either wildtype or JmjC* ) into the vector pCasper4 . All transgenic flies were generated by The Best Gene ( thebestgene . com ) . Lifespan studies were carried out as described by [41] . To test the ability of UAS-Lid ( wildtype and deletion ) transgenes to rescue the lid mutant phenotype , a UAS-Lid ( or deletion ) transgene was recombined onto the lid10424 chromosome . At least 2 independent P element insertions were tested for each to minimize chromosomal position effects . This lid10424 , UAS-Lid ( or deletion ) recombinant chromosome , balanced over CyO , was then mated to the lid10424/CyO; Actin-Gal4/TM6B strain . Rescue was assessed by scoring the presence of straight winged , non-TM6B , progeny . Somatic clones overexpressing UAS transgenes marked by the co-expression of GFP were generated as described in [42] . Longevity studies were carried out as described in [41] and paraquat assays as described in [43] . Histone binding assays: 1 µg of biotinylated histone peptides ( Fisher ) were incubated with 5 µg purified GST-PHD finger in 1 ml of binding buffer ( 50 mM Tris pH 7 . 5 , 200 mM NaCl , 2 mM dithiothreitol , 0 . 5% Nonidet P-40 ( v/v ) , 1 µM ZnSO4 , 1% BSA ) at 4°C overnight . Complexes were then immobilized using 10 µl Streptavidin-agarose beads ( Invitrogen ) for 1 hr at 4°C . Immobilized complexes were then washed three times with 1 ml of binding buffer , boiled and loaded on a 4–12% gel . Gels were stained with coomassie blue to visualize bound GST-PHD protein . GST-protein binding assays: 1 µg of purified GST-dMycC [7] was incubated with S35-labeled Lid or Lid deletion proteins made using rabbit reticulocyte lysate ( Invitrogen ) in 1xPBS , 1% BSA and 0 . 5%NP-40 , washed in 1xPBS , 0 . 5% NP-40 , boiled and loaded onto a 4–12% protein gel . GST-dMycC was visualized using coomassie brilliant blue and S35 detected via standard procedures . The Lid rabbit and dMyc antibodies have been described previously [7] , [44] . Anti-trimethylated H3K4 and H3K36me2 were obtained from Active Motif , and γ-tubulin from Sigma . Western analysis was carried out using standard protocols , infrared conjugated secondary antibodies ( LiCOR ) and Odyssey scanner and software . Immunofluorescence was carried out as described in [7] . Quantitation of Western blots was carried out using LiCOR odyssey v3 . 0 software .
|
Correct spatial and temporal control of gene expression is essential for development . One of the many ways that gene expression is regulated is by the addition , recognition , and removal of methyl groups from the histone proteins around which DNA is wrapped within the nucleus . Here we describe a systematic analysis of Little imaginal discs ( Lid ) , a protein that regulates transcription via a number of different mechanisms that involve regulated removal and recognition of histone methylation . We show that while Lid's histone demethylase activity is not essential for development , numerous other conserved domains of this protein are . Furthermore , we find a genetic interaction between lid and another histone demethylase , dKDM2 , that suggests this enzyme can compensate for the loss of Lid's enzymatic activity . These findings have significance for our insight into how gene expression is normally regulated and have implications for our understanding of how this goes awry during disease progression .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/histone",
"modification",
"molecular",
"biology/chromatin",
"structure",
"cell",
"biology/cell",
"growth",
"and",
"division",
"biochemistry/transcription",
"and",
"translation",
"cell",
"biology/gene",
"expression"
] |
2010
|
Essential Functions of the Histone Demethylase Lid
|
Promoters process signals through recruitment of transcription factors and RNA polymerase , and dynamic changes in promoter activity constitute a major noise source in gene expression . However , it is barely understood how complex promoter architectures determine key features of promoter dynamics . Here , we employ prototypical promoters of yeast ribosomal protein genes as well as simplified versions thereof to analyze the relations among promoter design , complexity , and function . These promoters combine the action of a general regulatory factor with that of specific transcription factors , a common motif of many eukaryotic promoters . By comprehensively analyzing stationary and dynamic promoter properties , this model-based approach enables us to pinpoint the structural characteristics underlying the observed behavior . Functional tradeoffs impose constraints on the promoter architecture of ribosomal protein genes . We find that a stable scaffold in the natural design results in low transcriptional noise and strong co-regulation of target genes in the presence of gene silencing . This configuration also exhibits superior shut-off properties , and it can serve as a tunable switch in living cells . Model validation with independent experimental data suggests that the models are sufficiently realistic . When combined , our results offer a mechanistic explanation for why specific factors are associated with low protein noise in vivo . Many of these findings hold for a broad range of model parameters and likely apply to other eukaryotic promoters of similar structure .
Combinatorial regulation of gene expression is an important mechanism for signal integration in prokaryotes and eukaryotes ( reviewed in [1] ) . Typically , specific motifs in the DNA sequence favor binding of particular transcription factors ( TFs ) and thus encode a cis-regulatory input function [2] . Protein-protein interactions among different TFs , which do not necessarily involve direct contacts with DNA , contribute to—frequently synergistic—regulatory function [1] . This is a very versatile mechanism for hierarchical control , e . g . , when TFs can only be recruited in a pre-defined sequence or when they are excluded under specific conditions [3] . Chromatin state and chromatin-modifying activities provide yet another layer of regulation , and recruitment of the latter is typically also mediated by TFs [4] . Hence , multiple , complex levels of combinatorial control characterize transcriptional regulation [5] . New high-throughput measurement methods have generated a wealth of information on transcriptional regulatory circuits at different levels such as chromatin states , promoter occupancy by TFs , and mRNA expression dynamics as the system's output . Analysis of combinatorial regulation at the genome-scale points to a modular organization of transcriptional regulatory networks , which could facilitate data integration . However , this requires a multi-level analysis [6] and dynamic processes may lead to large functional re-arrangements of transcriptional regulatory networks . Concomitantly , understanding network design principles needs a detailed investigation of dynamic processes [7] , [8] . Corresponding computational models aid in disentangling transcriptional network structures and in quantitatively analyzing the impact of promoter architecture on the regulatory outcome . Depending on network size , available experimental data , and model purpose , model types range from qualitative logical models to quantitative approaches based on thermodynamic considerations or ordinary differential equations ( ODEs ) ( reviewed in [9]–[11] ) . However , most previous work focused on stationary gene–regulatory input functions in real-life organisms and in rational promoter design [2] , [12]–[14] . Recently , stochastic kinetic models have received increased attention because we lack a deeper understanding of how gene network architecture shapes gene expression noise [15] . Stochasticity in gene expression arises from environmental effects and from intrinsic sources . It can have benefits and adverse effects for gene network function ( reviewed in [16] , [17] ) . Hence , noise in gene expression may be an evolvable trait that is intimately linked to promoter architecture [16] . For eukaryotic systems , irregular promoter activation due to chromatin modifications or transcriptional re-initiation are the main intrinsic noise sources [15] , [18] . Despite recent progress [15] , [18] , [19] , our understanding of how the dynamic interplay of transcription factors , chromosomal positioning , epigenetic control , and cis-regulatory promoter elements shapes expression dynamics and noise properties is still limited . Moreover , much of our knowledge derives from artificial expression systems and unnatural stimuli , and we need more studies in complex natural systems to reliably assess the impact of stochasticity on diseases and developmental pathways [16] . For this , integrated approaches have to consider potential tradeoffs between optimal noise properties , phenotypic fitness , high mRNA productivity , and robustness to perturbations [15] , [20] . Budding yeast ribosome biogenesis can be employed for such an integrated analysis because the system is quantitatively well-characterized , complex , and crucial for cell physiology . It needs to operate efficiently and reliably; for instance , ribosome biogenesis requires coordinated expression of several hundred genes and accounts for up to 80% of transcriptional activity during rapid growth [21] , [22] . Tight and coordinated transcriptional control therefore appears critical , and promoter architecture plays a key role in integrating TF interactions and different layers of control . The system employs the Forkhead ( FH ) -type TF Fhl1 ( accession number S000006308; all accession IDs refer to the Saccharomyces Genome Database ( SGD ) available at http://yeastgenome . org unless mentioned otherwise ) , which belongs to a family of transcriptional regulators with more than 100 members conserved from yeast to human . These regulators typically serve as converging points for signaling pathways , they possess variable activation/repression domains , and often act in concert with coactivators/corepressors and general regulatory factors ( GRFs ) ( for reviews cf . [23] , [24] ) . Moreover , binding of the GRF Rap1 ( SGD S000005160 ) and of Fhl1 in yeast highly correlates with low protein noise [25] and transcriptional co-regulation is particularly strong for RP genes [26] , [27] , but the causes for both are unclear . These features make the control of ribosome biogenesis an ideal example system to address three general questions: Do complex promoters provide advantages over alternative , simpler designs ? Why are complex designs frequently employed when reliable regulation is critical ? Is the complex promoter architecture especially suited to provide low variations in mRNA levels ? Here , we address these questions by developing , analyzing , and validating a set of dynamic mathematical models of the promoter of yeast ribosomal protein ( RP ) genes . The set includes the in vivo design and three functionally related , but progressively simpler synthetic architectures . We integrate selected information from large-scale studies and from targeted experiments to provide the necessary quantitative basis for these models , and to comprehensively characterize the stationary and dynamic regulation of promoter activities using deterministic and stochastic simulations . This enables us to pinpoint structural features underlying the observed behavior and to identify functional tradeoffs that impose constraints on promoter architecture .
To develop kinetic promoter models , we start from elementary interactions between transcription factors and DNA . Typical RP gene promoters contain paired binding sites for the GRF Rap1 [27] ( cf . Figure 1A and 1B ) . Rap1 binds DNA directly [28] , which is required for efficient expression of RP genes and maintains promoter regions essentially nucleosome-free [3] , [29] . This complex alone can recruit RNA polymerase II for basal transcriptional activity [27] , [30] . Recruitment of the FH-type transcription factor Fhl1 , which in turn binds Ifh1 ( SGD S000004213 ) via its Forkhead-associated ( FHA ) domain , leads to full activation [27] , [31] , [32] . When Fhl1 is bound in the absence of Ifh1 , even basal transcription is suppressed [33] . Upstream signaling pathways can convey nutrient status to Ifh1 such that it rapidly dissociates from the promoter . This leads to a substantial reduction of RP transcription , whereas occupation by Rap1 and Fhl1 remains unchanged [27] , [32] . High-confidence datasets show that this RP gene promoter architecture is very generic in yeast [31] , [32] . Although additional regulators can contribute to RP gene regulation , their effects are probably indirect , strain-specific , or they affect RP gene expression in the same qualitative fashion as Ifh1 [27] , [29] , [34] , [35] . Hence , the interactions between GRF , Fhl1 , Ifh1 , and RP gene promoters capture the key aspects of transcriptional control of RP genes . Model 1 represents the wild-type scenario as follows ( Figure 1B ) : Sequential recruitment of two Rap1 molecules leads to basal transcriptional activity . Subsequent Fhl1 binding in the absence of Ifh1 quenches basal transcription , while Fhl1-dependent recruitment of Ifh1 induces full activity . Since regulation of Ifh1 binding to the promoter critically determines promoter activity [31] , [32] we simulate regulation upstream of Ifh1 by varying the amount available for promoter binding , i . e . , the effective Ifh1 concentration ( see Protocol S1 for details ) . Since most physiological stimuli appear to regulate Ifh1 binding while Rap1 and Fhl1 serve as scaffold , it is unclear if the seemingly complex architecture of the natural RP gene promoter yields any functional advantage . In principle , one could envision the same coordinated regulation by controlling the activity , localization , or DNA binding affinity of a single TF such as Ifh1 . Note that more complex promoters in terms of combinatorial control exist even in yeast [36] . However , RP genes are special because they form an exceptionally tight cluster of coregulated genes in transcriptome studies [26] , [27] . To investigate differences in function and regulatory performance of structurally related , but simpler architectures , we developed three alternative promoter models ( Models 2–4 ) . They are progressive simplifications of the natural promoter configuration ( Figure 1B; see Protocol S1 for details ) . Models 2 and 3 follow the same logic of sequential TF recruitment as the wild-type model . In Model 2 , a second Ifh1 molecule replaces Fhl1 and transcription ceases when only a single Ifh1 is bound ( Figure 1B ) . By contrast , recruiting one Ifh1 molecule suffices for full activation in Model 3 and Model 4 . Compared to the wild type ( Model 1 ) , Model 2 is a biologically more parsimonious solution with only two different proteins , but it maintains the same kinetic order as Model 1 . Model 3 , in addition , has a reduced kinetic order . Finally , Model 4 is the structurally simplest promoter variant that can transmit environmental inputs to a target gene . Notably , it does not employ GRFs . Although the simplified models are synthetic , they correspond to promoter architectures encountered in vivo . Model 2 with its cooperative activation by homodimeric TFs resembles regulation by cI repressor in phage λ [37] . Certain Ternary Complex Factor-type promoters are structurally similar to Model 3 [38] . Promoter architectures with single TFs as in Model 4 are well-described in yeast , e . g . , involving the TF Gcn4 ( SGD S000000735 ) in amino acid biosynthesis [36] . Importantly , for the molecular species denoted as Ifh1 in simplified Models 2–4 we assume functional equivalence , but not structural identity to Ifh1 , which itself cannot bind to DNA [29] . Stochastic binding and dissociation events of TFs and of RNA polymerase determine whether a given RP gene is transcribed . We represented control events by sets of elementary chemical reactions and mass-action kinetics [39] , [40] with or without including gene silencing due to changes in chromatin structures ( see Material and Methods and below ) . In this modeling framework , derivation of promoter kinetics for both the deterministic regime ( based on ODEs ) and for the stochastic setting is straightforward [41] . To analyze mean promoter activities , or other average properties , we used a deterministic description and verified its qualitative consistency with stochastic simulations for selected models and parameter settings ( data not shown ) . Some simulations were performed without considering gene silencing , both to separate its effects from those of the promoter configuration alone and because an equivalent stationary behavior could have been achieved in its presence by adapting the binding constants ( see Materials and Methods and Protocol S1 for details ) . To address how upstream signaling pathways—through variation in Ifh1 levels—modulate RP gene transcription , and how this is influenced by the ambivalent coactivator/repressor Fhl1 , we compared model predictions of stationary promoter activity without chromatin remodeling . For realistic parameter values , promoter activities are very similar for all models ( Figure 2A ) because , in the more complex models 1–3 , most genes are occupied by Rap1 dimers and thus available for Ifh1 binding . Notably , the in vivo configuration ( Model 1 ) does neither provide the highest stationary activity , nor the steepest or the most graded response of all model variants . Hence , the stationary input-output characteristics with respect to Ifh1 alone do not explain the complexity of the in vivo architecture . Next , we focused on gene inactivation because rapid down-regulation of ribosome synthesis is important for cellular growth when nutrients become scarce . In this case , Ifh1 leaves the promoter and RP synthesis effectively ceases , whereas environmental conditions barely affect Fhl1 and Rap1 binding [3] , [42] , [43] . We emulated adverse environmental conditions by complete absence of Ifh1 . Only the simplest model without GRF ( design 4 ) enables a complete shut-off ( Figure 2B ) . All other configurations retain a basal activity due to RP gene complexes with two Rap1 molecules . For realistic values of basal promoter activity ( η ) , Fhl1 binds the majority of Rap12-RP gene complexes in design 1 and thereby efficiently quenches basal transcriptional activity when Ifh1 is absent ( Figure 2C and 2D ) . In models 2 and 3 , transcription could only be lowered by an inefficient 5–10-fold reduction of cellular Rap1 levels . While the qualitative model behavior results from the way Fhl1 and basal activation by Rap1 are represented , we need such realistically parametrized mathematical models to assess these control effects quantitatively . Thus , we suggest that the ambivalent coactivator/corepressor ( Fhl1 ) enables a rapid switch between full and low basal activity without invoking inefficient control by GRFs . This may apply to similar promoters with dual coactivator/repressor TFs constitutively bound GRFs other than Rap1 [24] , [38] . The analysis of model 4 demonstrates that a single-input promoter with efficient shut-off can be realized with a single transcription factor and without basal activity conferred by the GRF . We therefore analyzed the combined effects of Fhl1 and Ifh1 on promoter activity . By varying the Fhl1 concentration it is not only possible to adjust the degree of activation in the presence of Ifh1 and the degree of repression in its absence , but also the factor fold-change between the two states ( Figure 2C ) . In other words , independent regulation of Fhl1 and Ifh1 provides an ON–OFF switch with basal activity and tunable upper and lower activity bounds . Predicting this behavior requires quantitative knowledge on protein levels and kinetic constants since , for example , decreasing the affinity of Ifh1 by 100-fold renders Fhl1 predominantly a repressor at low Ifh1 levels ( Figure 2D ) . Promoter activity is sensitive to changes in Ifh1 over a wider concentration range compared to Fhl1; especially at wild-type Fhl1 levels , Ifh1 can robustly modulate the activity plateau ( Figure 2C ) . By contrast , Fhl1 determines sensitivity of promoter activity to Ifh1: low effective Fhl1 concentrations limit the maximum promoter activity and make the promoter unresponsive to Ifh1 changes . Hence , both Fhl1 and Ifh1 can serve as input signals for tuning the switch . These generic predictions are supported by experimental evidence that Fhl1 and Ifh1 respond to different regulatory inputs [36] , [44] , [45] . In addition , the models predict that effective regulator concentrations need to be considerably lower than total in vivo protein levels to establish a tunable switch ( Figure 2A and 2C ) . This agrees with reports of large changes in nuclear Ifh1 and Fhl1 concentrations [34] and with estimates that much of Ifh1 is unavailable for promoter binding in vivo ( [45] , J . Merwin and D . Shore , personal communication ) . Full exploitation of the complex promoter architecture's regulatory potential , hence , requires regulatory mechanisms that target both inputs individually . This suggests novel regulatory motifs in the control of yeast ribosome biogenesis . Model predictions may depend on the choice of binding affinities between TFs and DNA as well as between the TFs themselves . Naturally , the question arises to what extent the relative model performance can be generalized . Optimizing each model's parameters separately over a broad parameter range demonstrates that the relative performance of promoter variants regarding maximum activity and shut-off properties remains unchanged ( cf . Figure S5 and Protocol S1 for details ) . However , the use of such ‘optimal’ parameter sets can be problematic because the evolutionarily relevant objective function is unknown . As a complementary approach , we employed robustness analysis based on the natural promoter structure and choice of TFs because they represent the known outcomes of evolution . More specifically , we quantified the robustness of model predictions by assaying the sensitivity of achievable promoter activity to random perturbations in TF binding constants ( see Materials and Methods ) . This informs us to what extent a specific prediction depends on a particular choice of system parameters . Insensitivity to parameter variations justifies generalizations , especially because robustness to random perturbations is an important characteristic of functional biological networks [46] , [47] . Figure 3A shows the promoter activity for Model 1 as a function of binding affinities for Rap1 ( 1st step ) , Fhl1 , and Ifh1 . The pronounced vertical stratification demonstrates that strong Ifh1 binding is essential for high promoter activity . The affinity of Fhl1 has a less marked effect and the attainable promoter activity barely depends on the strength of the first and second Rap1 binding steps ( Figure 3C , Figure S1 , and Figure S2 ) . These features apply to all models ( data not shown ) and they are , thus , rather independent of the actual promoter configuration . The distribution of binding affinities associated with high promoter activity in Model 1 confirms that early binding steps are less sensitive to changes in binding affinities than later ones ( Figure 3C ) . Maximum sensitivity in a sequence of cooperative binding steps is known to require high association constants immediately before RNA polymerase binds [13] , [37] . Our analysis generalizes this result to promoters that have intermediate states with basal activity when realistic concentrations and binding constants are considered . However , the more complex designs were less robust when we mutated all binding affinities simultaneously . Only 4% of the mutated promoters showed high stationary activity for Model 1 , as opposed to 6% for Model 2 , 17% for Model 3 , and 22% for the simplest Model 4 ( Figure 3B ) . Ifh1 and Rap1 should contribute similarly to sensitivity and insensitivity in designs 1–3 . The major difference in robustness must , hence , be conferred by the additional intermediate state without transcriptional activity in models 1 and 2 ( see Figure 1 , e . g . , transcriptionally inactive , single-bound Ifh1 in model 2 ) . Apparently , the decreased robustness in Model 1 is the trade-off for added functionality . The ubiquitous presence of FH-type regulation and its usage at critical control points suggest that additional flexibility outweighs potential effects of reduced robustness . This holds for a broad , physiologically plausible range of transcription factor affinities . It will , therefore , generalize to other structurally related promoters that employ combined coactivator/corepressor TFs . Efficient regulation crucially depends on the ability to consistently respond to changes in the input ( s ) . Next , we therefore investigated the dynamic promoter responses to varying external conditions . More specifically , we investigated the dynamic promoter performance with and without gene inactivation due to chromatin modifications , which may lower the concentration of accessible genes at any given time point . For the steady-state analysis above , silencing could be mimicked by decreasing the affinity of TFs for DNA , but this does not hold for the dynamic behavior . Specifically , we mimicked environmental changes by applying a sinusoidal time-varying input of free Ifh1 with fixed amplitude and frequency . Such a periodic forcing function is the standard choice in frequency response analysis [48] because the system is stimulated by a single frequency , and not by a frequency spectrum as for other input shapes . This allows us to map output behavior to a unique input frequency . For linear models , input and output frequencies match and only a phase shift occurs , while nonlinear models produce a spectrum of output frequencies . By varying the input frequency , we can mimic noise effects ( high frequencies ) and observe how well the system tracks dynamic inputs ( lower frequencies ) . Figure 4A shows how the scaled amplitude of promoter activity oscillations and the frequency-dependent average promoter activity ( Figure 4B ) vary for oscillation periods between 1 s ( f = 1 Hz ) and approximately 27 hours ( f = 10−5 Hz ) when gene silencing is neglected . Along with the phase shift between input and output activity , this representation is related to the well-known Bode plot for linear systems in control theory . It combines the ratio of output and input amplitudes in a double logarithmic plot and the phase shift between output and input in a semi-logarithmic plot as function of the input frequency . As our models are nonlinear , the predicted promoter activities deviate from the sinusoidal shape of the input and show a more switch-like behavior , but the predominant frequency contribution to the output was always identical to the input frequency ( see Figure S3 ) . Shape modulation causes differences between the average promoter activities for dynamic and constant inputs , namely lower/higher activities for slow/fast Ifh1 oscillations , respectively ( Figure 4B ) . In all designs , promoter activity follows slow input signals quantitatively and closely for periods larger than 15 minutes ( f<10−3 Hz ) , while it rejects fast Ifh1 oscillations with a period below 15 min ( f>10−3 Hz ) for the chosen parameter settings . The frequency response of activity oscillations and the phase shift ( Figure S4A ) are characteristic of a first-order-type low-pass behavior , which enables faithful transmission of low-frequency signals and rejection of high-frequency noise in engineered and biological systems [11] , [48] , [49] . To analyze the impact of random chromatin modifications on promoter dynamics , we assumed a reversible and constitutive process that maintains a compact chromatin state ( assembled nucleosomes ) in the absence of TF binding ( see Protocol S1 ) . With a single TF ( Model 4 ) , dynamic gene inactivation substantially decreases the average promoter activity , alters the phase response , and suppresses activity fluctuations ( cf . Figure 4C and 4D and Figure S4C and S4D ) . The latter leads to a desirable noise filtering at high frequencies , but it also prevents faithful input tracking in the physiologically relevant frequency range . A stronger TF binding affinity can compensate for the low average activity ( data not shown ) . However , faster association rates may meet physico-chemical limitations [50] , while slower dissociation will increase the response time to input signals . In contrast , chromatin closure has almost no effect on the dynamics of Rap1-containing promoter architectures ( Models 1–3 ) , apart from a slight reduction in average promoter activity . Similarly , promoter activity in models 1–3 resists noise even for large , physiologically plausible fluctuations in Rap1 concentrations regardless of chromatin compaction ( Figure S4B and S4D and data not shown ) . Importantly , this superiority of Rap1-containing architectures is not restricted to a specific choice of parameters—and , hence , TF-binding site affinities—nor to a specific stimulus shape: We obtained the same qualitative behavior when optimizing the parameters of each model separately for the response to step changes in Ifh1 within a range of realistic kinetic constants and TF affinities ( Figure S5 and data not shown; see Protocol S1 for details ) . Moreover , optimal parameter sets obtained in independent optimization runs showed parameter variability in agreement with the above robustness analysis ( see Protocol S1 ) . Altogether , stably bound dimeric GRFs , in general , can protect the promoter from noise propagation due to unspecific chromatin modifications . GRFs ensure that the promoter remains in a poised state for rapid reactivation even after prolonged absence of TF binding . This obviates the need to first reactivate the genes in a sequence of—potentially slow—chromatin modification steps [15] , [16] , [18] before the transcriptional machinery can be recruited again . This interpretation is in line with the observation that Rap1 maintains RP gene promoters essentially nucleosome-free [3] , [29] and that it is necessary and sufficient for TFIID recruitment [51] . It is also consistent with the proposed barrier function of Rap1 in preventing spreading of silent chromatin [28] . Hence , we propose that the natural RP gene promoter architecture ensures efficient promoter activation and rapid responses to input signals even when unspecific chromatin modifications occur . The deterministic analysis suggested that Rap1-containing promoters are more resistant to noise from random chromatin modifications . To further investigate noise propagation , we analyzed the ‘extreme’ models 1 and 4 by stochastic simulations . In addition to chromatin modification , we considered inherent fluctuations of TF levels as noise sources ( see Figure 5A and Materials and Methods ) . A priori , it is therefore not obvious if the architecture with a single TF or the more complex design with three noisy TFs transmits more noise to downstream mRNA production . In particular , we focused on the stationary noise in RP mRNA levels as a function of four factors ( see Figure 1 for the corresponding reactions ) : ( i ) the level of Ifh1 as the main dynamic TF , assuming a constant coefficient of variation ( CV ) for this TF , ( ii ) the noise associated with a constant Ifh1 level , ( iii ) the equilibrium constant of chromatin compaction for a constant inactivation rate , and ( iv ) the velocity of compaction for a fixed equilibrium constant . Figure 5B and 5C show example simulations for models 1 and 4 , respectively , where model parameters are adjusted such that both models generate equivalent average mRNA numbers for the same compaction efficiency . Here , mRNA levels in the simple model drop to very low values much more frequently than for the GRF-containing design , causing increased variation in mRNA numbers ( see also Figure S6 ) . This is a first confirmation of the predictions on noise resistance from the deterministic analysis . To investigate gene expression noise more systematically , we explored the combined effects of variations for pairs of the above influence factors . Noise was quantified by calculating the CV of mRNA numbers for simulated trajectories in steady-state . In the presence of chromatin remodeling , the natural promoter architecture ( model 1 ) exhibits lower mRNA noise than the simple design in all conditions investigated ( Figure 5D–G ) . Notably , mRNA level variations in design 1 are essentially independent of either the velocity ( Figure 5D ) or the strength ( Figure 5F ) of compaction . Increasing Ifh1 noise levels only has a moderate effect in Model 1 and mRNA noise responds more to changes in Ifh1 numbers . By contrast , the simple architecture 4 is inherently more sensitive to the influence of Ifh1 noise levels and chromatin remodeling , especially if compaction is efficient or the chromatin opening/closing cycle is slow ( Figure 5E and 5G ) . The two designs display similar low mRNA noise levels , or even better performance of Model 4 , only when compaction is inefficient ( Ka≥2 . 51 , i . e . when RP genes are active more than 70% of the time in the absence of any TFs ) . We conclude that , despite its complexity , the natural design specifically prevents random fluctuations caused by chromatin remodeling . What are the sources for the lower noise in gene expression of the natural design ? Apparently , scenario 4 produces fewer mRNA molecules than scenario 1 because constitutive chromatin compaction inactivates a higher fraction of promoters . However , a systematic comparison of relative variability for a range of mRNA levels demonstrates that promoter configuration 1 is consistently associated with less noise than design 4 ( Figure 6 ) . The majority of mRNA variation—especially for low mRNA levels—results from irregular promoter activation as indicated by comparison with the expected noise levels due to discrete mRNA numbers ( leading to Poissonian fluctuations ) alone . Noise reduction for the natural architecture only minimally depends on the basal activity of the Rap12-RP gene complex ( not shown ) —it almost exclusively results from the promoter structure . Hence , noise creation and propagation at complex promoters is not solely determined by the binding strength of a particular TF , but also critically depends on the order of recruitment and on dynamic interactions with other TFs . Consequently , the domains mediating protein–protein interactions among cooperating TFs are selectable targets for the evolution of noise traits . More specifically , Rap1-containing promoters achieve low intrinsic noise in gene expression because they minimize stochastic noise induced by unspecific remodeling events , especially for realistic kinetic values and molecule numbers in yeast . Importantly , the simulation results in Figure 5 and Figure 6 demonstrate that this model prediction is robust even when key parameters are perturbed several fold from their nominal values . The regulation of RP genes is intricate because transcriptional co-regulation is particularly strong [26] , [27] . Co-regulated promoter activation is key to induce and maintain concerted expression of gene sets that are required simultaneously . To quantify the degree of mRNA coexpression from individual but identical RP gene promoters , we used the sum of squared pairwise differences between mRNA molecule numbers over time ( cf . Protocol S1 for details ) . The average sum is much smaller for Model 1 than for Model 4 and the differences between natural and simple design are highly significant ( p<10−36 , Welch's t-test ) . Hence , the natural promoter architecture is clearly superior in keeping absolute mRNA levels within tight bounds for large gene sets simultaneously . It enables efficient production of molecular machine precursors in stoichiometric quantities despite short mRNA half-lives that are required for quick adaptation to changes in external conditions . For ribosomal proteins , these features are essential because , when environmental conditions deteriorate , resource-intensive ribosome synthesis must be stopped immediately to reroute building blocks and energy to processes critical for survival . Finally , to critically test the predictive capabilities of the most realistic model ( Model 1 ) we used two independent data sets for model validation . In both cases , except for experiment-specific settings , model structures and parameters remained unchanged . More specifically , we compared model predictions with the experimentally observed dynamic response to IFH1 overexpression to evaluate if model structure and parameters would yield reliable predictions for a regulator contained in the model . First , we compared the predicted dynamics of Ifh1 binding to RP promoters and RP mRNA production for galactose-inducible IFH1 expression with experimental data [31] , [32] ( see Protocol S1 for details ) . Using a fit to the measured IFH1 mRNA profile [31] as input ( Figure 7A ) , Model 1 predicted promoter dynamics and RP mRNA production after induction of the GAL1-IFH1 construct ( GAL1: SGD S000000224 ) . Since the absolute level of basal GAL1-IFH1 expression under non-inducing conditions was not determined , we performed simulations for a range of plausible values . For basal IFH1 mRNA expression at 12% of the wild-type level on glucose we obtained good qualitative and quantitative agreement between model and experimental data ( Figure 7B and 7C ) , independently of the assumed basal activity η ( data not shown ) . The model does not capture the decrease in measured mRNA levels at the last time point . However , no such reduction was observed in a similar experiment [32] ( cf . filled squares in Figure 7C ) . We cannot exclude that deviations between model and data reflect , at least in part , unmodeled mechanisms . Interestingly , the model predicts that larger changes in Ifh1 occupancy at the promoter are not necessarily linked to a monotonic increase in the fold-change of RPmRNA levels ( Figure 7B and 7C and Figure S7 ) . Quantitative discrimination of model alternatives for this experimental setup , therefore , critically depends on accurate quantification of induction dynamics and IFH1 mRNA basal levels in absolute terms; more comprehensive experiments are required to evaluate model performance more stringently . To assess potential structural model uncertainties , we simulated the relation between Ifh1 promoter occupancy ( which is the key control variable in the model ) and RP mRNA production for stresses that might involve unmodeled regulators . In the model , stationary RP mRNA levels depend linearly on the fraction of Ifh1-bound RP promoters . This assumption leads to qualitatively correct predictions of changes in RP mRNA levels ( Figure 7D–F ) in response to heat shock , osmotic shock , and rapamycin addition [32] . Even quantitatively , the differences between measured and simulated responses were not statistically significant ( Welch's t-test , 95% confidence level ) in any of the three conditions . We obtained the same results for predicted Rap1 occupancies and for Fhl1 occupancy under osmotic shock ( Figure S8 ) . The difference between simulated and measured Fhl1 occupancies , however , was significant for heat shock ( p = 0 . 0055 ) and rapamycin addition ( p = 0 . 0084 ) . Underestimation of Fhl1 binding in these conditions may reflect the influence of additional regulators . Such quantitative discrepancies highlight which model aspects require improvement; they identify possible settings under which alternative regulators can be studied in future experiments . We conclude that Model 1 , which considers only Ifh1 and Fhl1 as dominating dynamic regulators , correctly captures salient features of RP gene promoter and mRNA behavior under stationary and dynamic conditions . Biologically meaningful and partly quantitative predictions are , hence , possible already with our simple model .
Complex promoters are involved in many cellular processes where correct timing of expression or precise and coherent regulation of gene sets is required . Their architectures , however , prevent intuitive explanations of promoter functions and advantages for controlling gene expression dynamics . Using the well-characterized yeast RP gene promoter as example , we derive a set of quantitative kinetic models for the natural and for three simplified synthetic promoter configurations . Our model comparison encompasses a broad range of performance characteristics , including dynamic responsiveness and noise transmission , which are not commonly covered in more traditional promoter models [9] , [10] . For the specific example of yeast RP gene promoters , we conclude that the natural design is particularly suited to combine tunable regulation of gene expression with a fast response to external signals , strong co-regulation of target genes , and low mRNA noise in the presence of chromatin remodeling . These are partially contradicting objectives , and a quantitative analysis is required to evaluate the corresponding trade-offs . In particular , the natural promoter can serve as switch between activating and repressing modes with tunable upper and lower activity bounds . Despite the limited quantitative data available for model development , the most realistic models' qualitative—and to a certain extent quantitative—features and predictions comply with our knowledge on RP gene regulation in yeast . Importantly , several predictions are new and experimentally testable: ( i ) the importance of Forkhead proteins for superior shut-off properties and ( ii ) the differential regulation of Fhl1 and Ifh1 required for tunable switch function . Specifically , the role of Rap1-Fhl1 scaffolds in achieving low transcriptional noise mechanistically explain why RP promoters recruiting these factors are associated with low protein noise in vivo [25] and why RP genes exhibit prominent transcriptional co-regulation [26] , [27] . This study's more general results on complex promoter architectures primarily concern the relations between promoter structure and noise resistance . Importantly , GRF-containing architectures render promoter activity robust to influences of unspecific chromatin remodeling , independent of the compaction efficiency and speed . They maintain genes in a poised state for rapid ( re- ) activation even during prolonged absence of the main activating TF . Therefore , complex promoters can contribute much less noise to mRNA levels than simpler designs . This is particularly relevant for highly expressed and unstable proteins , where forced mRNA fluctuations dominate intrinsic noise [17] and , subsequently , total protein noise [25] . Although many GRF-containing promoters are found in highly expressed “housekeeping” genes [52] , not all of these exhibit the same exceptionally low noise as RP genes in vivo [25] . This corroborates that synergistic action with ambivalent TFs such as Fhl1 is crucial . Two important aspects warrant further investigation . First , TFs frequently interact with histones and histone ( de ) acetylases that co-regulate promoter activity [29] , [53] , [54] . Evidence from yeast indicates that dynamic recruitment of chromatin modifiers such as NuA4 can contribute to low noise [25] . Secondly , TATA boxes in promoters of many highly expressed proteins promote transcriptional re-initiation [55] , but also increase intrinsic expression noise through a stable scaffold [15] , [18] . Experimental data [30] , [51] and our simulations demonstrate that GRF binding yields a similarly stable scaffold that leaves genes poised for transcription . Yet RP gene promoters are typically TATA-less [30] , suggesting that the Rap1-Fhl-Ifh1 and similar architectures achieve high expression rates with minimal transcriptional bursts from re-initiation . Our analysis relies on a realistic biological example and , correspondingly , some quantitative model features may be specific for the RP gene system . However , robustness analysis and model optimizations demonstrate that many stationary and qualitative dynamic features are inherent properties of the promoter structure; they are preserved within a broad range typical of physiological parameter values and TF levels . These findings may apply to structurally related promoter architectures , especially those involving certain Forkhead proteins [24] , [56] , [57] or ternary-complex TFs [38] . Indeed , some TCF-type promoters are characteristic of immediate-early genes in mammalian cells [38] . Performance requirements similar to RP genes hold for the synthesis of other molecular machines and for the temporal coordination of the Clb2 cluster in the yeast cell cycle [24] . These experimental observations are consistent with our proposal that the general architecture is especially suited for rapid gene ( re- ) activation and strong transcriptional co-regulation . We expect our study to aid in understanding complex promoter architectures not only in terms of stationary logical functions [2] , [13] , [14] , but also regarding key qualitative aspects of gene expression dynamics .
We modeled molecular interactions of transcription factors at the promoter using chemical reaction kinetics , which lead to ordinary differential equation ( ODE ) models . All deterministic simulations were performed in MATLAB 7 R14 ( The MathWorks , Natick , Mass . ) . For stochastic simulations , we employed extended promoter models that also account for the noise in transcription factor levels , synthesis and degradation of RPmRNAs , and competitive binding of Rap1 to non-RP target genes . Stochastic simulations were performed on a PC cluster using a C-based implementation of the approximate R-leaping algorithm of Auger et al . [58] . Raw simulation results were processed and analyzed in MATLAB 7 R14 ( The MathWorks , Natick , Mass . ) . Details regarding the chemical reaction networks , choices of kinetic constants and initial values as well as settings of the simulation algorithms are described in Protocol S1 . SBML files for models 1 and 4 are provided as Protocols S2 , S3 , S4 , S5 . To assess the influence of TF levels on steady-state promoter activity , total concentrations of the TF under question were varied and the model was simulated until it reached steady state . We assigned activity levels to the resulting complexes between RP gene and TFs depending on composition ( Protocol S1 ) . We investigated the robustness of promoter activity by randomly varying the values of equilibrium constants for binding of each TF and simulating the model into steady state using measured TF concentrations [59] . For each model , 10 , 000 parameter sets were independently sampled from a log-uniform distribution spanning values between 10−1 and 101 nM−1 . Performing the same analysis for a range from 10−2 to 102 nM−1 did not alter the relative sensitivity of promoter activity qualitatively ( data not shown ) . To establish Bode-type plots for the frequency responses of the promoters , we first simulated the ODE models into steady state in the absence of the stimulant ( either Ifh1 or Rap1 ) . Subsequently , a sinusoidal input in the free stimulant concentration was applied such that the concentration oscillated between its total concentration and zero for 50 cycles at the respective frequency . For each model , 4096 points of the simulated trajectories of the relevant molecular species were used to determine their corresponding frequency , amplitude , and phase values by Fast Fourier Transformation . Despite the nonlinear nature of the models , the predominant frequency contribution to the output was always identical to the input frequency . From this data , the associated promoter activities and phase shifts between input and output were calculated . In the stochastic simulation studies addressing RPmRNA noise for promoter designs 1 and 4 , the kinetic constants of gene inactivation/reactivation were increased or decreased up to tenfold while maintaining the nominal Ka value constant . Similarly , the gene inactivation rate constant was kept at its nominal value while varying the probability of an RP gene being active in the absence of TF binding between fa = 1% and fa = 99% by adjusting Ka ( see Protocol S1 for details ) . For fair comparison of mRNA noise between models ( Figure 5B and 5C ) , we chose Ifh1 levels yielding the same mean mRNA number ( 45–46 ) , corresponding to 615 ( Model 1 ) and 1430 ( Model 4 ) molecules of Ifh1 , respectively . Details on simulation conditions , choice of experimental data , and fitting of the IFH1 mRNA time course are described in Protocol S1 .
|
Combinatorial regulation of gene expression is an important mechanism for signal integration in prokaryotes and eukaryotes . Typically , this regulation is established by transcription factors that bind to DNA or to other regulatory proteins . Modifications of the DNA structure provide another layer of control , for instance , in gene silencing . However , it is barely understood how complex promoter architectures determine key features of promoter dynamics such as gene expression levels and noise . Here , we employ realistic mathematical models for prototypical promoters of yeast ribosomal protein genes as well as simplified versions thereof to analyze the relations among promoter design , complexity , and function . By comprehensively analyzing stationary and dynamic promoter properties , we find that functional tradeoffs impose constraints on the promoter architecture . More specifically , a stable configuration in the natural design results in low transcriptional noise and strong co-regulation of target genes in the presence of gene silencing . Combined , our results offer a mechanistic explanation for why specific factors are associated with low protein noise in vivo . We expect that many of these findings apply to other promoters of similar structure .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"computational",
"biology/transcriptional",
"regulation"
] |
2009
|
Precise Regulation of Gene Expression Dynamics Favors Complex Promoter Architectures
|
Evolutionarily conserved mitogen activated protein kinase ( MAPK ) pathways regulate the response to stress as well as cell differentiation . In Saccharomyces cerevisiae , growth in non-preferred carbon sources ( like galactose ) induces differentiation to the filamentous cell type through an extracellular-signal regulated kinase ( ERK ) -type MAPK pathway . The filamentous growth MAPK pathway shares components with a p38-type High Osmolarity Glycerol response ( HOG ) pathway , which regulates the response to changes in osmolarity . To determine the extent of functional overlap between the MAPK pathways , comparative RNA sequencing was performed , which uncovered an unexpected role for the HOG pathway in regulating the response to growth in galactose . The HOG pathway was induced during growth in galactose , which required the nutrient regulatory AMP-dependent protein kinase ( AMPK ) Snf1p , an intact respiratory chain , and a functional tricarboxylic acid ( TCA ) cycle . The unfolded protein response ( UPR ) kinase Ire1p was also required for HOG pathway activation in this context . Thus , the filamentous growth and HOG pathways are both active during growth in galactose . The two pathways redundantly promoted growth in galactose , but paradoxically , they also inhibited each other's activities . Such cross-modulation was critical to optimize the differentiation response . The human fungal pathogen Candida albicans showed a similar regulatory circuit . Thus , an evolutionarily conserved regulatory axis links metabolic respiration and AMPK to Ire1p , which regulates a differentiation response involving the modulated activity of ERK and p38 MAPK pathways .
Organisms sense and respond to diverse stimuli through the action of signal transduction pathways . During complex behaviors like cell differentiation , multiple pathways choreograph changes in the cell cycle and cell polarity to produce a new cell type . In many cases , it is not clear what pathways are involved or how different pathways collaborate to produce major changes in cellular reprogramming . Here , we investigate the problem of cell differentiation in fungal species that differentiate to the filamentous/hyphal cell type . In pathogens , filamentous growth is critical for virulence [1] , [2] . Therefore , identifying the pathways that regulate filamentous growth , and understanding how they function in an interconnected manner , is important for studies in eukaryotic cell differentiation and fungal pathogenesis . In the budding yeast Saccharomyces cerevisiae , different MAPK pathways mediate the response to different stimuli . An ERK-type MAPK pathway called the filamentous growth pathway induces differentiation to the filamentous/invasive/pseudohyphal cell type in response to nutrient limitation [3] , [4] . A p38-type MAPK pathway , called the high osmolarity glycerol response ( HOG ) pathway mediates osmoadaptation [5] , [6] . The two pathways utilize some of the same components , including a core module consisting of the Rho-type GTPase Cdc42p , the p21 activated ( PAK ) kinase Ste20p , the MAPKKK Ste11p , and the adaptor protein Ste50p ( Fig . S1 , [5] , [7]–[9] and references therein ) . Plasma membrane ( PM ) regulators of the filamentous growth pathway , Msb2p , Sho1p , and Opy2p [10]–[12] , also regulate the Ste11p branch of the HOG pathway [13]–[15] . The signaling mucin Msb2p may preferentially regulate the filamentous growth pathway , as another signaling mucin , Hkr1p , has been shown to mainly regulate the HOG pathway [16] , [17] . A second branch of the HOG pathway ( Sln1p branch ) converges on the MAPKK Pbs2p and is regulated by the two-component sensors Sln1p and Ypd1p , the protein kinase Ssk1p , and the redundant MAPKKKs Ssk2p and Ssk22p ( Fig . S1 , [13] , [18]–[21] ) . Thus , different MAPK pathways mediate different responses through the action of common or shared signaling modules . To date , the filamentous growth and HOG pathways are thought to control different responses [22]–[25] . The filamentous growth pathway induces differentiation into chains of branched interconnected filaments by regulating changes in the cell cycle [26]–[28] , cell adhesion [29] , [30] , and cell polarity [31]–[33] . By comparison , the HOG pathway induces transient growth arrest [6] , [34] , [35] by the phosphorylation of translation initiation factors [36]–[38] . The HOG pathway also controls the production of osmolyte ( [39] and references therein ) and regulates changes in chromatin architecture [40]–[43] , yet does not trigger a morphogenetic response . In fact , osmotic stress transiently depolarizes the actin cytoskeleton [44]–[46] . Further evidence that the pathways operate separate programs comes from the fact that the HOG pathway shuts off the filamentous growth pathway in response to osmotic shock [16] , [47]–[49] . The filamentous growth pathway can likewise attenuate the HOG pathway [10] . The paradigm that signaling pathways ( even ones that share components ) function as separate entities is complicated by the fact that under certain conditions the concerted action of multiple MAPK pathways occurs . The HOG and protein kinase C ( PKC , [50] ) pathways are together required in some settings [51] , where they act cooperatively [52] or sequentially [53] . During mating [54] , [55] , the PKC pathway maintains cell integrity [56] , [57] , and the HOG pathway balances changes in osmolarity [35] . Severe stress , like enzymatic digestion of the cell wall [58] , [59] or defects in protein glycosylation [60] trigger multiple pathways as part of an ill-defined response . Diverse stimuli also trigger a general stress response , called the environmental stress response ( ESR , [61] , [62] ) but how this is related to MAPK-dependent responses is not clear . Here , comparative RNA sequencing ( RNA seq ) analysis was used to examine the transcriptional response to inducers of the HOG ( salt ) and filamentous growth ( galactose ) pathways . The response to glycosylation deficiency was also examined , because that stress requires the action of both pathways . Analysis of expression profiling data uncovered galactose as an inducer of the HOG pathway . Accordingly , the AMPK Snf1p and an intact TCA cycle were required to activate the HOG pathway during growth on galactose . The HOG pathway has recently been shown to be activated in response to ER stress and require the unfolded protein response ( UPR ) regulator Ire1p [63] , [64] . Growth of cells in galactose induced Ire1p-dependent activation of the HOG pathway . Therefore , the HOG and filamentous growth pathways are both induced during growth on galactose . Both pathways contributed to growth under this condition , and remarkably , both pathways attenuated each other's activities . Modulation of the filamentous growth pathway by the HOG pathway was required to optimize the differentiation response . The regulatory circuit described here connects general regulators of cellular responses – AMPK , Ire1p , and MAPK ( p38 and ERK ) – in a regulatory axis that controls cell differentiation . This axis was conserved in other fungal species and may underlie differentiation-type responses in metazoans , which contain evolutionarily conserved regulatory pathways .
Comparative RNA sequencing ( RNA seq , [65] ) was performed to examine the response of S . cerevisiae cells to different stimuli . The response to osmotic stress ( YEPD+0 . 4M KCl [23] ) , the non-preferred carbon source galactose ( YEP-GAL , 2% GAL [66] ) , and an inhibitor of N-linked glycosylation ( YEPD+2 . 5 µg tunicamycin [67] ) were examined . Each stimulus induced the expression of overlapping and non-overlapping genes ( Fig . 1A ) . As reported , salt induced targets of the HOG pathway ( Table S1 , [12] , [24] ) , galactose induced the GAL genes and other starvation-responsive genes ( Table S1 , [68] , [69] ) , and tunicamycin induced targets of the UPR and other genes ( Table S1 , [70] , [71] ) . The different stimuli also induced an overlapping gene set ( Fig . 1A , 504 genes ) . Common genes included targets of the ESR ( 150 of 504 total , Table S1 [61] , [62] ) and targets of the HOG pathway . Likewise , a partially overlapping set of repressed genes was also identified that included ESR targets ( Fig . S2A ) . To explore the role of the HOG pathway in response to different stresses , RNA seq profiles were compared between wild-type cells and cells lacking the HOG pathway MAPKK Pbs2p [pbs2Δ] . An overlapping set of Pbs2p-dependent genes was induced in response to all three conditions ( 32 genes , Fig . 1 , B and C ) . Many of the genes ( Fig . 1C , 27/32 , asterisks ) are targets of the ESR . The HOG response to the different stimuli was complex: Pbs2p-dependent targets showed different levels of induction by the different inducers and different induction profiles . For example , different subsets of Pbs2p-dependent targets were unique to each stimulus ( Fig . 1 , B and C; Table S1 ) or common to two of the three stimuli ( Fig . 1B ) . Thus , the HOG pathway actuates common and unique outputs in response to different stimuli . The finding that galactose induced HOG pathway targets in a Pbs2p-dependent manner was unexpected . Pbs2p-dependent targets of the HOG pathway that were induced during growth in galactose regulate mitochondria/respiration , carbohydrate metabolism ( including gluconeogenesis , glyoxylate cycle , glycogen metabolism ) , and amino acid/nitrogen metabolism ( Fig . S2B; Table S1 ) . Quantitative PCR ( qPCR ) analysis confirmed galactose- and Pbs2p-dependent induction of HOG targets , which included established targets of the HOG pathway ( STL1 , ENA1 , GPD1 , CTT1 , and HSP12 ) to varying levels corresponding to the RNA seq analysis ( Fig . 1D , [23] , [24] , [41] , [72] ) . Likewise , the HOG pathway reporter p8XCRE-lacZ [73] was induced by galactose ( by 2 . 3-fold ) in a Pbs2p-dependent manner ( Fig . 1E; salt induced the reporter by 3-fold ) . Therefore , the HOG pathway induces a transcriptional response during growth in galactose . The filamentous growth pathway is induced during growth in galactose [16] , [74] , and cells undergo filamentous growth in this setting . The filamentous growth pathway shares components with the HOG pathway . Comparative RNA seq between wild-type cells and the ste12Δ mutant showed that the HOG and filamentous growth pathways induce different target genes . Known targets of the filamentous growth pathway were identified ( FLO11 , CLN1 , PGU1 , YLRO42C , BAR1 , MSB2 , Table S1 , [11] , [22] , [27] ) as well as new targets , including genes that regulate progression through the G2/M phase of the cell cycle ( CLB1 , CLB2 and SWE1 , [75] , [76]; Fig . 1F ) , bud-site-selection ( Fig . 1F , BUD8 , RAX2 and RSR1 , [77]–[79] ) , a PM regulator of the PKC pathway ( Fig . 1F , WSC2 [80] ) , and components of the filamentous growth pathway ( Fig . 1F , SHO1 and TEC1 [12] , [23] , [81] ) , possibly leading to positive feedback [82] . These genes were not Pbs2p-dependent , and the filamentous growth pathway did not show induction of HOG pathway targets ( Table S1 ) . Therefore , the HOG and filamentous growth pathways mediate non-overlapping responses during growth on galactose . To further examine HOG pathway activation during growth in galactose , phosphorylation of the MAPK Hog1p ( P∼Hog1p ) was measured , which provides a readout of HOG pathway activity [73] , [83] . This assay had the advantage of evaluating the kinetics and genetic pathways required for pathway activation . Consistent with the RNA seq analysis , growth of cells in galactose induced the activation of the HOG ( Fig . 2A , P∼Hog1p ) and filamentous growth pathways ( Fig . 2A , P∼Kss1p ) . Depending on the condition , different levels of basal P∼Hog1p and P∼Kss1p were detected , which may represent differences in baseline activity between the pathways . Other nutrient-limiting conditions were also tested . Non-fermentable carbon sources , acetate and ethanol-glycerol ( Fig . S3A ) , and limitation of environmental nitrogen ( Fig . S3B ) , also activated the HOG pathway . Other poor nutrients , like raffinose ( Fig . S3C ) , limiting glucose ( Fig . S3C ) , and phosphate limitation ( Fig . S3D ) did not induce the HOG pathway , which indicates that a particular metabolic context evokes the response ( see below ) . Because galactose caused robust induction of the HOG and filamentous growth pathways , that inducer was used for subsequent experiments . An established trigger of the HOG pathway is an increase in external osmolarity [84] . HOG pathway activation by high osmolarity ( KCl , 0 . 4 M ) and galactose ( 2% ) was compared . In response to osmotic stress , P∼Hog1p was detected by 5 min . By 60 min , P∼Hog1p was reduced due to pathway attenuation ( Fig . 2 , B and C , upper panel [84] , [85] ) . By comparison , P∼Hog1p was detected during growth in galactose at 240 min ( Fig . 2 , B and C , middle panel ) , and the signal persisted until 720 min ( Fig . 2D ) . Salt ( 0 . 4 M KCl ) induced higher levels of P∼Hog1p than growth in galactose ( Fig . 2E ) , which is consistent with the RNA seq analysis ( Table S1 ) . Salt ( 0 . 4 M KCl ) added to cells grown in galactose caused a rapid increase in P∼Hog1p ( Fig . 2 , B and E ) . Transfer of cells from galactose ( YEP-GAL ) to glucose ( YEPD ) , which leads to glucose repression ( see below ) , caused a reduction in P∼Hog1p levels by 20 min ( Fig . 2 , B and F ) , which was comparable to the reduction in P∼Hog1p levels in response to osmotic shock ( Fig . 2 , B and C ) . Therefore , the amplitude and duration of HOG pathway signaling differs depending on whether the inducer is osmotic stress or galactose . Two branches of the HOG pathway regulate the response to osmotic stress , which converge on the MAPKK Pbs2p ( Fig . S1 , [13] , [18] , [19] ) . In response to salt , the branches are redundant , in that mutants lacking both branches ( ssk1Δ ste11Δ ) show the same defect as mutants lacking the MAPKK ( pbs2Δ ) or MAPK ( hog1Δ , Fig . 2G ) . During growth in galactose , but not salt , the MAPKKKs Ssk2p and Ssk22p were required for HOG pathway activation ( ssk2Δ ssk22Δ; Fig . 2G ) . Thus , Ssk2p and Ssk22p have a role in the response to galactose that differs from their role in response to osmotic stress . The different branches of the HOG pathway play different roles under different conditions , such as in the response to different concentrations of salt [23] . Nitrogen limitation showed a similar requirement for Ssk2p and Ssk22p ( Fig . S3E ) . The Ste11p branch alone was not required ( ste11Δ , Fig . 2G ) , but when the ste11Δ mutant was combined with the ssk1Δ mutant , a defect was observed ( ssk1Δ ste11Δ , Fig . 2G ) . Thus , the Ste11p branch plays a minor role in the HOG response to galactose . One might expect that the Msb2p/Sho1p branch , which induces the filamentous growth pathway in response nutrient limitation , would transmit nutrient signals to Pbs2p/Hog1p ( Fig . S1 ) . In fact , the Sln1p branch played the major role in this nutritional response , whereas the Sho1p branch played a minor role . Glucose is the preferred carbon source in yeast [86] , [87] . When glucose is abundant , yeast cells exclusively utilize that nutrient over non-preferred carbon sources like galactose . Glucose repression prevents the transport and utilization of other carbon sources [86] , [88]–[90] . As shown above , glucose added to cells grown on galactose resulted in attenuation of the HOG response ( Fig . 2F ) . To further test whether glucose prevents the HOG response to galactose , cells were grown in media containing both glucose and galactose as a carbon source . Under this condition , HOG pathway signaling was also attenuated ( Fig . 3A ) . These experiments indicate that galactose metabolism is required for HOG pathway activation . Consistent with this possibility , mutants defective for galactose transport and utilization ( Fig . 3B; gal3Δ , gal4Δ , gal7Δ , and gal10Δ [88] , [90]–[93] ) were defective for HOG pathway activation . Galactose utilization increases the respiratory capacity by shunting ATP production through the electron transport chain [86] , [94] , [95] . An inhibitor of respiration , antimycin [96]–[98] , prevented HOG pathway activation by galactose ( ANT , Fig . 3C ) . Antimycin did not prevent HOG pathway activation in response to salt ( Fig . S4A ) . Metabolic respiration produces intermediates that are utilized by the tricarboxylic acid ( TCA , or citric acid ) cycle to generate precursors for electron transport . Strains lacking TCA cycle enzymes aconitase ( aco1Δ ) , fumarase ( fum1Δ ) , malate dehydrogenase ( mdh1Δ ) , alpha keto-glutarate ( kgd1Δ ) and iso-citrate dehydrogenase ( idh1Δ ) were defective for HOG pathway activation by galactose ( Fig . 3D ) . These mutants were not required to mediate an osmotic response ( Fig . S4B , shown for aco1Δ ) . Therefore , metabolic respiration of galactose underlies activation of the HOG pathway . The AMP-dependent protein kinase ( AMPK ) Snf1p is a major regulator of the response to poor carbon source utilization [99] , [100] . The main function of Snf1p is the de-repression of glucose-repressed genes [86] , [89] , [99] . Snf1p was required for HOG pathway activation by galactose ( Fig . 3E ) . Snf1p functions with the regulatory subunit Snf4p [101] , which was also required for HOG pathway activation by galactose ( Fig . S4C ) . Snf1p and Snf4p were not required for HOG pathway activation in response to osmotic stress ( Fig . S4D ) . Snf1p phosphorylates the transcriptional repressor Mig1p to relieve glucose repression [102]–[104] , which leads to induction of the GAL genes and other genes [89] , [102] , [105] , [106] . Loss of Mig1p restored HOG pathway activity in the snf1Δ mutant ( Fig . 3E , mig1Δ snf1Δ ) . Cells lacking Mig1p alone did not influence HOG pathway activity ( Fig . 3E , mig1Δ ) . Therefore , Snf1p regulates the HOG pathway through its major role in relieving de-repression of glucose-repressed genes . Snf1p also regulates nitrogen assimilation pathways [107] but was not required to activate the HOG pathway in response to nitrogen deficiency ( Fig . S4E ) . In summary , metabolic respiration triggers AMPK-dependent activation of the HOG pathway . Protein glycosylation is an oligosaccharide modification of proteins that occurs in the endoplasmic reticulum ( ER ) and Golgi apparatus [108] . Defects in protein glycosylation trigger a global response that involves the action of several MAPK pathways , including the filamentous growth [10] , [60] and HOG pathways [63] , [64] . Comparative RNA seq analysis identified HOG pathway targets induced by treatment with tunicamycin , an inhibitor of N-linked glycosylation ( Fig . 1 , B and C; Table S1 ) . To further explore the HOG and filamentous growth pathway response to glycosylation deficiency , a conditional mutant , pmi40-101 [60] , was used that is defective for an early step in N- and O-linked glycosylation [109] , [110] . Growth of the pmi40-101 mutant in media lacking mannose induces its glycosylation defect and showed elevated HOG and filamentous growth pathway activity ( Fig . 4A ) . Defects in O-linked glycosylation also modestly activated the HOG pathway ( Fig . S5A ) . In response to glycosylation deficiency , HOG pathway activation did not require Snf1p ( Fig . S5B ) , which is consistent with the idea that Snf1p regulates the HOG pathway by the de-repression of glucose-repressed genes . Defects in protein glycosylation induce problems with protein folding in the ER , which activates the UPR [111] , [112] . The UPR is regulated by the kinase Ire1p [113] , which has recently been shown to mediate the HOG pathway response to glycosylation deficiency ( Fig . 4B [63] , [64] ) . As expected from these reports , Ire1p was required to mediate the HOG response to glycosylation deficiency ( Fig . 4B ) . Protein glycosylation deficiency , as induced by tunicamycin or the pmi40-101 mutant , also induced a transcriptional reporter of the UPR ( UPRE-lacZ , [113] ) in an Ire1p-dependent manner ( Fig . 4C ) . An increase in metabolic respiration might also trigger ER stress that leads to Ire1p-dependent activation of the HOG pathway . HOG pathway activation in response to galactose was reduced in the ire1Δ mutant ( Fig . 4D ) . Galactose also induced expression of the UPRE-lacZ reporter in an Ire1p-dependent manner ( Fig . 4C ) . Induction of the UPRE-lacZ reporter by galactose was abolished by treatment with antimycin ( Fig . 4C ) . Therefore , an increase in metabolic respiration stimulates the UPR , which leads to Ire1p-dependent activation of the HOG pathway . The HOG pathway is activated by several stimuli , including salt , nitrogen ( this study ) , myriocin [14] , [114] , and oxidative stress [115] , [116] . The UPR was not induced by these stimuli ( Fig . 4C ) . Therefore , two different types of inducers activate the HOG pathway . One type is Ire1p-dependent ( induced by increased metabolic respiration and glycosylation deficiency ) , and another is Ire1p-independent ( induced by salt and other stresses ) . The filamentous growth and HOG pathways are activated during growth in galactose ( Fig . 1A; Fig . 2A; Table S1 ) . However , the HOG pathway inhibits the filamentous growth pathway in response to osmotic stress [12] , [16] , [47]–[49] . To determine whether the HOG pathway inhibits the filamentous growth pathway during growth in galactose , cells were examined by microscopy . Cells lacking an intact HOG pathway showed hyper-polarized growth ( Fig . 5A , pbs2Δ ) , indicative of a hyperactive filamentous growth pathway . Comparative RNA seq showed that transcriptional targets of the filamentous growth pathway ( PGU1 , SVS1 , MSB2 , and KSS1; Table S1 ) were up-regulated in the pbs2Δ mutant in galactose . In line with the RNA seq data , the pbs2Δ mutant showed elevated activity of a filamentous growth pathway reporter ( Fig . 5B , FRE-lacZ ) . Tyrosine phosphatases Ptp2p and Ptp3 negatively regulate the Hog1p pathway [85] , [117] . The ptp2Δ ptp3Δ double mutant showed elevated P∼Hog1p levels and correspondingly lower levels of P∼Kss1p under pathway inducing conditions ( galactose ) ( Fig . 5C ) . This was also observed under basal conditions ( glucose ) ( Fig . 5C ) . Accordingly , the ptp2Δ ptp3Δ double mutant ( and the ptp3Δ single mutant ) showed reduced invasive growth and crosstalk reporter activity ( Fig . S6 ) . Therefore , the HOG pathway inhibits the filamentous growth pathway during growth in galactose . The filamentous growth pathway also inhibits the HOG pathway [10] . Consistent with this finding , the level of P∼Kss1p was elevated in the pbs2Δ mutant at 0 h , 2 h , and 4 h ( Fig . 5D ) of growth in galactose . Similarly , the level of P∼Hog1p was elevated in the kss1Δ mutant at 6 h and 8 h ( Fig . 5E ) . RNA seq analysis showed STE12 was up-regulated by galactose in the pbs2Δ mutant ( Table S1 ) . This was confirmed by qPCR ( 1 . 52 log2 fold ) ( Fig . 5F ) and was reflected at the protein level ( Fig . 5G ) . Likewise , Hog1p protein levels were modestly affected in the ste12Δ mutant ( Fig . 5G ) . Thus , the HOG pathway inhibits the filamentous growth pathway during growth in galactose , which may occur at the STE12 level . These results show that the HOG and filamentous growth pathways modulate each other's activities in a complex pattern in galactose . To this point , our results suggest an apparent paradox . The HOG and filamentous growth pathways are both activated during growth on galactose , yet the pathways dampen each other's activities . To determine the roles of the pathways in this setting , mutants in the filamentous growth and HOG pathway pathways were examined for growth in galactose . Mutants lacking the filamentation MAPKK ( ste7Δ ) and HOG MAPKK ( pbs2Δ ) were not defective for growth on galactose ( Fig . 6A ) . However , the ste7Δ pbs2Δ double mutant showed a growth defect ( Fig . 6A ) . This defect was not specific for the MAPKKs , because another mutant that blocks the activity of both pathways showed an equivalent growth defect ( ste11Δ ssk1Δ , Fig . 6A ) . The ste7Δ pbs2Δ double mutant also showed morphological defects during growth in galactose ( Fig . 6B ) . Therefore , the HOG and filamentous growth pathways have a redundant function in promoting proper growth and morphogenesis during growth in galactose . What is the benefit to the cross-modulation between the two pathways ? One possibility is that modulation of the pathways' activities may be important to optimize the response . To test this possibility , the pbs2Δ mutant was examined in detail by microscopic examination . A subset of pbs2Δ cells showed morphological defects [∼5% compared to <0 . 5% of wild-type cells , >1000 cells counted ( Fig . 6C ) ] . Thus , hyper-activation of the filamentous growth pathway can lead to morphogenetic defects . To further explore this possibility , the pattern of septins , which mark the mother-bud neck [118] , [119] , was also examined . Septin staining showed an irregular pattern pbs2Δ cells with morphological defects ( Fig . 6C ) . This defect is indicative of problems with cell-cycle progression or proper growth . To determine if the hyper-polarized growth of the pbs2Δ mutant comes from hyper-activation of the filamentous growth pathway , cell morphology was quantitated by microscopy . In rich media ( YEPD ) , wild-type cells grow predominately in the vegetative ( round ) form ( 8+/−2% elongated cells; 200 cells counted for all trials ) . By comparison , the pbs2Δ mutant shows a cell-elongation morphology ( >83 . 5+/−5% elongated cells ) that was abolished in the pbs2Δ ste7Δ double mutant ( 7+/−3% elongated cells ) . Therefore , the enhanced polarized morphology seen in pbs2Δ cells , and concomitant morphological abnormalities , is due to Ste7p . This result is complicated because in YEP-GAL , the ste7Δ pbs2Δ double mutant shows morphological defects not seen in either single mutant ( Fig . 6B ) . Thus , Pbs2p may attenuate morphogenesis in multiple ways . This finding extends to other negative-regulatory inputs to the filamentous growth pathway as well ( Chavel et al . IN PRESS ) . Although only a low percentage of cells exhibit morphological defects , it is likely that even minor mis-coordination of basic cellular processes would be detrimental to cell health . Thus , modulation of the filamentous growth pathway by the HOG pathway is necessary for proper cell growth and morphogenesis . As a second test , the response of a population of cells was examined . Yeast cells expand in biofilms/mats through the action of the filamentous growth pathway , which regulates expression of the cell-adhesion molecule Flo11p [120] . When hyper-activated , the filamentous growth pathway causes an increase in FLO11 expression that prevents biofilm/mat expansion [121] , [122] . We found that the pbs2Δ mutant formed smaller biofilms/mats than wild-type cells during expansion on galactose media ( Fig . 6D , see quantitation in graph ) . Thus , modulation of the filamentous growth pathway by the HOG pathway is required to coordinate cell growth and optimize colonial behavioral responses . The signaling circuit characterized here might be specific to S . cerevisiae or extend to other species . To address this question , pathways of the fungal pathogen Candida albicans were examined . Like budding yeast , C . albicans has a Kss1p-type pathway ( Cek1p pathway [123]–[125] ) , and a p38-type pathway ( CaHOG pathway [126] ) . The CaHOG MAPK CaHog1p is activated by osmotic stress ( Fig . 7A [127] ) and was also induced by tunicamycin , myriocin , and growth in galactose ( Fig . 7A , 5 h at 30°C ) . Thus , the versatility of HOG pathway in sensing diverse stresses is conserved among several fungal species . We also tested whether CaIre1p is required to mediate the HOG pathway response to galactose . The ire1Δ/ire1Δ double mutant was defective for producing the elevated levels of P∼CaHog1p seen during galactose treatment in wild-type cells ( Fig . 7B ) . A strain lacking CaIre1p but containing a complemented version ire1Δ/IRE1 , restored P∼CaHog1p activity ( Fig . 7B ) . Thus , CaIre1p mediates the HOG pathway response to galactose . Previous reports have shown that the C . albicans HOG pathway negatively regulates the Cek1p pathway [128] . In C . albicans , growth at high temperatures ( 37°C ) is a potent inducer of dimorphism [129] , [130] . Interestingly , growth of C . albicans cells in galactose induced P∼Cek1p levels only at 37°C in wild-type cells ( Fig . 7C ) . The C . albicans hog1Δ/hog1Δ mutant grown in galactose showed elevated P∼Cek1p levels at 30°C and 37°C ( Fig . 7C ) . Similarly , the hog1Δ/hog1Δ mutant showed hyper-invasive growth compared to wild-type cells ( Fig . 7D ) . Thus , CaHog1p inhibits Cek1p pathway activity during growth in galactose . The Cek1p pathway might also inhibit the CaHOG pathway . In the cek1Δ/cek1Δ mutant , elevated P∼CaHog1p levels were observed during growth in galactose at 30°C and 37°C ( Fig . 7E ) . Therefore , the CaHog1p pathway is activated by galactose in a Ire1p-dependent manner . Under this condition , the CaHog1p pathway and Cek1p pathways are both activated and both modulate each other's activities . These results indicate that the signaling axis described in S . cerevisiae extends to the opportunistic pathogen C . albicans .
Differentiation into specialized cell types occurs during development and in response to extrinsic cues . Fungal species differentiate into the filamentous/hyphal cell type , which in pathogens occurs during colonization of the host . Using a genomic survey , RNA seq analysis , we identify a new role for a p38 MAPK pathway ( HOG ) in differentiation to the filamentous cell type in yeast . The HOG pathway is activated during growth in poor carbon sources through a regulatory circuit involving the AMPK Snf1p . Since deletion of MIG1 alleviates the need for Snf1p , the regulation of Hog1p is likely to be downstream of Mig1p repressed genes . HOG pathway activation in this context also required the ER stress kinase Ire1p . The connection between Ire1p and the HOG pathway has been reported [63] , [64] . Our study therefore connects AMPK and Ire1p in a regulatory circuit that governs p38 ( Fig . 7F ) . The HOG pathway and the ERK-type filamentous growth pathway induced target genes to promote growth in galactose , but the pathways also modulated each other's activities . Such modulation optimized cell growth and morphogenesis to facilitate production of the filamentous cell type ( Fig . 7F ) . Our findings therefore elucidate a signaling network that occurs during differentiation and highlights the critical role for pathway modulation in proper cell-type specification . Nutrient sensing in yeast has been extensively studied . Well-established pathways mediate the response to carbon and nitrogen limitation . We show that limiting nutrients , like non-preferred carbon and nitrogen sources , activate the p38-type HOG pathway in yeast . As a result , glucose repression , the AMPK Snf1p , metabolic respiration , and the TCA cycle feed into HOG pathway signaling ( Fig . 7F ) . The HOG pathway is well known for its ability to sense and respond to changes in external osmolarity . Multiple inducers activate the HOG pathway , including citric acid [131] , hypoxia [132] , cold stress [133] and defective sphingolipid biosynthesis [114] . Here we demonstrate a connection between nutrients and activation of the HOG pathway ( Fig . 7F ) . In mammals , p38-type stress activated protein kinase ( SAPK ) pathways mediate AMPK-dependent metabolic reprogramming . The p38 pathway can alter the balance between survival and apoptosis [134] . p38 also regulates respiration in muscles , and gluconeogenesis in liver , and when mis-regulated can lead to problems that range from diabetes [135]–[137] to tumor malignancy [138] . We show that in yeast , metabolic respiration feeds into the HOG response and requires the UPR regulator Ire1p . Several inducers of the Ire1p have been identified [139] but to our knowledge , this is the first example of a connection between defects in metabolic respiration and Ire1p ( Fig . 7F ) . Hints at this connection come from studies in mammalian cells . Ire1p can regulate AMPK function [140] , and glucose levels are connected in some manner to Ire1p activity [141] . Elements of the ER stress pathway drive metabolic reprogramming in triple-negative breast cancer cells [142] . In tumor microenvironments , Ire1p is required to promote the balance between lipid and protein biosynthesis potentially at the level of ER production [143] . It is plausible that Ire1p regulates p38 activity in these settings as well . It is unlikely that the AMPK-Ire1p-p38 circuit governs all p38-type responses; however , the connections reported here may underlie nutrient-dependent p38-type responses in many settings . Metabolic respiration ( and defects in protein glycosylation ) induce the HOG ( p38 ) and filamentous growth ( ERK ) pathways . The two pathways do not depend on each other for activation , and they induce non-overlapping targets ( herein and [23] , [82] ) . Both pathways are redundant for full growth on galactose , and the two pathways modulate each other's activities . This complicated functional interplay between the pathways is critical for proper cell growth and optimal differentiation to the filamentous cell type . Generally speaking , p38 and ERK pathways can be activated in the same cells to together orchestrate complex responses that include cell differentiation [144] , [145] . An important insight from our study is that p38 and ERK pathways modulate each other's activities to produce an optimal response . Such cross-wiring to fine tune responses underscores the importance of precision in cell differentiation .
Yeast and bacterial strains were grown by standard methods [146] , [147] . Yeast strains were grown on YEP ( yeast extract and peptone ) medium containing 2% glucose ( D ) or 2% galactose ( GAL ) unless otherwise indicated . Cells were grown at 30°C . Strains are listed in Table 1 . Primers used in the study are listed in Table 2 . The plate-washing assay [148] and the single cell invasive growth assay [60] were performed as described . Biofilm assays were performed as described [120] , except that galactose was used as a carbon source . Plasmid YCp-Cdc12p-GFP was provided by J . Pringle [149] , pFRE-lacZ by H . Madhani [150] , pUPRE-lacZ by David Eide [151] , and p8X-CRE-lacZ by H . Saito [73] . Plasmid selection was maintained in synthetic complete medium containing 2% glucose ( SD ) or 2% galactose ( S-GAL ) that lacked uracil ( -URA ) or leucine ( -LEU ) . Gene disruptions were performed according to standard genetic techniques [152] , [153] . Total RNA was isolated by acid phenol method from 10 ml cultures of WT and pbs2Δ mutant grown in YEPD ( 5 . 5 hrs ) , YEP-GAL ( 5 . 5 hrs ) , YEPD+Tunicamycin ( 3 hrs ) and YEPD+salt ( 10 min ) . Isolated RNA was purified over a RNAeasy column ( Qiagen ) . RNA concentration was measured using the NanoDrop2000 spectrophotometer ( Thermo Scientific , Wilmington , DE ) and purity was established by the A260/A280 ratio . RNA was detected by running samples on an 8M Urea 6%polyacylamide gel , stained by ethidium bromide . Three independent inductions were evaluated for RNA-seq analysis . Total RNA integrity was checked using an Agilent 2200 TapeStation ( Agilent Technologies , Inc . , Santa Clara , CA ) and quantified using a Trinean DropSense96 spectrophotometer ( Caliper Life Sciences , Hopkinton , MA ) . RNA sequencing was performed at the Fred Hutchinson Cancer Research Center ( Seattle , WA ) . RNA seq was performed in triplicate by sequencing RNA prepared from 3 different ( independent ) cultures . RNA-seq libraries were prepared from total RNA using the TruSeq RNA Sample Prep Kit ( Illumina , Inc . , San Diego , CA , USA ) . Library size distributions were validated using an Agilent 2200 TapeStation ( Agilent Technologies , Santa Clara , CA , USA ) . Additional library QC , blending of pooled indexed libraries , and cluster optimization was performed using Life Technologies' Invitrogen Qubit 2 . 0 Fluorometer ( Life Technologies-Invitrogen , Carlsbad , CA , USA ) . RNA-seq libraries were pooled ( 24-plex ) and clustered onto a flow cell lane using an Illumina cBot . Sequencing was performed using an Illumina HiSeq 2500 in Rapid Mode employing a paired-end , 50 base read length ( PE50 ) sequencing strategy . Image analysis and base calling were performed using Illumina's Real Time Analysis v1 . 17 software , followed by ‘demultiplexing’ of indexed reads and generation of FASTQ files , using Illumina's CASAVA v1 . 8 . 2 software ( http://www . illumina . com/software . ilmn ) . For analysis of the RNA seq data , reads of low quality were filtered out prior to alignment to the reference genome ( S . cerevisiae assembly R64-1-1 , Ensembl release 75 ) using TopHat v2 . 0 . 9 [154] . Counts were generated from TopHat alignments for each gene using the Python package HTSeq v0 . 5 . 4 ( http://www-huber . embl . de/users/anders/HTSeq/doc/overview . html ) . Genes with low counts across all samples were removed , prior to identification of differentially expressed genes using the Bioconductor package edgeR v3 . 4 . 2 [155] . A false discovery rate ( FDR ) method was employed to correct for multiple testing [156] . Differential expression was defined as |log2 ( ratio ) |≥0 . 585 ( ±1 . 5-fold ) with the FDR set to 5% . cDNA was synthesized using iScript cDNA synthesis kit ( BioRAD , Carlsbad CA ) according to manufacturer's protocol . PCR reactions were set-up using iQ SYBR Green Supermix ( BioRAD , Carlsbad , CA ) . qPCR was performed using the following amplification cycles: initial denaturation for 8 min at 95°C , followed by 35 cycles ( denaturation for 15 sec at 95°C and annealing for 1 min at 60°C ) . Expression of genes was quantified using the 2−ΔΔCt method [157] where ACT1 ( actin ) was used for normalization of expression values . To analyze HOG and filamentous growth pathway activity by phosphoblot analysis , cells were induced under the following conditions . For S . cerevisiae , cells were grown in YEPD to mid-log phase , and 0 . 4 M KCl was added for 5 min . For C . albicans , 0 . 4 M NaCl was used . For galactose , cells were grown in YEP-GAL medium to mid-log phase ( 5 . 5 hrs ) . For tunicamycin , cells were grown to mid-log phase in YEPD , and tunicamycin was added to cells for 3 hrs . [2 . 5 µg Sigma CAT . # T7765] . Antimycin [Sigma CAT # A8674] was added to mid-log phase cells for 3 hrs . ( 2 . 5 µg , 3 hrs ) , H2O2 [Sigma CAT #216763] was added to cells at a concentration of 5 mM for 20 min . Myriocin [Sigma CAT #M1177] . Yeast strains were grown in nitrogen free media ( yeast nitrogen base without amino acids and without ammonium sulfate [1 . 7 g/L] BD Franklin Lakes , NJ; #233520 ) [31] supplemented with glucose as a carbon source . For phosphate free medium ( yeast nitrogen base without amino acids and without phosphate [5 . 6 g/L] , MP Biomedicals LLC , Solon , OH; #114027812 ) was supplemented with amino acids as a nitrogen source and glucose as a carbon source . For induction of Candida albicans pathways , strains were maintained at 30°C . Cells were grown to mid-log phase ( ∼5 hrs ) in YEPD or YEP-GAL and treated with 0 . 5 M NaCl for 10 min , myriocin ( 5 mM for 10 min ) , and tunicamycin ( 2 . 5 µg for 3 hrs ) . The pmi40-101 mutant was grown in YEPD medium supplemented with or without 50 mM mannose for 5 hrs . For strains that exhibited growth defects in galactose , input cell number ( OD600 ) was increased to be equivalent to wild-type cells at mid log phase . Cell extracts were prepared for immunoblot analysis according to established procedures [158] . Mid-log phase cells were harvested by centrifugation , and proteins were precipitated by trichloroacetic acid [TCA] . Cells were lysed in the TCA buffer ( 10 mM Tris HCl pH 8 . 0; 10%TCA; 25 mM ammonium acetate; 1 mM EDTA ) containing glass beads using FastPrep-24 Instrument ( MP Biomedicals LLC , Solon , OH ) . After high-speed centrifugation the pellet was thoroughly mixed in the resuspension buffer ( 0 . 1M Tris HCl pH 11 . 0; 3%SDS ) and boiled for 5 min and centrifuged for 30 sec at 16000 g . To the supernatant , equal volume of 2× SDS loading dye ( 100 mM Tris HCl pH 6 . 8; 4%SDS; 0 . 2% Bromophenol Blue; 20% glycerol; 200 mM ß-mercaptoethanol ) was added . Protein samples were separated on 10% SDS polyacrylamide gels ( SDS-PAGE ) and transferred to nitrocellulose membranes ( Protran BA85 , VWR International Inc . , Bridgeport NJ ) . The membrane was blocked in immunoblot buffer ( 5% nonfat dry milk , 10 mM Tris-HCl [pH 8] , 150 mM NaCl and 0 . 05% Tween 20 ) for 16 h at 4°C . WesternBright MCF fluorescent Western blotting kit from Advansta Inc . ( Menlo Park , CA; LPS #K-12045-D20 ) was used for detection . Pgk1p antibodies ( Life Technologies , Camarillo , CA; #459250 ) were used as a loading control . P∼Hog1p was detected using phospho-p38 antibodies ( Cell Signaling Technology , Danvers , MA; #9211 ) . S . cerevisiae Hog1p was detected by Hog1p antibodies ( Santa Cruz Biotechnology , Santa Cruz , CA; #yC-20 ) . C . albicans Hog1p was detected by ( Santa Cruz Biotechnology , Santa Cruz , CA; #y-215 ) . Cdc2 p34 antibody that recognizes PSTAIRE motifs in cyclin dependent kinases was used as a loading control for Candida protein extracts ( Santa Cruz Biotechnology , Santa Cruz , CA; #sc-53 ) . Phosphorylated Kss1p was detected by p42/p44 antibodies ( Cell Signaling Technology , Danvers , MA; #4370 ) and total Kss1p was detected by ( Santa Cruz Biotechnology , Santa Cruz , CA; #6775 ) . Secondary antibodies , goat α-mouse IgG–HRP ( Bio-Rad Laboratories , Hercules , CA; #170-6516 ) , goat α-rabbit IgG-HRP ( Jackson ImmunoResearch Laboratories , Inc . , West Grove , PA; #111-035-144 ) , donkey α-goat IgG-HRP ( Santa Cruz Biotechnology , Santa Cruz , CA; #sc-2020 ) were used and incubated for 1 hr at 20°C . Ponceau S ( Sigma , St . Louis , MO; #P7170 ) was used to confirm equal loading among samples . ß-galactosidase assays were performed as described [11] . Cells were grown in selective media ( SD-URA ) for 16 hrs and sub-cultured in YEPD or YEP-GAL media for 5 . 5 hrs . Three independent experiments were performed and the average values are represented . Error bars indicate the standard deviation between trials . Differential-interference-contrast ( DIC ) and fluorescence microscopy was performed with an Axioplan 2 fluorescent microscope ( Zeiss ) with a PLAN-APOCHROMAT 100×/1 . 4 ( oil ) objective ( N . A . 0 . 17 ) . Digital images were obtained with the Axiocam MRm camera ( Zeiss ) . Image Acquisition and analysis was carried out using Axiovision 4 . 4 software ( Zeiss ) . Heat maps were generated using MeV ( MultiExperiment Viewer ) ( http://www . tm4 . org/mev . html ) . ImageJ analysis was used to quantitate band intensity for protein gels and immunoblots ( http://imagej . nih . gov [159] ) using the invert function and by subtraction of background signals . SGD was used for yeast gene annotation and analysis ( http://www . yeastgenome . org ) . RNA seq data was evaluated and represented by Microsoft EXCEL software .
|
In fungal species , differentiation to the filamentous/hyphal cell type is critical for entry into host cells and virulence . Comparative RNA sequencing was used to explore the pathways that regulate differentiation to the filamentous cell type in yeast . This approach uncovered a role for the stress-response MAPK pathway , HOG , during the increased metabolic respiration that induces filamentous growth . In this context , the AMPK Snf1p and ER stress kinase Ire1p regulated the HOG pathway . Cross-modulation between the HOG and filamentous growth ( ERK-type ) MAPK pathways optimized the differentiation response . The regulatory circuit described here may extend to behaviors in metazoans .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"fungal",
"genetics",
"microbiology",
"osmotic",
"shock",
"systems",
"ecology",
"fungi",
"nutrition",
"microbial",
"genomics",
"mycology",
"medical",
"microbiology",
"yeast",
"systems",
"biology",
"signal",
"transduction",
"cell",
"biology",
"ecology",
"phenotypes",
"gene",
"regulatory",
"networks",
"genetics",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"genomics",
"nutrients",
"computational",
"biology",
"organisms"
] |
2014
|
Metabolic Respiration Induces AMPK- and Ire1p-Dependent Activation of the p38-Type HOG MAPK Pathway
|
HIV-1 replicative capacity ( RC ) provides a measure of within-host fitness and is determined in the context of phenotypic drug resistance testing . However it is unclear how these in-vitro measurements relate to in-vivo processes . Here we assess RCs in a clinical setting by combining a previously published machine-learning tool , which predicts RC values from partial pol sequences with genotypic and clinical data from the Swiss HIV Cohort Study . The machine-learning tool is based on a training set consisting of 65000 RC measurements paired with their corresponding partial pol sequences . We find that predicted RC values ( pRCs ) correlate significantly with the virus load measured in 2073 infected but drug naïve individuals . Furthermore , we find that , for 53 pairs of sequences , each pair sampled in the same infected individual , the pRC was significantly higher for the sequence sampled later in the infection and that the increase in pRC was also significantly correlated with the increase in plasma viral load and with the length of the time-interval between the sampling points . These findings indicate that selection within a patient favors the evolution of higher replicative capacities and that these in-vitro fitness measures are indicative of in-vivo HIV virus load .
Measuring the fitness of HIV-1 is notoriously difficult . At the between-host level , fitness can be interpreted as the transmission potential which is defined as the expected number of transmissions in the course of an infection [1] . This quantity can however only be measured in cohorts of untreated patients with known infection status that are followed over long time periods [1] . At the within-host level , fitness is determined by the average number of secondary infected cells resulting from a single infected cell in vivo . This hypothetical quantity is difficult to determine [2] but can be approximated by in-vitro measurements of the replicative capacity ( RC ) ( see [3] ) . However , the in-vivo relevance of such in vitro fitness values is largely unclear . In a recent publication , some of the authors of this article described a computational method to predict RC values on the basis of viral amino-acid sequences [3] . To this end , a machine-learning algorithm based on a quadratic fitness model was applied to a training data set of 65 , 000 amino-acid sequences of the pol gene and the associated RC values . The resulting RC-predictor could explain roughly 40% of the deviance of RC values in a test-data set consisting of 5 , 000 sequences , which had not been used for the inference of this predictor . In the present study , we apply this computational predictor to clinical data from the Swiss HIV Cohort Study ( SHCS ) ( www . shcs . ch ) in order to obtain an assessment of the RC-predictor in an independent dataset and to study its correlation with plasma HIV RNA viral load , a known surrogate marker associated with disease progression [3] .
The Swiss HIV cohort study was approved by individual local institutional review boards of all participating centers ( www . shcs . ch ) . Written informed consent was obtained for each SHCS study participant . Fitness is measured as the log replicative capacity of HIV-derived amplicons [representing all of Protease ( PR ) and most of Reverse Transcriptase ( RT ) ] inserted into a constant backbone of a resistance test vector . The models are then trained to predict this fitness from the amino-acid sequence of the amplicons . Details on the experimental measurement of the RC values and on inferring the predictor have been published in [3] . Here , we briefly reiterate the principles of the models fitted . In essence , the predictor is based on fitting the data consisting of amino acid sequences s and the corresponding log-RC values ( w ) with the following model ( M1 ) sij denotes the presence ( sij = 1 ) or absence ( sij = 0 ) of allele j at position i . ( or more generally , if an ambiguity in the population sequencing is consistent with several amino acids at a given position , sij denotes the probability of allele j at position i ) . The model parameters I , mij and εij;kl can be interpreted as intercept , main effects , and epistatic effects . As the number of parameters exceeds the number of data-points , the model M1 has been fitted to the data on the basis of a machine learning approach ( generalized kernel ridge regression ) . With this approach over-fitting is no concern because the sub-dataset on which the predictor is evaluated is independent from the sub-dataset from which the predictor is inferred ( see supplementary material of Hinkley et al . [3] for a detailed description of the fitting procedure ) . We assessed the RC-predictor by using two datasets collected from untreated , chronically infected patients . The latter criterion was introduced because HIV RNA levels are usually very high during acute HIV infection , and it was ensured by discarding data points measured within the first 180 days after the first positive HIV test . The patients were enrolled in the Swiss HIV Cohort Study , a longitudinal multicenter observational cohort study ( SHCS ) ( www . shcs . ch ) [4] . These datasets consist of clinical data ( Table 1 ) and the corresponding viral amino acid sequences from the SHCS drug resistance database [5] . We focus on patients , for whom amino-acid sequences of the entire protease and the first 303 amino acids of the reverse transcriptase were available . We only consider sequences , which have been obtained from therapy-naïve patients infected with HIV-1 subtype B because the training set originated solely from subtype B strains . The first set consists of nucleotide sequences with the corresponding HIV RNA virus load measurements ( plasma viral load set; n = 2073 patients ) . Selection of viral load measurements is restricted to values obtained within 30 days before or after the genotypic tests , but before initiation of antiretroviral therapy . The second set contains 53 patients for whom genetic sequences are available at two time points , which are at least 6 months apart ( median [interquartile] distance between the two measurements: 3 . 9 [1 . 9; 7 . 4] years; longitudinal set ) ( see [6] for more details on this dataset ) . Relationships between HIV RNA and pRC were modelled by the use of univariable and multivariable linear regression . Model assumptions were verified by inspecting residual versus fitted plots and by checking for unequal variance across fitted values ( heteroskedasticity ) and outliers . Because these diagnostics suggested the presence of heteroskedasticity we performed “robust” versions of linear regressions , which estimate a weighted variance based on the Huber−White method . Statistical calculations were carried out with Stata 11 . 2 ( Stata Corp . , College Station , TX , USA ) . The level of significance was set at 0 . 05 , and all p-values are two sided .
Demographic and clinical characteristics of our study population are displayed in table 1 . We assessed the predicted RC ( pRC ) with respect to two clinically relevant quantities or processes: Firstly , the relation between pRC and virus-load measurements measured around the same time and , secondly , the temporal change of pRC within ART-naive individuals . In the plasma viral load dataset ( 2073 patients ) , values for RC predictions ( pRC ) were ranging from −1 . 07 to 1 . 43 units ( median [interquartile range] 0 . 62 [0 . 40; 0 . 81] ) , and corresponding median [interquartile] HIV RNA levels were 4 . 7 log10 copies/mL [4 . 1; 5 . 2] . Using univariable linear regression analysis , we find a highly significant effect of the pRC value on virus load ( F−Test p<0 . 001; see Figure 1A ) : a 1 unit increase in pRC is associated with an 0 . 57 increase [95% confidence interval 0 . 45; 0 . 69] in log10 HIV RNA . The fraction of variance in virus load explained through the pRC ( R2 ) is 4 . 4% . Although somewhat attenuated , this effect of pRC on virus load remains highly significant ( p<0 . 001; 0 . 29 [0 . 18; 0 . 40] log10 copies/mL HIV RNA per 1 unit increase in pRC ;table 1 ) if we control in a multivariable regression model for age , ethnicity , risk group , sex , CDC C stage and CD4 count at time of viral sequencing , and the laboratory that generated the sequence data . The association between HIV RNA and pRC changes only minimally when the fully adjusted regression model is re-estimated on individuals without any evidence for transmitted drug resistance mutations as defined by the most recent WHO surveillance list [7] ( n = 1909; regression coefficient [95% confidence interval] 0 . 30 [0 . 18; 0 . 42] log10 copies HIV RNA per unit change pRC ) . For the longitudinal dataset , we find that the pRC value increases in the course of an infection . Among the 53 patients with two viral sequences available taken at least 6 months apart , the median [interquartile] difference in pRC is 0 . 10 units [0 . 04; 0 . 25] and is statistically significantly different from 0 ( p sign rank<0 . 001 ) . Unadjusted linear regression estimates this increase in pRC at 0 . 020 units per year [95% confidence interval 0 . 006; 0 . 035] ( figure 1B ) . At the same time , HIV RNA also tended to be higher at the second , later time point , with a median of 0 . 42 log10 copies/mL [−0 . 28; 0 . 88] ( sign rank p = 0 . 005 ) . Consequently , we find a statistically significant association between the change in pRC correlates and the change in HIV RNA over time in these 53 patients when applying a linear regression model to the data , which predicts a rise of 0 . 90 [0 . 01; 1 . 79] log10 copies/mL in HIV RNA per 1 unit increase in pRC over time ( figure 1C ) . This finding suggests that within-host evolution seems to be characterized by a trend towards higher replication rates , and consequently higher plasma HIV RNA viral loads . The above analyses were based on untreated patients sampled after the acute phase of the infection . We find similar results if we exclude patients , which have been sampled in the AIDS phase ( defined as patients with at least one CDC stage C event , n = 206 ) . In particular , we still find a highly significant ( p<0 . 001 ) correlation between pRC and RNA load ( slope: 1 unit increase in pRC is associated with an 0 . 54 increase [95% confidence interval 0 . 41; 0 . 66] in log10 HIV RNA ) and a significant ( p = 0 . 0058 ) increase of RC over time ( increase in pRC at 0 . 020 units per year [95% confidence interval 0 . 006; 0 . 035] ) . Only the significance-level of the correlation between the temporal change of pRC and the temporal change of RNA load changes from ‘significant’ ( p = 0 . 04 ) to ‘trend’ ( p = 0 . 058 ) ; however even in this case the point estimates for the regression coefficient are very similar in both cases ( 0 . 9[0 . 01; 1 . 79] vs . 0 . 84[−0 . 03; 1 . 70] ) .
How do the pRCs analyzed here relate to previous findings ? For example , the 6 sequences ( in our data-set ) carrying the lamivudine mutation M184V , which has a large negative fitness effect on the virus [8] and has been associated with an 0 . 3 log10 copies lower HIV RNA relative to wild type [9] , had a median [interquartile range] pRC of 0 . 1 [−1 . 3; 0 . 6] , compared to 0 . 6 [0 . 4; 0 . 8] in the 1909 sequences without any transmitted resistance mutations ( Wilcoxon rank sum p<0 . 001 ) . Overall , the pRC varied over a range of 2 . 5 units from minimum to maximum . Our unadjusted and adjusted regression models would therefore predict a difference in HIV RNA of approximately 1 . 4 and 0 . 73 log10 copies/mL between the lowest and the highest pRC value . Yet HIV RNA viral loads varied over 6 logs from 1 . 9 to 7 . 9 log10 copies/mL in our dataset . This discrepancy is not very surprising given that our predictor for RC only takes the variation of 400 amino acid positions ( roughly 10% of the genome of HIV ) into account . However , the finding of a correlation of pRC and HIV RNA is robust , as confirmed by several sensitivity analyses , and it is consistent with a number of previous studies , which have also shown a correlation between in vitro measurements of RC and virus load [10] , [11] , [12] , [13] , [14] . Our findings thus support the notion that virus load is to a large extent controlled by virus genetics [15] , [16] , [17] . The fraction of variance explained by pRC ( 4 . 4% ) is much lower than the fraction of variance in virus load explained by virus genetics in previous studies [15] , [16] , [17] , but it should be borne in mind that the estimates of studies [15] , [16] , [17] are based on the variation in the entire genome ( Note that this is the case even for Alizon et al . [15] , because , even though the phylogenies used in that study were inferred from the pol-gene , they reflect the relatedness of the entire genome provided that recombination is not too common on an epidemiological level ) . It should also be noted that our results argue that at least a part of the virus' genetic control of the virus load established in patients appears to be mediated by the replicative capacity of the virus . This finding that virus load is controlled by RC contrasts the interpretation that virus load is mainly determined by the activation-rate of CD4 cells[18] . However , the relative importance of these different factors remains an open question . The increase of pRCs over time is also consistent with previous observations [19] , and supports the view that , within a single host , HIV is selected for higher replicative capacities over time . Overall our results show on the basis of a computational predictor , firstly that in vitro replicative capacity increases in the course of infection , which is consistent with the interpretation that RC is a determinant of fitness at the within-host level , and secondly that RC is linked to virus load , which has been shown to be a in vivo determinant of viral fitness at an epidemiological level [1] . In our view , it is remarkable that predicted RC based on partial pol sequences representing only 10% of HIVs genome correlates with virus load . Accordingly , taking into account the variation in the entire HIV genome ( as will become possible in the future ) may help to develop much more accurate predictors of virus fitness and virus load .
|
Determining how well different genotypes of HIV can replicate within a patient is central for our understanding of the evolution of HIV . Such in vivo fitness is often approximated by in vitro measurements of viral replicative capacities . Here we use a machine-learning algorithm to predict in vitro replicative capacities from HIV nucleotide sequences and compare these predicted replicative capacities with clinical data from HIV-infected individuals . We find that predicted replicative capacity correlates significantly with the concentration of HIV RNA in the plasma of infected individuals ( virus load ) . Furthermore , we show that the predicted replicative capacity increases in the course of an infection . Finally , we found that the temporal increase of replicative capacity correlates significantly with the temporal increase of virus load within a patient . These results indicate that ( predicted ) replicative capacity is a useful measure for viral fitness and suggest that virus genetics determines virus load at least to some extent via replicative capacity .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"genetics",
"biology",
"microbiology",
"evolutionary",
"biology",
"population",
"biology",
"genetics",
"and",
"genomics"
] |
2011
|
Assessing Predicted HIV-1 Replicative Capacity in a Clinical Setting
|
In order to eliminate viral infections , hundreds of interferon-stimulated genes ( ISGs ) are induced via type I interferons ( IFNs ) . However , the functions and mechanisms of most ISGs are largely unclear . A tripartite motif ( TRIM ) protein encoding gene TRIM69 is induced by dengue virus ( DENV ) infection as an ISG . TRIM69 restricts DENV replication , and its RING domain , which has the E3 ubiquitin ligase activity , is critical for its antiviral activity . An in vivo study further confirmed that TRIM69 contributes to the control of DENV infection in immunocompetent mice . Unlike many other TRIM family members , TRIM69 is not involved in modulation of IFN signaling . Instead , TRIM69 interacts with DENV Nonstructural Protein 3 ( NS3 ) directly and mediates its polyubiquitination and degradation . Finally , Lys104 of NS3 is identified as the target of TRIM69-mediated ubiquitination . Our study demonstrates that TRIM69 restricts DENV replication by specifically ubiquitinating a viral nonstructural protein .
Recently , mosquito-borne viral diseases become global threats to human health . As the most significant mosquito-borne viral pathogen , Dengue virus ( DENV ) is responsible for outbreaks of dengue fever ( DF ) , dengue shock syndrome ( DSS ) , and dengue hemorrhagic fever ( DHF ) . DENV causes millions of infections in over 100 countries annually , resulting in more than 25 , 000 deaths [1 , 2] . A DENV vaccine was recently licensed for use after several decades of efforts , however , it confers only partial cross protection for all DENV serotypes [3 , 4] . Additionally , there is still no antiviral drugs have been approved to treat DENV induced diseases [5 , 6] . Similar with other mosquito borne flaviviruses , DENV genome is quickly translated into a polyprotein after entering into the host cells . Then the viral polyprotein is cleaved into three structural proteins ( capsid protein C , membrane protein M , and envelope protein E ) and seven nonstructural proteins ( NS1 , NS2A , NS2B , NS3 , NS4A , NS4B and NS5 ) [7–10] . The nonstructural proteins are not the component of the mature , infectious virions , but involved in polyprotein processing , viral RNA synthesis , and virus morphogenesis [11] . NS3 , a multifunctional protein , has a superfamily 2 ( SF2 ) DEAH-box helicase domain that possesses RNA 5’-triphosphatase ( RTPase ) , RNA-stimulated nucleoside triphosphatase ( NTPase ) , RNA annealing , and 3’-tailed dsRNA unwinding activities [12–17] . It forms a protease complex together with NS2B , helps to cleave the DENV polyprotein and many other proteins , such as STING [18–20] . NS5 , containing an N-terminal methyltransferase and a C-terminal RNA-dependent RNA polymerase , is indispensable for viral RNA replication [21 , 22] . NS4B was recently reported to promote DENV replication and alleviate RIG-I dependent activation of interferon responses by induction of mitochondria elongation [23] . The interaction of NS3 and NS5 is important for DENV RNA replication [24 , 25] . NS4B also interacts with NS3 , facilitating dissociation of NS3 helicase from ssRNA [26–28] . Due to its central role in DENV replication , NS3 is an intriguing target for anti-DENV therapy . Although the functions of other nonstructural proteins remain unclear , they do indeed play important roles in viral replication , assembly and maturation [29] . The tripartite motif family members ( TRIMs ) share three conserved domains , an N-terminal Really Interesting New Gene ( RING ) domain , one or two B-Boxes ( B1/B2 ) and a coiled-coil ( CC ) domain . TRIM proteins are implicated in multiple cellular functions , ranging from transcriptional regulation to post-translational modifications involved in various cellular processes , such as cell differentiation , apoptosis and oncogenesis [30] . TRIM proteins have been long predicted to be part of the innate immune pathway . In line with this , recent studies show that an increasing number of TRIM proteins are recognized as ISGs and mediate antiviral activities [31–33] . The antiviral activities of TRIM proteins depend , for the most part , on their function of E3-ubiquitin ligases activity . TRIM38 sumoylates cGAS and STING during the early phase of virus infection to promote the stability of these two proteins [34] . TRIM56 inhibits bovine viral diarrhea virus ( BVDV ) replication by targeting intracellular viral RNA replication [35] . TRIM5a is responsible for post-entry restriction of diverse retroviruses , including N-MLV and HIV-1 [32 , 36] . TRIM5α blocks HIV-1replication by targeting the capsid and promoting its rapid , premature disassembly [37] . At the same time , TRIM5α stimulates innate immune signaling by catalyzing the synthesis of unanchored K63-linked poly-ubiquitin chains that bind and activate TAK1-dependent NF-κB [32] . TRIM22 inhibits HIV-1 by down-regulating the viral long terminal repeat-directed transcription [38 , 39] . Some TRIM proteins restrict viral replication by directly targeting viral proteins . TRIM22 has been reported to interact with HIV-1 Gag protein , EMCV 3C protease , influenza virus nucleoprotein and HCV NS5A , resulting in inhibition of viral replication [40–43] . TRIM79α inhibits tick-borne encephalitis virus via targeting the viral RNA-dependent RNA polymerase , NS5 , for lysosomal degradation [44] . Although knowledge on the cellular roles of TRIM E3 ubiquitin ligases has rapidly grown over the last years , many aspects of their molecular functions remain unclear . Here , we identified another TRIM family member , TRIM69 ( also known as RNF36 , HSD34 , and Trif ) as an IFN-inducible virus restriction factor . As an E3 ubiquitin ligase [45] , TRIM69 plays crucial roles in apoptosis [46] , tumor control [47] and zebrafish development [48 , 49] . However , TRIM69 has not been reported to have any function on antiviral immunity . In this study , we demonstrated that TRIM69 is an IFN-stimulated gene and restricts DENV replication in vitro and in vivo . TRIM69 directly interacts with viral NS3 and results in NS3 degradation by proteasomes . Thus , TRIM69 is a novel IFN inducible restriction factor for DENV .
To evaluate the mechanisms of how host cells resist a pathogenic microorganism , RNA-Seq was performed to screen out the host factors involved in DENV-2 infection . 152 mRNAs were significantly changed after DENV infection in 293T cells ( 99 mRNAs were induced , while the others were decreased ) . As expected , many genes related to antiviral innate immune signal pathway were found out as shown by gene cluster analysis ( S1 Fig ) . Furthermore , 52 of the 99 upregulated genes upon DENV-2 infection were predicted as ISGs by searching the Interferome V2 . 01 database ( www . interferome . org ) . Many well-known ISGs , such as DDX58 , IRF9 , ISG15 and STAT1 , were screened out after DENV-2 infection ( S1 Table ) . The mRNA expression of TRIM69 , together with five other putative ISGs ( LGALS3BP , C19ORF66 , DDX60 , FBXO15 , and HELZ2 ) was confirmed to be significantly upregulated after DENV-2 infection ( Fig 1A , S2A Fig and S1 Table ) . The protein level of TRIM69 was also increased with DENV-2 infection in a virus dose-dependent manner ( Fig 1B ) . Consistent with this , TRIM69 was also upregulated in A549 , HUVEC and PBMC cells infected with DENV-2 ( Fig 1C ) . In addition , the expression of TRIM69 was increased in peripheral blood cells from DENV-2-infected mice ( Fig 1D ) . When HUVEC and HFF cells were stimulated with SeV , TRIM69 was also upregulated ( Fig 1E ) . A previous study reported the expression of TRIM family genes in response to interferons in immune cells . TRIM69 was identified as one of 27 TRIM genes which were induced by interferons [50] . Consistent with their results , we also found that the expression of TRIM69 mRNA and protein were induced in 293T , HUVEC and HFF cells upon IFN-β stimulation ( Fig 1F and 1G ) . Four of other five selected ISGs , LGALS3BP , C19ORF66 , DDX60 , and HELZ2 , were also induced in 293T cells stimulated with IFN-β ( S2B Fig ) . Taken together , these data further confirmed that TRIM69 is an ISG induced by type I IFN and virus infection . To explore the function of TRIM69 on DENV replication , cells were transfected with TRIM69-Myc and then infected with DENV-2 . Transient overexpression of TRIM69 did not cause noticeable cell toxicity at 72 h post transfection ( Fig 2A ) . qRT-PCR and Western Blot results suggested that the viral RNA and proteins were significantly decreased in TRIM69 overexpressed cells compared with control cells ( Fig 2B and 2C ) . Immunofluorescence ( IF ) assay also confirmed that the viral NS3 and NS4B protein levels were significantly decreased in TRIM69 overexpressed cells ( Fig 2D ) . Consistently , the released viruses in cell supernatants were also decreased in TRIM69 overexpressed cells ( Fig 2E ) . In addition , we also used a luciferase-based DENV replicon ( DGL2 ) , derived from DENV-1 [51 , 52] , to analyze the function of TRIM69 on virus replication . The replicon replication was also impaired by TRIM69 ectopic expression ( Fig 2F ) . To further confirm this phenotype , two TRIM69 knockdown shRNAs ( sh69-1 and sh69-2 ) were constructed . Silencing TRIM69 by shRNA transfection did not influence cell viability ( Fig 2G ) . The abundance of both DENV NS3 and NS4B proteins was significantly increased in two TRIM69 shRNAs transfected 293T cells after DENV-2 infection ( Fig 2H ) . The virus titers in cell supernatants were also increased in TRIM69 silenced cells ( Fig 2I ) . In addition , a TRIM69 knockout stable cell line was generated by CRISPER/Cas9 system , and the replication of DGL2 replicon was significantly increased in TRIM69 knockout cells compared with controls ( Fig 2J ) . Furthermore , the replication of DGL2 was also increased in TRIM69 silenced Huh7 . 0 stable cell line ( Fig 2K ) . Since TRIM69 is induced by IFNs and plays a role to restrict DENV infection , we wondered whether TRIM69 is critical for the efficacy of IFN on DENV inhibition . Viral replication assays suggested that IFN-β treatment could not efficiently suppress DGL2 replication in TRIM69 silenced cells ( Fig 2L ) . This demonstrates that TRIM69 is critical for IFN mediated anti-DENV activity . To investigate whether mouse TRIM69 has the same function as its human homolog , mTRIM69 was overexpressed or silenced in mouse B16F10 cells . The results suggested that viral proteins expression and viral titers of DENV were significantly decreased in mTRIM69-myc transfected B16F10 cells compared with controls ( S3A Fig ) . Furthermore , DENV infection was also obviously increased in mTRIM69 silenced cells ( S3B and S3C Fig ) . Altogether , these data illustrate that both human and mouse TRIM69 protein acts as antiviral factors for DENV replication . TRIM69 is a TRIM family member containing a RING domain with E3 ubiquitin ligase activity . We next tested whether the E3 ubiquitin ligase activity of TRIM69 is necessary for DENV inhibition . TRIM69 CA , a mutant TRIM69 with the catalytic amino acids Cys61 and Cys64 of the RING domain substituted by two Alanines , loses its E3 ubiquitin ligase activity [53] . Cell lines stably expressing TRIM69-Flag and TRIM69 CA-Flag were generated using pLV-Flag vector by antibiotics selection . Western Blots confirmed that stable cell lines expressed higher level of TRIM69 ( or TRIM69 CA ) compared with endogenous TRIM69 ( Fig 3A ) . After DENV-2 infection , the abundance of viral NS4B , as shown by Western Blots ( Fig 3A ) and IF assay ( Fig 3B ) , was decreased in TRIM69 expressing cell line , but not in TRIM69 CA cells . The virus titers from the cells stably expressing TRIM69 were lower than the control or TRIM69 CA ( Fig 3C ) . In line with this , DGL2 replication was also impaired in TRIM69 , but not TRIM69 CA , overexpressed cells ( Fig 3D ) . When the cells treated with MG132 , a proteasome inhibitor , the content of NS4B was recovered in TRIM69 overexpressed cell after DENV infection ( Fig 3E ) . These results indicate that the E3 ubiquitin ligase activity of TRIM69 is critical for its antiviral activity . Previous study suggested DENV causes a transient infection in immunocompetent mice with detectable virus in various organs [54] . To explore the function of TRIM69 on DENV in vivo , shm69-1 and shNC lentiviruses were generated and used to inject into mice via caudal vein . 7 days post lentivirus infection , mice were challenged with DENV-2 by intravenous injection . qRT-PCR and western blot suggested that mouse TRIM69 was silenced by shm69-1 lentiviruses in mouse lung , spleen and kidney ( Fig 4A and 4B ) . In consistent with the in vitro data , both DENV RNA level ( Fig 4C ) and virus titers ( Fig 4D ) were significantly increased in organs from TRIM69-silenced mice . These data further confirmed that TRIM69 is an important host antiviral factor against DENV in vivo . To test whether TRIM69 also restrict other virus infection , TRIM69 overexpressing or control cells were infected with influenza virus H1N1 ( an RNA virus ) or herpes virus HSV-1 ( a DNA virus ) , respectively . The results suggested that TRIM69 did not interfere with H1N1 or HSV-1 infection ( S4A and S4B Fig ) . TRIM69-silenced mice showed similar susceptibility with wide type mice to H1N1 infection ( S4C and S4D Fig ) . These data suggest that TRIM69 may play a specific antiviral activity against DENV . Several members of TRIM family proteins were reported to restrict viral replication by modulate the IFN pathways . We next tested whether TRIM69 is involved in IFN or ISG activation . Results suggested that overexpressing or silencing TRIM69 did not significantly influence SeV-induced IFN or ISG production ( S5 Fig ) . This is also consistent with previous report by Versteeg G et al . , that TRIM69 does not modulate either IFN production or ISG expression [33] . To further elucidate the mechanisms of TRIM69 on DENV inhibition , immunoprecipitation and mass spectrometry ( IP-MS ) were performed to find out proteins that interact with TRIM69 during DENV infection ( Fig 5A and S6A Fig ) . Three viral proteins , NS3 , NS4B and NS5 , were pulled down by TRIM69-Flag coupled beads but not with beads alone ( S6B Fig ) . Three peptides of NS3 , one peptide of NS4B , and one peptide of NS5 were identified by IP-MS ( Fig 5B ) . These three viral proteins were co-expressed in 293T cells with or without TRIM69-Myc . The result suggested , the abundance of NS3 , but not NS4B or NS5 , was significantly reduced in TRIM69 overexpressed cell compared with controls ( Fig 5C ) . Moreover , the ectopic expression of NS3 was increased via TRIM69 knockdown by sh69-2 ( Fig 5D ) . This suggests that NS3 is a target for TRIM69 . NS3 forms a protease complex with NS2B , not only responsible for cleavage of viral polyprotein , but also for immune evasion [19 , 20] . NS2B3 could specifically cleave human STING , play a role to escape STING mediated antiviral pathway . We then tested whether TRIM69 also influences the cleavage activity of NS2B3 on STING . We found that TRIM69 significantly reduced the amount of NS2B3 protein , thereby impaired the cleavage of STING ( S7A Fig ) . These data further suggest that TRIM69 targets NS3 and modulates NS3 function . To confirm the interaction between DENV NS3 and TRIM69 , the cellular distribution of NS3 and TRIM69 was examined by confocal microscopy . When co-expressed with TRIM69 , NS3 is re-distributed from a predominantly diffuse cytoplasmic localization to punctate sites co-localizing with TRIM69 ( Fig 6A ) . The co-localization of TRIM69 and NS3 was specific , as another viral protein , NS4B , did not co-localize with TRIM69 ( Fig 6A ) . Co-IP assays were performed to further confirm the physical interaction between TRIM69 and NS3 . IP of NS3 with Flag antibody successfully coprecipitated TRIM69-Myc ( Fig 6B ) . Likewise , the reciprocal test using Myc antibody could immunoprecipitate TRIM69 with NS3 ( Fig 6C ) . Furthermore , endogenous TRIM69 also interacted with NS3 ( Fig 6D and 6E ) from DENV infected cells . Finally , a GST pulldown assay also confirmed that purified TRIM69 protein interacts with GST-NS3 directly ( Fig 6F ) . The interaction between mTRIM69 and NS3 was also investigated in mouse cells . mTRIM69 also co-localized and interacted with NS3 in B16F10 cells ( S8A and S8B Fig ) . These data suggest that TRIM69 interacts with DENV NS3 . Since TRIM69 is an E3 ligase , we next investigated whether NS3 is ubiquitinated by TRIM69 . As shown in Fig 7A , overexpressing TRIM69 , but not TRIM69 CA , led to NS3 degradation; however , this degradation was blocked by MG132 . When the ectopically expressed NS3 was immunoprecipitated by Flag , we observed ubiquitination modifications on NS3 , and the ubiquitination of NS3 was obviously increased in the presence of TRIM69-Myc , but not of TRIM69 CA-Myc ( Fig 7B ) . We also detected more endogenous ubiquitin conjugated to NS3 in the presence of TRIM69 , but not TRIM69-CA ( Fig 7C ) . Consistent with this , the ubiquitin ligated to NS3 was significantly reduced when TRIM69 was knockdown ( Fig 7D ) . Finally , an in vitro ubiquitination assay further confirmed that TRIM69 can directly ubiquitinate NS3 in the presence of ubiquitin E1 and E2 in a cell-free system ( Fig 7E ) . DENV-2 NS3 contains 46 lysine residues . Seven ( Lys15 , Lys90 , Lys104 , Lys170 , Lys489 , Lys515 , and Lys584 ) of these were predicted to be potential ubiquitination sites by the UbPred program ( http://www . ubpred . org/ ) . To determine the NS3 ubiquitination sites by TRIM69 , we replaced each of the seven NS3 lysine residues noted above individually with arginine . Immunoprecipitation with anti-Flag and immunoblot analysis of ubiquitin demonstrated that K104R substitution significantly decreased the ubiquitination of NS3 by ectopically expressed TRIM69 ( Fig 8A ) . Furthermore , NS3 WT , K90R , and K104R were transfected into 293T cells together with or without TRIM69-Myc . The immunoblot analysis showed that the expression of NS3 WT and K90R were significantly reduced via TRIM69 ectopic expression , however , K104R was not ( Fig 8B ) . To further confirm the Lys104 is a potential ubiquitination site of NS3 for TRIM69 , we constructed a mutant DENV-1 DGL2 replicon ( NS3-K104R ) via site-directed mutagenesis , in which the Lys104 of NS3 being replaced by arginine . DGL2 NS3-WT and NS3-K104R were transfected into 293T cells individually together with or without TRIM69-Myc . The results revealed that the replication of DGL2 reduced with TRIM69 ectopic expression , however , the replication of NS3-K104R did not ( Fig 8C ) . All the data illuminate that Lys104 is an ubiquitination site of NS3 for TRIM69 .
This work has illustrated a TRIM family member , TRIM69 , as a key host factor needed to restrict DENV infection . TRIM69 mRNA was found to be overexpressed in 293T cells infected with DENV-2 as determined by RNA-Seq analysis . A significant amount of differential expressed genes found in virus infected cells have been described as signaling pathway molecules involved in antiviral innate immunity ( S1 Fig ) . 52 out of the 99 upregulated genes are predicted ISGs , such as TRIM69 , LGALS3BP , C19ORF66 , DDX60 , and HELZ2 ( S2 Fig ) . All of these putative ISGs were upregulated after DENV infection ( S2 Fig ) . Consistent with our results , two recent studies also report that both C19ORF66 and HELZ2 are induced by IFN and suppress DENV replication [55 , 56] . A previous report identified that TRIM69 is induced in peripheral blood cells upon type I IFN stimulation [50] , here we show that TRIM69 is also upregulated in 293T , HUVEC , and HFF cells by IFN-β stimulation or virus infection ( Fig 1 ) , and has antiviral properties against DENV infection . Many TRIM family proteins are involved in regulating signaling pathways such as Toll-like receptors ( TLRs ) and RIG-I-like receptors ( RLRs ) which are needed for viral detection and innate immune responses [57] . For example , TRIM12c interacts with TRAF6 , leading to a cooperative activation of IFN and NF-κB pathways [58] . TRIM38 negatively regulates TLR3/4-mediated innate immune and inflammatory responses [59] . TRIM13 acts as a negative regulator for MDA5-mediated type I interferon production [60] . Since TRIM69 is an ISG , we also tested whether or not it participates in IFN-induced signal pathway . Unlike other reported TRIM family members , TRIM69 did not influence SeV-induced IFN-β production or IFN-β/SeV induced ISRE promoter activation , which were consistent with the findings from a previous screening [33] . They found roughly half of the 75 TRIM family members modulated the interferon response , but TRIM69 was not in the list [33] . These results suggest that TRIM69 has no influence on IFN production or IFN function . In line with this , we also found that TRIM69 did not influence other virus infection , such as H1N1 or HSV-1 ( S4 Fig ) . These results suggest that TRIM69 may use a specific mechanism to restrict DENV infection , independent of interferon pathway . Some TRIM proteins have been demonstrated to have direct antiviral activity , including TRIM5α , TRIM22 , and TRIM79α [41 , 44 , 61–64] . To further investigate the mechanism of TRIM69 on inhibiting DENV replication , IP-MS analysis was performed to search for host and viral proteins interacting with TRIM69 . DENV NS3 was found directly interacts with TRIM69 and degraded via TRIM69 ectopic expression ( Fig 5 ) . Then , we tested whether TRIM69 also influences the function of NS3 . A recent study suggested that a K27-linked ubiquitination of NS3 enhance the interaction of NS3 and NS2B , thereby promotes the cleavage STING by NS2B3 complex[65] . We found that overexpression of TRIM69 impaired the cleavage of STING by NS2B3 ( S7A Fig ) . This is reasonable , since TRIM69 targets NS3 to degradation , NS2B3 level will also be decreased ( S7A Fig ) . While , we found TRIM69 seems not influence the interaction efficiency of NS2B and NS3 ( S7B Fig ) . A possible reason is that our other preliminary experiments suggest TRIM69 may influence K11-linked ubiquitination of NS3 , rather than previously reported K27-linked ubiquitination [65] . And K11-linked poly ubiquitination can mediates protein degradation in a proteasome dependent manner [66] . Further experiments will be required to address the detailed ubiquitination form on NS3 mediated by TRIM69 . In this study we found , TRIM69 acts as an IFN-β-stimulated ISG and has antiviral activity via its RING domain . As an E3 ubiquitin ligase , TRIM69 was reported to restrict DENV replication by direct ubiquitination of NS3 which leads to NS3 degradation . The viral protease NS3 is highly conserved throughout the Flavivirus genus and necessary for viral replication and immune evasion [67 , 68] . We next will further investigate whether TRIM69 acts as a broad-spectrum restriction factor for all the closely related mosquito-borne flaviviruses .
The HUVEC ( Human Umbilical Vascular Endothelium Cells ) and PBMC ( human Peripheral Blood Mononuclear Cells ) were obtained from BeNa Culture Collection ( Bejing , China ) . All samples were anonymized and the projects using of human biological specimens were approved by an institutional review board ( IRB ) of Soochow University . Animal experiments were conducted according to the Guide for the Care and Use of Medical Laboratory Animals ( Ministry of Health , People’s Republic of China ) and approved by the Animal Care and Use Committee as well as the Ethical Committee of Soochow University ( SYSK- ( S2012-0062 ) ) . 293T , Vero , HeLa , A549 , Huh7 . 0 , and HFF cells were obtained from ATCC ( Manassas , USA ) and grown in DMEM ( Life Technologies , Grand Island , USA ) supplemented with 10% FBS and antibiotics/antimycotics . HUVEC , PBMC and B16F10 cells were grown in 1640 ( Life Technologies ) supplemented with 10% FBS and antibiotics/antimycotics . DENV type 2 ( DENV-2 ) New Guinea C ( NGC ) strain was propagated in mosquito C6/36 cells ( ATCC CRL-1660 ) . Cells were infected with DENV at a multiplicity of infection ( MOI ) of 1 , unless otherwise stated . Influenza A virus ( H1N1-A/PR/8/34 ) and Sendai virus ( SeV ) was propagated in 10 days old embryonated eggs ( Bejing Laboratory Animal Research Center , Beijing , China ) , and the virus titer was detected by hemagglutination assay using chicken red blood cells ( BeNa Culture Collection ( Beijing , China ) ) . Human herpesvirus 1 ( HSV-1 ) was propagated in Vero cells . The following antibodies were purchased from Cell Signaling Technology ( CST , Danvers , USA ) , rabbit anti GAPDH ( Cat # 2118 ) , mouse anti Flag ( Cat # 8146 ) , mouse anti HA ( Cat # 2367 ) and anti-rabbit IgG HRP-linked Antibody ( Cat # 7074 ) . Others were obtained as follows; mouse anti Myc ( Cat # M20002 Abmart , Shanghai , China ) , anti-Flag M2 Affinity Gel ( Cat # A2220 Sigma-Aldrich , St Louis , USA ) , rabbit polyclonal anti DENV2 NS3 ( Cat # PA5-32199 Thermo Fisher Scientific , Waltham , USA ) , rabbit polyclonal anti DENV2 NS4B ( Cat # GTX113374 GeneTeX , Irvine , USA ) , rabbit polyclonal anti ZIKV NS3 ( Cat # GTX133309 GeneTeX , Irvine , USA ) , rabbit anti TRIM69 ( RNF36 , Cat # ab111943 Abcam , Cambridge , UK ) , and HRP Goat anti-mouse IgG ( Cat # 405306 Biolegend , San Diego , USA ) . Lipofectamine 2000 was purchased from Life Technologies . MG132 , puromycin , and dimethylsulfoxide ( DMSO ) were purchased from Sigma-Aldrich . N-ethylmaleimide ( NEM ) was obtained from Thermo Fisher Scientific . Protease inhibitor ( PI ) was from CST . Recombinant human IFN-β was from PeproTech ( Rocky Hill , USA ) . TRIM69-Flag and TRIM69 CA-Flag were kindly provided by Dr . Linfang Wang ( Tsinghua University , Beijing , China ) [53] . To construct TRIM69-Myc , TRIM69 ORF was amplified from TRIM69-Flag and cloned into the Xho I and EcoR I sites of the pCMV-Myc vector . TRIM69 and TRIM69CA ORFs were also constructed into a puromycin resistant vector pLV-Flag for stable cell-line construction . To construct sh69-1 and sh69-2 , target sequences ( 5’-GGGAAACTGATCTGCTTTC-3’ or 5’-GGACAAGTTGGTAGAGAAG-3’ ) were inserted into vector RNAi-Ready pSIREN-RetroQ-ZsGreen and Lenti-U6-shRNA-GFP-puro individually . To construct shm69-1 and shm69-2 , target sequences ( 5’-GCTCGTGGAGAAGATTAAGAA-3’ or 5’-CGTTTCTTTACGGAGGAGCTT-3’ ) were inserted into vector Lenti-U6-shRNA-GFP-puro individually . For construction of CRISPR/Cas9 construct of TRIM69 , the gRNA sequence ( 5’-GGCTCAAGAGGCTTCACCTC-3’ ) was cloned into pX462 . All NS genes were amplified from cDNA of DENV-2 NGC strain . A DNA-based DENV replicon DGL2 ( for DENV type 1 ) was generously provided by Dr . Takayuki Hishiki ( Kyoto University , Kyoto , Japan ) [51] . 293T cells were transfected with pLV-TRIM69-Flag or pLV-TRIM69CA-Flag , and selected with puromycin ( 2 μg/mL ) for at least 3 weeks . The overexpressions of TRIM69 and TRIM69CA in selected stable cell lines were confirmed by Western Blot . Similarly , shTRIM69 ( U6-shRNA-GFP-puro ) and TRIM69-/- cell line ( pX462 ) were constructed by puromycin selection followed by single cell clone culture and Western Blot identification . Total RNA was collected from DENV-2 infected ( or non-infected ) 293T cells at 48h post-infection . cDNA libraries were prepared through the sequential use of the RNeasy Mini Kit with On-Column DNase Digestion Set ( QIAGEN , Venlo , Netherlands ) , Dynabeads mRNA DIRECT purification Kit and Total RNA-Seq Kit v2 ( Thermo Fisher Scientific ) . The transcription sequences were sequenced using an Illumina Hiseq2000 , and the total base number was more than 20 Gb per sample . RNA-Seq de novo assembly was performed using Trinity . Get ORF in EBOSS were used to find protein from contigs . 100 ng expression plasmid , 50 ng IFN-β-Luc/ISRE-Luc , and 10 ng pRL-TK ( internal control ) were co-transfected into 293T cells plated in 96-well plates . Then the cells were treated with SeV infection or IFN-β ( 200 U/ml ) stimulation where indicated . 24 hours later , cells were harvested and the DLR assays were performed with a luciferase assay kit ( Promega , Madison , WI ) . All reporter assays were completed at least in triplicate , and the results were shown as average values ±standard deviations ( SD ) from one representative experiment . In 96-well plates , 50 ng of DGL2 replicon plasmid was transfected into 293T cells stably expressing TRIM69 ( or TRIM69 CA ) or TRIM69 silenced ( shTRIM69 or TRIM69-/- ) cells . For the Gaussia luciferase assay , culture supernatants were collected at different time points and luciferase was measured using BioLux Gaussia Luciferase Assay Kit ( New England Biolabs ) according to manufacturer’s instructions . Hela cells were transfected with or without the plasmid TRIM69-Myc . The cells were then infected with DENV-2 for 24 h . Cells on coverslips were fixed in 4% formaldehyde for 10 min , sequentially permeabilized with 0 . 4% Triton X-100 , blocked with 5% FBS , incubated with primary antibodies ( rabbit anti-NS3 , rabbit anti-NS4B and mouse anti-Myc ) at 4 °C overnight , and incubated with TRITC-Goat anti-mouse IgG ( H+L ) ( Jackson , Cat # 115-025-003 ) and Goat anti-Rabbit IgG-FITC ( Southern Biotech , Cat # 4030–02 ) for 1 h at room temperature . Nuclei were counter stained with DAPI ( 0 . 5μg/ml ) . Finally , the images were obtained by confocal microscopy . For co-localization study , plasmids with NS3-Flag or NS4B-Flag were transfected into Hela cells together with or without hTRIM69-Myc . To investigate the co-localization of mTRIM69 and NS3 , NS3-Flag was transfected into mouse B16F10 cells together with or without mTRIM69-Myc . All the cells were treated with MG132 ( 20 μM ) for 4 h before fixation . Total RNA from the indicated cells treated by different treatments were extracted using the total RNA kit I ( OMEGA ) and reverse-transcribed using the PrimeScript Master Mix kit ( TaKaRa ) . The resulting cDNAs were mixed with RT-PCR primers and SYBR Premix Ex Taq II ( TaKaRa ) and amplified for 40 cycles ( 95 °C 15 s , 60 °C 30 s , and 72 °C 15s ) . The qPCR primers for human TRIM69 , and human β-actin were listed below: TRIM69 Forward: 5’-TCTGTGGGGCAGTCTAAGGA-3’ , Reverse: 5’-CCATGGACACATGTTGCTGC-3’; and β-actin Forward: 5’-GGGCATGGAGTCCTGTGGCA , Reverse: 5’-GGGTGCCAGGGCAGTGATCTC-3’ . The qPCR primers for mouse Trim69 , and mouse β-actin were listed below: Trim69 Forward: 5’-GAGGAGATGGAGGTGAATC-3’ , Reverse: 5’-TTGTGATGTCTGTGAGGAA-3’; and β-actin Forward: 5’-CGTTGACATCCGTAAAGAC-3’ , Reverse: 5’-GAGCCAGAGCAGTAATCT-3’ . All the qPCR results are represented as relative fold changes after normalized to β-actin controls . The titers of DENV-2 in cell-free supernatants were determined with a median tissue culture infective dose ( TCID50 ) assay according to standard protocols on Vero cells [69] . Briefly , Samples were serially diluted and inoculated into Vero cells in 96-well plates . After 5-day incubation , cells were examined for cytopathic effects ( CPE ) under a light microscope . The virus titer ( TCID50/ml ) was calculated using the Reed-Muench method . 1 TCID50/ml was equivalent to 0 . 69 pfu/ml [69 , 70] . TRIM69-Flag plasmid was transfected into 293T cells and then infected with DENV-2 . The cells were treated with MG132 ( 20 μM ) for 4 h before lysed with RIPA buffer ( 25 mM Tris•HCl pH 7 . 4 , 150 mM NaCl , 1% NP-40 , 1 mM EDTA , 5% glycerol ) together with Protease Inhibitors ( CST ) . Samples were centrifugated for 10 min to remove cellular debris . The lysates were incubated with Flag Ab conjugated agarose beads ( Sigma-Aldrich ) overnight at 4 °C . After immunoprecipitation , proteins were separated on SDS–PAGE gels ( Invitrogen ) and stained with coomassie blue staining . Gel slices were excised and proteins were reduced with 10 mM DTT prior to alkylation with 55 mM iodoacetamide . Peptides were extracted and analyzed by nano-LC-MS/MS ( ekspertnanoLC , TripleTOF 5600-plus , AB Sciex , USA ) . For co-immunoprecipitation ( Co-IP ) assays , NS3-Flag was transfected together with or without human or mouse TRIM69-Myc constructs . The lysate was incubated with Myc Ab overnight at 4 °C . Then protein A/G was added into the lysate and incubated for 4 hours . The beads were then washed four times and western blot analysis was performed to detect NS3 and TRIM69 . TRIM69 antibody and NS3 antibody were used to in the Co-IP of endogenous TRIM69 and DENV NS3 . For GST pulldown assay , recombinant GST-NS3 and GST control were incubated with immunoprecipitation purified TRIM69-Myc protein . The proteins pulled out by GST agarose were analyzed by western blots . NS3-Flag and HA-Ub were co-transfected into 293T cells together with or without TRIM69-Myc . The cells were then treated with MG132 ( 20 μM ) for 4h and lysed by RIPA buffer with PI and NEM . All the samples were heated at 95 °C for 5 min prior to affinity purification in 1% SDS to remove NS3 interacting proteins . Then the Flag Ab conjugated agarose beads were added into the samples separately . Following incubation overnight at 4 °C , the samples were examined via western blotting . In vitro ubiquitination assay was performed using an E3 Ligase Auto-Ubiquitylation Assay Kit ( Abcam ) according to manufacturer’s instructions . Briefly , immunoprecipitated NS3 were incubated with purified recombinant TRIM69 ( or TRIM69 CA ) , E1 ( Hdm2 ) , and E2 ( UbcH5a ) in the presence of ATP . The in vitro ubiquitination of NS3 was analyzed by western blots . The lentiviral shRNA against mouse TRIM69 and matched control lentiviral vector were transfected into 293T cells together with the relative packaging plasmids . Lentiviruses were produced from the cells after 72 h transfection , and purified by ultracentrifugation . Then the 5×107 pfu of lentiviruses derived from shm69-1 were injected into mice caudal vein . 7 day post lentiviruses injection , mice were challenged with DENV-2 ( 1×107 pfu ) via intravenous injection . 3 days post DENV infection , mice were sacrificed and the Lung , Spleen and Kidney organs were dissected to monitor the DENV replication . Prism 7 software ( GraphPad Software ) was used for charts and statistical analyses . The significance of results was analyzed by an unpaired two-tailed ANOVA test or Student’s t-test with a cutoff P value of 0 . 05 .
|
Mosquito-borne viruses , such as Dengue virus ( DENV ) , have become global threats to human health in recent years . However , no antiviral drugs have been approved to treat DENV induced diseases , and the safe and effective vaccines are still under development . It is of great importance to explore the detail mechanisms of host-virus interaction . In this report , we found that an interferon inducible host protein , TRIM69 , is upregulated upon DENV infection . TRIM69 acts as a restriction factor for DENV replication both in vitro and in vivo . As an E3 ubiquitin ligase , TRIM69 directly binds to viral Nonstructural Protein 3 ( NS3 ) , which leads to NS3 ubiquitination and degradation . Thus TRIM69 is a novel interferon inducible host antiviral factor by targeting a specific viral protein for its degradation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods",
"and",
"materials"
] |
[
"dengue",
"virus",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"293t",
"cells",
"pathogens",
"enzymes",
"biological",
"cultures",
"microbiology",
"enzymology",
"ubiquitin",
"ligases",
"viruses",
"rna",
"viruses",
"immunoprecipitation",
"ligases",
"research",
"and",
"analysis",
"methods",
"genetic",
"interference",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"ubiquitination",
"viral",
"replication",
"cell",
"lines",
"precipitation",
"techniques",
"biochemistry",
"post-translational",
"modification",
"flaviviruses",
"virology",
"viral",
"pathogens",
"interferons",
"genetics",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2018
|
Interferon-stimulated TRIM69 interrupts dengue virus replication by ubiquitinating viral nonstructural protein 3
|
Members of the family of calcium dependent protein kinases ( CDPK’s ) are abundant in certain pathogenic parasites and absent in mammalian cells making them strong drug target candidates . In the obligate intracellular parasite Toxoplasma gondii TgCDPK3 is important for calcium dependent egress from the host cell . Nonetheless , the specific substrate through which TgCDPK3 exerts its function during egress remains unknown . To close this knowledge gap we applied the proximity-based protein interaction trap BioID and identified 13 proteins that are either near neighbors or direct interactors of TgCDPK3 . Among these was Myosin A ( TgMyoA ) , the unconventional motor protein greatly responsible for driving the gliding motility of this parasite , and whose phosphorylation at serine 21 by an unknown kinase was previously shown to be important for motility and egress . Through a non-biased peptide array approach we determined that TgCDPK3 can specifically phosphorylate serines 21 and 743 of TgMyoA in vitro . Complementation of the TgmyoA null mutant , which exhibits a delay in egress , with TgMyoA in which either S21 or S743 is mutated to alanine failed to rescue the egress defect . Similarly , phosphomimetic mutations in the motor protein overcome the need for TgCDPK3 . Moreover , extracellular Tgcdpk3 mutant parasites have motility defects that are complemented by expression of S21+S743 phosphomimetic of TgMyoA . Thus , our studies establish that phosphorylation of TgMyoA by TgCDPK3 is responsible for initiation of motility and parasite egress from the host-cell and provides mechanistic insight into how this unique kinase regulates the lytic cycle of Toxoplasma gondii .
The phylum Apicomplexa encompasses numerous obligate intracellular parasites that pose a significant health risk to animals and humans . Among these , Toxoplasma gondii is one of the most widespread , infecting all warm-blooded animals including approximately one third of the human population . Humans become infected congenitally or by ingestion of either environmental oocysts , which are shed in the feces of cats , or tissue cysts in the undercooked meat of infected animals . Most infections are asymptomatic during the acute stage but as to evade the immune response the parasite converts to a latent encysted form , thus establishing a chronic infection . In immunocompromised individuals and lymphoma patients , new infections or rupture of pre-existing cysts can lead to life-threatening toxoplasmic encephalitis [1–3] . Additionally , in congenital infections , toxoplasmosis can lead to blindness , severe neurological problems , or even death , given the immature nature of the fetal immune system [4] . A significant portion of the pathogenesis observed during toxoplasmosis is a direct consequence of the repeating cycles of invasion , division and egress that drive propagation of the parasite through the infected organism [5] . As the parasites escape their host cell during egress , the host membrane is ruptured resulting in cell death and an ensuing inflammatory response , both of which contribute to the pathogenesis of this infection . Active egress from the host cell involves parasite motility , cytoskeletal rearrangements within the parasite , and secretion from specialized organelles known as the micronemes [6–9] . A pore forming protein secreted from the micronemes , the perforin-like protein TgPLP1 , facilitates egress by permeabilizing both the parasitophorous vacuolar membrane ( PVM ) and host plasma membrane [10] . Secretion of TgPLP1 and the initiation of motility during egress are regulated by calcium signaling , which is evident by the fact that treatment of intracellular parasites with calcium ionophores induces microneme secretion , motility and egress [6–9] . Calcium signaling in this parasite is quite distinct from what is typically observed in mammalian cells , involving plant-like factors such as the phytohormone abscisic acid ( ABA ) [11] and members of the family of Calcium Dependent Protein Kinases ( CDPK ) [12] . In particular , TgCDPK1 has been shown to be upstream of a signaling pathway regulating microneme secretion during egress and invasion [13] . Recently , three research teams , ours among them , identified a second calcium dependent protein kinase , TgCDPK3 , as being critical for ionophore-induced egress ( iiEgress ) [14–16] . Through a series of selection and screens we isolated independent mutants that exhibit delayed iiEgress , resistance to extracellular exposure to calcium ionophores , which usually renders parasites non-invasive , and a reduction in in vivo virulence [17] . Whole genome sequencing of one of these mutant strains ( MBE1 . 1 ) revealed a missense mutation that results in threonine for isoleucine ( T239I ) change within the catalytic domain of TgCDPK3 [18 , 19] . As expected given the position of the mutated amino acid , this mutation significantly reduces the in vitro kinase activity of recombinant TgCDPK3 [14] . The critical role of TgCDPK3 as mediator of egress was validated when introduction of a wild type copy of TgCDPK3 was found to complement the phenotypes observed in MBE1 . 1 . Localization of TgCDPK3 to the periphery of the parasite [14] would suggest that it could phosphorylate membrane-associated proteins that influence egress , such as members of the motility machinery and those that regulate calcium signaling and fluxes . To experimentally determine the substrates of TgCDPK3 the relative phosphorylation site usage in wild type and Tgcdpk3 mutant parasites was determined by quantitative mass-spectrometry using stable isotope labeling with amino acids in cell culture ( SILAC ) [20] . Comparisons of phosphorylation sites in wild type ( WT ) and mutant strains were made for intracellular parasites with and without ionophore . This analysis revealed 156 sites that are differentially phosphorylated between WT and mutant parasites . Importantly , most of the differential phosphorylation between the mutant and wild type strains is rescued in the complemented strain . A third of the phosphosites detected ( 51 of 156 ) showed a significant difference between WT and mutant parasites even in the absence of ionophore , indicating that TgCDPK3 regulates biological processes independent of iiEgress . This category includes proteins important for ion-homeostasis and metabolism , which is supported by the observation that basal calcium levels are increased in Tgcdpk3 mutant parasites [20] . Among ionophore induced phosphosites that are more abundant in the WT than in the mutant strains are many that could play a role in egress or parasite motility such as Myosin A , F , and G , proteins of the inner membrane complex ( IMC ) [21] and a recently discovered protein that associates with cortical microtubules , TrxL-1 ( TGGT1_115220 ) [22] . Interestingly a recent study showed that one of these candidates , Myosin A is phosphorylated in a calcium dependent manner at specific sites and that this phosphorylation event is important for parasite egress although the responsible kinase was not identified [23] . The list of proteins less phosphorylated in TgCDPK3 mutants also includes calcium-signaling proteins including a putative calmodulin ( TGGT1_042450 ) and two calcium-dependent kinases ( TgCDPK2a and TgCDPK3 itself ) . These results show that TgCDPK3 plays a pivotal role in regulating tachyzoite functions including , but not limited to , egress . Given the complexity of the TgCDPK3-related phosphoproteome the mechanistic reason for the egress defect observed in parasites lacking TgCDPK3 function remains unexplained . In this study we define the TgCDPK3 interactome through the implementation of a proximity based interaction protein trap and identify Myosin A ( TgMyoA ) as a TgCDPK3 substrate . We show that TgCDPK3 specifically phosphorylates TgMyoA at Serines 21 and 743 in vitro and that these phosphorylation events are important for parasite egress in vivo .
To identify putative substrates and interacting proteins of TgCDPK3 we utilized the BioID system , which relies on fusing a protein of interest to a mutant version of the bacterial BirA biotin ligase ( BirA* ) [24] . This mutant version of BirA lacks specificity and thus promiscuously biotinylates any protein within 10 nm of the fusion protein . Accordingly , we generated a construct in which BirA* is fused to the C-terminus of TgCDPK3 followed by a hemagglutinin ( HA ) epitope tag ( TgCDPK3-BirA*-HA , aka BirA* fusion ) . The BirA* fusion construct was transfected into the Tgcdpk3 mutant strain MBE1 . 1 [14] . As a control we transfected a construct carrying an HA tagged TgCDPK3 ( TgCDPK3-HA ) into MBE1 . 1 as well . Western blot using anti-HA antibodies showed that our recombinant strains correctly express either TgCDPK3-HA or TgCDPK3-BirA*-HA , both migrating at the expected size ( Fig 1A ) . Immunofluorescence assays showed that the fusion protein is targeted to parasite periphery similarly to what is observed with TgCDPK3 ( Fig 1B ) . Since we expressed the BirA* fusion protein in a strain lacking TgCDPK3 function we were able to test its functionality by its ability to complement the egress phenotype observed in Tgcdpk3 mutant strains [14] . After 2 minutes of exposure to the calcium ionophore A23187 , MBE1 . 1 parasites remained mostly intracellular ( 99 . 6% ) while those expressing TgCDPK3-HA or BirA* fusion protein showed 100% and 96 . 4% egress respectively ( Fig 1C ) . Thus , we have generated a strain expressing a BirA* fusion protein , which localizes correctly and is functional in the context of egress . To identify putative TgCDPK3 interacting proteins we grew TgCDPK3-HA and BirA* fusion-expressing parasites in the presence of biotin . Lysates of both cultures were treated with RIPA buffer and the supernatant was subjected to affinity purification with streptavidin conjugated magnetic beads to trap the biotinylated proteins . Western blot of the precipitated material showed that , in addition to proteins that were common between the TgCDPK3-HA and BirA* fusion protein expressing parasites , there were several proteins that appeared to be biotinylated solely in the BirA* fusion protein expressing parasites ( Fig 1D ) . Having confirmed the presence of various proteins exclusively biotinylated in the parasites expressing BirA* fusion protein , we scaled up the affinity purification of biotinylated proteins from parasites grown with biotin and subjected the resulting material to mass spectroscopy analysis . This analysis identified six proteins that were common between the two strains ( S2 Table ) and fourteen proteins that were detected only in the MBE1 . 1 + TgCDPK3-BirA* parasite sample including TgCDPK3 , which was expected as the BirA* fusion would biotinylate itself ( Table 1 ) . Remarkably , seven of the proteins identified through our approach were previously shown through a proteomic study to be differentially phosphorylated between wild type and Tgcdpk3 mutant parasites ( Table 1 , in bold ) . Having identified these proteins through two independent approaches strongly suggests that they might be direct substrates of TgCDPK3 . Among the proteins that interact with TgCDPK3-BirA* the top hit was Myosin A ( TgMyoA ) , which was also identified as less phosphorylated at serine 20 or 21 in the phosphroproteome of parental and Tgcdpk3 mutant parasites [20] . To further confirm that TgMyoA is less phosphorylated in Tgcdpk3 mutant parasites we performed Phos-tag gel electrophoresis , which involves use of Phos-tag biomolecule that specifically binds phosphorylated proteins and retards their migration in the gel [25] . Towards this goal , we harvested intracellular MBE1 . 1 ( Tgcdpk3 mutant ) or MBE1 . 1+TgCDPK3-HA parasites in presence of intracellular buffer and re-suspended in either intracellular or extracellular buffer followed by incubation at 37°C for 2 minutes and examined the phosphorylation status of TgMyoA ( Fig 2A ) . The results showed that TgMyoA’s migration is significantly slower in extracellular conditions , indicating that it is phosphorylated when parasites transition from intra- to extra-cellular conditions . Importantly , this shift in migration of TgMyoA is reduced in the Tgcdpk3 mutant strain , MBE1 . 1 , confirming that TgMyoA is less phosphorylated in the absence of TgCDPK3 function ( Fig 2A ) . As a next step we set out to determine whether TgCDPK3 can phosphorylate TgMyoA and TgGAP45 , another protein that forms part of T . gondii’s motility complex [7] , and was also exclusively identified in the sample from biotin exposed TgCDPK3-BirA* parasites . Towards this goal we performed an in vitro phosphorylation assay using purified recombinant TgCDPK3 and a non-biased overlapping peptide array covering the entire TgMyoA and TgGAP45 sequences . Each peptide was 15 amino acids in length and tiled peptides were shifted by 3 amino acids . The peptides ( 273 for TgMyoA and 78 for TgGAP45 ) were spotted on a modified cellulose membrane using routine Fmoc ( N- ( 9-fluorenyl ) methoxycarbonyl ) chemistry , deprotected and exposed to activated recombinant TgCDPK3 in presence of [γ-32P]ATP and calcium . Peptide spots phosphorylation was quantified using phosphoimaging . For TgMyoA , two peptides ( 13ATALKKRSSDVDHAVD28 and 736AALRLLKSSKLPSEE750 ) showed phosphorylation signal >100 PSL/mm2 ( Fig 2B ) . By contrast , none of the peptides spanning GAP45 showed significant phosphorylation signal ( S1 Fig ) . TgCDPK3 is a serine threonine kinase and in each of the two peptides of TgMyoA that were phosphorylated there are 3 potential phosphorylation sites ( Fig 2B ) . To determine the specific residues that are phosphorylated we generated mutated versions of both peptides that contained single , double or triple mutations where serine ( S ) or threonine ( T ) were mutated to the non-phosphorylable residue alanine . In vitro phosphorylation of these mutant peptides with purified recombinant TgCDPK3 showed that in the peptide 13ATALKKRSSDVDHAVD28 , mutation of either T14 or S20 does not affect phosphorylation signal while mutation of S21 results in complete loss of phosphorylation ( Fig 2C ) . In the second peptide 736AALRLLKSSKLPSEE750 mutation of S744 or S748 does not significantly affect phosphorylation while mutation of S743 leads to 96 . 4% loss of phosphorylation signal ( Fig 2D ) . These results indicate that TgCDPK3 can specifically phosphorylate S21 and S743 residues of TgMyoA . Previous studies have shown that in T . gondii , TgMyoA is phosphorylated at multiple sites including S21 [20 , 26] . However , TgMyoA S743 has not been previously reported as phosphorylated in Toxoplasma parasites . To address whether S743 is phosphorylated in vivo we immuno-precipitated the motor complex with an antibody against TgGAP45 and analyzed the phosphorylation status of TgMyoA by mass spectrometry . The analysis indicated that S743 is indeed phosphorylated in intracellular parasites as evidenced by phosphorylation status of the first serine of the peptide 743SSKLPSEEYQLGKTMVFLK760 ( S2 Fig ) . Interestingly , it has been previously reported that genetic disruption of TgMyoA results in a delay of ionophore-induced egress reminiscent of what is observed in Tgcdpk3 mutant parasites [9] . To determine the importance of phosphorylation of S21 and S743 of TgMyoA during induced egress , a process that is regulated by TgCDPK3 , we complemented a TgMyoA null mutant strain with either wild type TgMyoA or TgMyoA in which either S21 or S743 , or both were mutated to alanine ( Fig 3A ) . Immunofluorescence assays of parasites expressing the wild type or mutant MyoA indicate that the transgenic proteins correctly localize to the inner membrane complex ( Fig 3B ) . Importantly , western blot analysis showed that wild type and mutant TgMyoA are expressed at similar levels in these transgenic parasites ( Fig 3C ) . We exposed these transgenic parasite lines as well as the TgMyoA knockout strain ( MyoA KO ) to A23187 for 2 minutes to determine the efficiency of ionophore-induced egress . As expected , the TgMyoA KO exhibited a strong egress defect ( 1% egress ) , which was complemented by expression of wild type TgMyoA ( 97 . 2% egress , Fig 3D ) . By contrast , the TgMyoA mutants TgMyoA ( S21A ) , TgMyoA ( S743A ) , and TgMyoA ( S21A+S743A ) only partially rescued the egress phenotype with 69 . 8% , 59 . 5% , and 53% egress , respectively ( Fig 3D ) . These results suggest that the presence of a phosphorylatable serine at positions 21 and/or 743 of TgMyoA contributes to Toxoplasma egress from host cells . We next tested whether mutating S21 and/or S743 of TgMyoA to the phosphomimetic residue aspartic acid could rescue the egress defect of Tgcdpk3 mutant parasites . Because phosphomimetic residues ( aspartic acid or glutamic acid ) do not fully approximate the electronegativity produced by phosphorylation , we employed the strategy of mutating two neighboring pairs of amino acids to overcome the charge differential [27 , 28] . Accordingly , we transfected the Tgcdpk3 mutant strain MBE1 . 1 with either a FLAG tagged wild type copy of TgMyoA or FLAG tagged TgMyoA in which serine residues 20 and 21 , 743 and 744 , or 20 , 21 , 743 and 744 were mutated to aspartic acid ( Fig 4A ) . Immunofluorescence assays and Western blots assays indicated that all versions of TgMyoA were correctly targeted and expressed at equal levels ( Fig 4B and 4C ) . At 2 minutes of exposure to A23187 , which is sufficient to induce egress of 100% of wild type parasites ( Fig 1C ) , MBE1 . 1 mutant parasites expressing wild type or phosphomimetic MyoA showed only 0 . 4% and 3 . 6% egress respectively . Nonetheless , by 6 minutes of ionophore treatment we saw a significant difference between the MBE1 . 1 mutant parasites expressing an exogenous copy of wild type MyoA and those expressing the phosphomimetic versions of the protein ( Fig 4D ) . Induction of egress with A23187 for 6 minutes showed that nearly all ( 97 . 3% ) MBE1 . 1 parasites expressing the exogenous wild type copy of MyoA remained inside of the cells after six minutes of treatment . This indicates that overexpression of TgMyoA does not rescue the egress defect associated with lack of TgCDPK3 activity . By contrast , expression of either TgMyoA SS ( 20–21 ) DD , TgMyoA SS ( 743–744 ) DD and TgMyoA S ( 20-21-743-744 ) D in MBE1 . 1 significantly complemented the ionophore induce egress phenotype ( 87% , 91 . 3% , and 86 . 7% egress at 6 minutes post induction respectively , Fig 4D ) . Thus , mimicking constitutively phosphorylated TgMyoA partially overrides the need for TgCDPK3 function during calcium-stimulated egress . TgMyoA is an important component of glideosome and plays a critical role in parasite motility [29 , 30] . Thus , it is plausible one of the roles of TgCDPK3 during induced egress is to initiate motility via the phosphorylation of TgMyoA . Interestingly , previous studies have shown that chemical inhibition of TgCDPK3 affects initiation of motility in extracellular parasites [15] . Additionally , we have previously shown that Tgcdpk3 mutant strains have reduced efficiency of invasion [14 , 17] , a process that depends on motility . To further examine the role of TgCDPK3 in parasite motility , we tested the efficiency of the Tgcdpk3 mutant strain ( MBE1 . 1 ) and the complemented strain ( MBE1 . 1+TgCDPK3 ) in transitioning from a non-motile to a motile state . This was accomplished by recording and analyzing live video microscopy of parasites for two minutes after changing the media from one that mimics intracellular conditions ( IC buffer ) to one that mimics extracellular conditions ( EC buffer ) [31] . While 85 . 8% of the complemented parasites had become motile by two minutes after switching the media , only 22 . 8% of the TgCDPK3 mutant parasites were motile during the same time period ( Fig 5A ) . Toxoplasma parasites normally exhibit three types of motility patterns referred to as helical , twirling and circling [32] . Therefore , we scored the type of motility exhibited by those parasites of either strain that were moving to determine whether TgCDPK3 played a role in a specific type of movement . The results showed that the proportion of parasites exhibiting each type of movement was similar between the two strains ( Fig 5B ) . As we noted that the mutant parasites appear to move at a slower pace than wild type ones , we also examined speed of parasite movement when switched from intracellular to extracellular conditions . TgCDPK3 mutants moved with an average speed of 0 . 21 μ/s while complemented parasites exhibited a much higher speed of 1 . 15 μ/s ( Fig 5C ) . Therefore , the large difference between the mutant and complemented strains in the percentage of parasites that quickly initiated motility and also faster speeds confirm that TgCDPK3 plays a role in parasite motility . We next wanted to determine if phosphomimetic mutants of TgMyoA could rescue the motility defect in TgCDPK3 mutant parasites . For this we analyzed MBE1 . 1 parasites expressing an extra copy of either TgMyoA ( WT ) or TgMyoA S ( 20-21-743-744 ) D by video microcopy . The analysis showed that only 17 . 3% of MBE1 . 1 + TgMyoA ( WT ) parasites became motile once transitioned from IC to EC buffer ( Fig 5D ) . By contrast , MBE1 . 1 parasites complemented with TgMyoA S ( 20-21-743-744 ) D showed increased levels of motility initiation with 56 . 6% of them becoming motile during the first two minutes after switching the buffer ( Fig 5D ) . Quantification of three kinds of gliding motility between the two strains showed that the proportion of each movement was similar between the two strains ( Fig 5E ) . However , when examined for speed , MBE1 . 1 + TgMyoA ( WT ) showed a pace of 0 . 1 μ/s while MBE1 . 1 + TgMyoA S ( 20-21-743-744 ) D moved at a slightly higher speed of about 0 . 29 μ/s ( Fig 5F ) . These results indicate that phosphomimetics of TgMyoA can rescue the motility defect of Tgcdpk3 mutants at least in number of motile parasites suggesting they can compensate lack of TgCDPK3 function in initiating parasite motility . However , as it is the case for iiEgress , this rescue is not complete , as the speed of the MBE1 . 1 + TgMyoA S ( 20-21-743-744 ) D parasites is still approximately threefold less than the wild type parasites .
Although it is well established that TgCDPK3 is important for parasite egress , the particular mechanism by which this calcium-stimulated kinase regulates this key event of the lytic cycle is not known . A recent study revealed 156 phosphorylation sites out of more than 12 , 000 quantified that were differentially phosphorylated between wild type and Tgcdpk3 mutant parasites , with many of them related to motility , ion-homeostasis and metabolism [20] . While some of these differentially phosphorylated sites might be direct substrates of TgCDPK3 , one might expect that a number of these sites related to downstream effects and compensatory mechanisms related to a loss in TgCDPK3 signaling . In addition , TgCDPK3 is involved in several processes such as calcium homeostasis and parasite division , which might involve distinct substrates from those involved in egress . Therefore , additional efforts were needed to identify the specific protein ( s ) whose phosphorylation by TgCDPK3 is key for induced egress and initiation of motility . With this in mind we successfully adapted the BioID system to identify putative substrates and interactors of TgCDPK3 . This approach , which is based on fusing the protein of interest to a promiscuous allele of the biotin ligase BirA , has the advantage that it can identify not only direct interactors but also proteins that are nearby or interact loosely or transiently , such as enzyme substrates . Identification by proximity labeling does not prove an enzyme/substrate relationship , and it is plausible that some of these interactions are structural in nature . However , in combination with our previous phosphoproteome analysis [20] , which identified seven of the thirteen proteins identified through BioID as less phosphorylated in Tgcdpk3 mutant parasites , it provides important indirect evidence for a kinase/ substrate relationship . Having been linked to TgCDPK3 in two independent and distinct approaches makes these seven proteins strong candidates for being TgCDPK3 substrates during the events regulated by this kinase . These seven putative substrates include TgCDPK2a , GAPDH1 , a MORN repeat containing protein , IMC4 , TgMyoA and TgMyoG , and a hypothetical protein of unknown function . Interestingly , many of these proteins are known or would be predicted to be within the periphery of the parasite , which strengthens the argument that they might be TgCDPK3 substrates . For example , IMC4 is part of the inner membrane complex [33] , which is a continuous layer of flat vesicles sutured together and to which the motor protein TgMyoA is anchored [30] . The glyceraldehyde 3-phosphate dehydrogenase 1 ( TgGAPDH1 ) , which is normally cytoplasmic , redistributes to the periphery of the parasite during egress [31] . Additionally , both the MORN repeat-containing protein and the hypothetical proteins ( TgGT1_310420 ) are predicted to be myristoylated , which suggests membrane localization . The function of the hypothetical protein is not known , but MORN proteins are involved in cell division in eukaryotes including T . gondii [34 , 35] . Interestingly , among the putative substrates identified in both the proteome and the BioID approaches , is a second calcium dependent protein kinase ( TgCDPK2a ) , which suggests that these kinases might work together as co-regulators of a protein network or as part of a signaling cascade . Nonetheless , at present no information as to either the localization or the function of TgCDPK2a is available . Of special interest among the proteins identified through BioID are TgMyoA and GAP45 , both of which form part of the motor complex driving the parasite’s gliding motility . The so-called glideosome resides in the space between the parasite plasma membrane and the IMC and it is a complex of several proteins including TgMyoA , two associated light chains , myosin light chain TgMLC1 and essential light chain TgELC1 and the glideosome associated proteins TgGAP40 , TgGAP50 , TgGAP45 or TgGAP70 [7 , 30 , 36 , 37] . Given the proximity of TgCDPK3 to the glideosome and the facts that induced egress is dependent on motility and that motility is a calcium-dependent process , a functional connection between TgCDPK3 and the motility machinery is a plausible one . Recent studies using a small molecule invasion enhancer that causes an increase in intracellular Ca2+ showed that TgMyoA is phosphorylated in a calcium dependent manner at specific residues , serine 20 , 21 and 29 and that phosphorylation of serine 21 is important for ionophore induced egress and motility [26] . However the kinase that mediates this phosphorylation process had not been known . Interestingly , TgCDPK3 had been considered as a likely candidate given the remarkably similar egress phenotype seen in both the TgmyoA and Tgcdpk3 mutant strains . Consistent with this idea peptides containing phosphorylated serine 20 or 21 were found to be less abundant in Tgcdpk3 mutant strains as compared to the parental or complemented ones in proteomic studies [20] . This finding suggested that the phosphorylation status of TgMyoA , at least for Ser20/21 may be coupled to TgCDPK3 signaling . Our BioID and mutagenesis data suggests that TgMyoA is directly regulated by TgCDPK3 and is thus a bona fide substrate for TgCDPK3 . Our data strongly argues for a direct relation between TgMyoA phosphorylation and TgCDPK3 . Based on our BioID results , TgMyoA either interacts with or is in close proximity to TgCDPK3 . Consistent with this idea; we show that recombinant TgCDPK3 can indeed phosphorylate TgMyoA in vitro with preference for serines 21 and 743 . While CDPKs can act non-specifically in vitro , it is important to note that we did not detect significant phosphorylation of any of the other known phosphorylated amino acids of TgMyoA and of none of those from TgGAP45 , indicating that we are observing some level of specificity in our peptide array assay . Interestingly , TgGAP45 was not observed as less phosphorylated in our mutant strains [20] . Therefore , any interaction between TgCDPK3 and TgGAP45 is likely to be structural and not enzymatic and the phosphorylation state of TgGAP45 , which is important for its function [38] , is likely regulated by a different kinase . Creating a version of TgMyoA that ‘looks’ phosphorylated overrides the need for TgCDPK3 , strongly indicating that this is the kinase responsible for modifying TgMyoA . Nonetheless , the complementation of iiEgress by the phosphomimetic versions of TgMyoA is partial . While the levels of egress and motility exhibited by the phosphomimetic expressing strains are significantly higher than that of the CDPK3 mutant strain , they don’t reach wild type levels . There are several plausible reasons for this incomplete complementation of the iiEgress phenotype , including the fact that phosphomimetic mutations are not a perfect simulation of phosphorylated serine [27 , 28] . Also , phosphomimetic mutations that result in constitutively active serine or threonine do not allow for dynamic changes of alternating phosphorylation and dephosphorylation events that might be occurring in vivo and are important for function . Another possibility we must consider is that other amino acids within TgMyoA are also regulated by TgCDPK3 , but were not revealed in our in vitro assays . Finally , and most likely , TgMyoA might not be the only substrate through which TgCDPK3 is exerting its regulation of iiEgress and/or other kinases might work redundantly along with TgCDPK3 . We have previously shown that disruption of TgCDPK3 results in dysregulation of calcium homeostasis , a phenotype not observed previously with any TgMyoA mutants , which could affect sensitivity to the ionophore . We haven’t specifically tested whether TgMyoA has an effect on resting calcium levels in this study but its involvement is unlikely given its predicted role as part of the molecular motor that drives movement of the parasite . Several of the putative substrates we identified including TgCDPK2A , Myosin-G and GAPDH1 are good candidates for influencing egress and future work will focus on understanding their potential contribution to TgCDPK3 regulated events . An interesting question that remains unanswered is the particular timing of the phosphorylation of TgMyoA by TgCDPK3 . Does TgCDPK3 phosphorylate TgMyoA during intracellular growth or does it occur upon induction of egress ? Based on the results obtained using Phos-tag ( Fig 2A ) it appears that there is a significant level of TgCDPK3-dependent phosphorylation of TgMyoA upon the transition from intracellular to extracellular conditions , which mimics what the parasite encounters during egress . Nonetheless , based on phosphoproteomic comparison between wild type and Tgcdpk3 mutant parasites , phosphorylated Ser21 is more abundant in the wild type strain even in intracellular parasites not exposed to ionophore , which would suggest this phosphorylation event occurs in the absence of egress induction . Interestingly , it has been reported that phosphorylation of several amino acids in TgMyoA is dependent on Ser21 being phosphorylated first [26] . Thus , a plausible model that would explain these various results is that TgCDPK3 phosphorylates TgMyoA at Ser21 in response to calcium fluxes that occur during intracellular growth , and that upon induction of egress either TgCDPK3 or another kinase further phosphorylates TgMyoA in a phospho-Ser21 dependent manner . Thus , in the absence of TgCDPK3 phosphorylation of TgMyoA is significantly altered during egress due to lack or reduction of Ser21 phosphorylation . Another important standing question is how phosphorylation of TgMyoA at those two particular sites influences its function at the mechanistic level . The importance of phosphorylation is well established for class II myosins , which are found in skeletal muscle . Nonetheless , TgMyoA is quite divergent structurally from other myosins [39 , 40] and therefore its regulation is likely to be unique . TgMyoA is a single headed motor protein [41] that belongs to the class XIVa myosin family which is unique to Apicomplexans and ciliates and the conserved motor domain shares only about 23–34% homology with mammalian myosins [23 , 39] . Class XIVa myosins also lack the conserved glycine at the lever arm pivot point and have a shorter C-terminal tail , which has been shown to be important for motor function in class II myosins [41] . In TgMyoA , Ser 21 is located in the N-terminal region whose role remains undefined , while Ser 743 lies within the motor domain . It is feasible that phosphorylation of these residues either results in structural modification of TgMyoA that in turn allows new protein-protein interactions or activates its enzymatic activity both of which could be important for mechanochemical function of TgMyoA . The recent successful expression and purification of recombinant TgMyoA [42] will be particularly useful to investigate how phosphorylation influences the function and biochemistry of this unique and key motor protein . Those new in vitro methods along with our novel discovery that TgCDPK3 phosphorylates TgMyoA within the parasite to initiate egress , will provide a more complete understanding of how motility is tightly regulated during the lytic cycle of this important human parasite .
Toxoplasma gondii tachyzoites were maintained by passage through human foreskin fibroblasts ( HFF , obtained from the American Tissue Culture Collection ATCC ) in a humidified incubator at 37°C with 5% CO2 . Normal growth medium consisted of DMEM supplemented with 10% fetal bovine serum , 2 mM L-glutamine and 50μg/ml of penicillin-streptomycin . Purification of parasites was performed as previously described [43] . Primers used in generating plasmid constructs described in this section are listed in supplemental table S1 ( S1 Table ) . To generate the BirA* fusion ( TgCDPK3-BirA-HA ) , CDPK3-BirA*-HA was commercially synthesized ( GenScript , USA ) , amplified by PCR using specific primers ( S1 Table ) and directionally cloned downstream of the Tgcdpk3 promoter in the vector , pTgcdpk3CDPK3-HA [43] using NcoI and PacI sites . The non-phosphorylable and phosphomimetic mutants of TgMyoA were made using Lightning site directed mutagenesis kit ( Agilent Technologies ) with primers listed in S1 Table and pmyoA-FLAGTgMyoA-WT/graBle [23] as the parent plasmid . All resulting constructs were verified by restriction digestion and sequencing . Plasmid constructs were linearized with the restriction enzyme KpnI , purified and electroporated into T . gondii tachyzoites according to established protocols [44 , 45] . Parasites transfected with BirA* fusion construct were cultured in presence of 50 μg/ml mycophenolic acid ( MPA ) and 50 μg/ml xanthine and cloned by limiting dilution to obtain stably transformed clones . When using vectors carrying the Ble gene as a selectable marker , transfected parasites were added onto an HFF monolayer and allowed to grow without any drug selection until the monolayer was lysed . Freshly egressed parasites were then washed with Hanks’s balanced salt solution containing 10 mM HEPES and 0 . 1 mM EGTA ( HHE ) and extracellular parasites were treated with 50 μg/ml phleomycin in DMEM for 4 hours at 37°C with 5% CO2 . The parasites were then added onto a HFF monolayer and cultured in the presence of 5 μg/ml phleomycin to select drug resistant parasites , which were cloned by limiting dilution . Affinity purification of biotinylated proteins was performed according to previously described protocols with minor modifications [24 , 46] . Briefly , parasites were cultured in growth medium containing biotin ( 150 μg/ml ) for 48 hours . Freshly egressed parasites ( 2 . 5 x 109 ) were then washed with phosphate buffered saline ( PBS ) and lysed with 1 ml RIPA buffer ( 20 mM Tris-HCl ( pH 7 . 5 ) , 150 mM NaCl , 1 mM Na2EDTA , 1 mM EGTA , 1% NP-40 , 1% sodium deoxycholate , 2 . 5 mM sodium pyrophosphate , 1 mM β-glycerophosphate , 1 mM Na3VO4 ) supplemented with complete protease inhibitor ( Roche ) and centrifuged at 16000 g for 15 minutes at 4°C . The supernatant was then transferred to a fresh tube and incubated with magnetic streptavidin beads ( Dynabeads Myone streptavidin C1 from Invitrogen ) at 4°C for 12 hours with gentle shaking . Beads were then collected with magnets and washed twice with wash buffer 1 ( 2% SDS ) , once with wash buffer 2 ( 0 . 1% deoxycholate , 1% Triton X-100 , 500 mM NaCl , 1 mM EDTA and 50 mM HEPES , pH 7 . 5 ) , once with wash buffer 3 ( 250 mM LiCl , 0 . 5% NP-40 , 0 . 5% deoxycholate , 1 mM EDTA and 10 mM Tris pH 8 . 1 ) , twice with wash buffer 4 ( 50 mM Tris , pH7 . 4 and 50 mM NaCl ) and twice with PBS , in that order . The beads were finally re-suspended in 1 ml PBS and 10% of each sample was then boiled at 98°C for 5 minutes to separate bound proteins from magnetic beads and eluted proteins were analyzed by either silver staining or Western blotting using streptavidin-HRP before mass spectrometry . Mass spectrometric analysis was carried out on a Thermo-Fisher Scientific LTQ Orbitrap Velos Pro mass spectrometer ( Thermo-Fisher Scientific , Waltham , MA ) interfaced with a Waters Acquity UPLC system ( Waters , Milford , MA ) . The proteins bound to streptavidin beads ( biotinylated proteins ) and IgG beads ( TgMyoA ) were directly digested by trypsin . Samples were first reduced with 10 mM DTT in 10 mM ammonium bicarbonate and then alkylated with 55 mM iodoacetamide ( prepared freshly in 10 mM ammonium bicarbonate ) . Alkylated samples were digested by trypsin ( Promega , Madison , WI ) overnight at 37°C . Tryptic peptides were first injected onto a C18 trapping column ( NanoAcquity UPLC Trap column 180μm x 20mm , 5μm , Symmetry C18 ) and subsequently onto an analytical column ( NanoAcquity UPLC column 100μm x 100mm , 1 . 7μm BEH130 C18 ) . Peptides were eluted with a linear gradient from 3 to 40% acetonitrile in water with 0 . 1% formic acid developed over 90 minutes at room temperature at a flow rate of 500 nL/min , and the effluent was electro-sprayed into the LTQ Orbitrap mass spectrometer . Blanks were run prior to the sample to make sure there were no significant background signals from solvents or the columns . Database search against Toxoplasma gondii GT1 strain annotated proteins from ToxoDB ( release 10 . 0 , updated January 31 , 2014 ) was performed using Sequest ( Thermo-Fisher Scientific ) search engine to identify biotinylated proteins and TgMyoA post-translational modification analysis was performed using the Thermo-Fisher Scientific Proteome Discoverer software ( v2 . 0 ) . The N-HIS-tagged TgCDPK3 expression construct described previously [14] was transformed into BL21-Rosetta ( DE3 ) pLysS cells , which were then induced to express recombinant protein at 37°C with IPTG . His tagged recombinant protein was purified under native conditions using QIAexpress Ni-NTA fast start kit ( Qiagen ) according to manufacturer’s protocol . The kinase activity of recombinant TgCDPK3 was examined using peptide substrate syntide-2 ( PLARTLSVAGLPGKK , AnaSpec , Inc . ) and exhibited a specific activity of 22 . 9 μmol/min/mg . Peptide arrays were synthesized using SPOTs synthesis method and spotted onto a derivatized cellulose membrane ( Intavis ) as described previously [47] . The peptide membrane was blocked at room temperature for 30 minutes in binding buffer containing 5% BSA . Recombinant TgCDPK3 ( 5nM ) was added to 50mM HEPES , pH 7 . 4 , 100mM NaCl , 10mM MgCl2 , 100μM ATP , 1mM CaCl2 , 6μCi/ml [γ-32P]ATP and incubated at room temperature for 15 minutes . The membrane was washed three times with 100mM sodium phosphate pH 7 . 0 , 1M NaCl , 10mM EDTA and visualized using phosphorimaging ( Fuji phosphor imager ) . The phosphorylation of each peptide was detected and quantified using Multi Gauge version 3 . 0 ( Fujifilm ) . Parasite lysates were heated at 100°C for 5 minutes in SDS-PAGE sample buffer with 2% 2-mercaptoethanol and resolved on 4–20% gradient gel ( Bio-Rad , Hercules , CA ) . Proteins were transferred from the gel onto nylon membranes using a semidry transfer apparatus ( Bio-Rad , Hercules , CA ) at 12 V for 30 minutes . After blocking with 5% ( w/v ) skim milk powder in TBS , membranes were treated with rabbit anti-HA tag antibody ( Cell Signaling Technology ) , for 1 hour . Membranes were then washed and incubated with horseradish peroxidase ( HRP ) conjugated goat-anti rabbit IgG ( Sigma ) . After washing , membranes were treated with SuperSignal West Pico chemiluminescent substrate ( Pierce Chemical ) and imaged using FluorChem E ( Proteinsimple ) [43] . Intracellular parasites 24 hours post-infection were harvested in intracellular buffer [31] , filtered with 3-μm Nucleopore membrane , pelleted and re-suspended in either intracellular buffer or extracellular buffer [31] . Parasites were then incubated at 37°C for 2 minutes and immediately placed on ice followed by centrifugation at 1000 g for 10 min at 4°C . The parasite pellet was then lysed with RIPA buffer containing phosphatase inhibitor , PhosSTOP ( Roche ) followed by addition of SDS sample buffer containing β-meracaptoethanol and heated at 100°C for 5 min . To examine phosphorylation status of Myosin-A , Phos-tag gel electrophoresis was carried out according to manufacturers instructions ( Wako Chemicals , USA ) . Briefly 200 μM Phos-tag ( Wako Chemicals , USA ) and 100 μM MnCl2 were added to conventional 7 . 5% ( w/v ) acrylamide resolving gel and the gel was run at constant voltage at RT . The gel was washed three times in SDS-PAGE running buffer containing 10 mM EDTA and once each in running buffer and transfer buffer before transferring to a PVDF membrane for immunoblotting using anti-MyosinA antibody . Immunofluorescence staining of intracellular parasites was performed according to previously described procedures [48] . The primary antibodies used were: mouse anti-HA ( Cell Signaling Technology ) , and rabbit anti-GAP45 and rabbit anti-MLC1[23] . Secondary antibodies used include: Alexa Fluor-594- or Alexa Fluor-488-conjugated goat anti-rabbit or goat anti-mouse ( Molecular Probes ) . Slides were viewed using a Nikon Eclipse E100080i microscope and digital images were captured with Hamamatsu C4742-95 charge-coupled device camera using NIS elements software . The efficiency of egress after calcium ionophore treatment was determined using established protocols [14] . Percent egress was determined by dividing the number of lysed vacuoles by the total number of vacuoles for a sample . Parasite motility assay was performed according to previously described methods [15 , 23] with some modifications . Briefly , 24-well plates were pre-coated with 75 μg/ml of BSA in water at 37°C for 30 minutes and washed three times with intracellular buffer [31] . Intracellular parasites 24 hours post-infection were harvested in presence of intracellular buffer , filtered with 3-μm Nucleopore membrane , pelleted and re-suspended in intracellular buffer . The parasites were then added onto wells and allowed to settle for 20 minutes at 37°C and the plate was transferred onto a heated chamber ( set at 37°C ) of inverted microscope ( Leica AF6000 ) . The intracellular buffer in the well containing extracellular parasites was gently aspirated and extracellular buffer [31] was added . Forty seconds after exchanging the buffer , parasite motility was imaged for 2 minutes at 2 frames per second using LAS X software . The movies were then manually analyzed to determine parasites exhibiting either twirling or helical or circular gliding and the number of parasites performing each type of motility was normalized to the total number of parasites in each movie . The speed of the parasite gliding was determined by measuring the distance travelled in a given time by three motile parasites in each of two separate movies per strain . The experiments were repeated 3 times .
|
Toxoplasma gondii can cause severe disease and death in the immunocompromised and in those infected congenitally . Due to limitations of existing drugs there is a need for studying proteins that are unique and essential to the parasite . We recently established that TgCDPK3 , a member of a family of calcium dependent protein kinase present in plants and some parasites but absent in human cells , regulates parasite egress from the host cell . While it has been hypothesized that TgCDPK3 controls rapid exit from the host by phosphorylating proteins needed for activating motility , the particular substrates of this kinase remained unknown . We have now applied an interaction trap system to identify the proteins that are modified by this kinase , which include a parasite motor protein Myosin A ( TgMyoA ) . We show that TgCDPK3 specifically phosphorylates TgMyoA and this phosphorylation is important for parasite egress and motility .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Phosphorylation of a Myosin Motor by TgCDPK3 Facilitates Rapid Initiation of Motility during Toxoplasma gondii egress
|
Schistosomiasis is one of the most prevalent parasitic diseases worldwide and is a public health problem . Schistosoma mansoni is the most widespread species responsible for schistosomiasis in the Americas , Middle East and Africa . Adult female worms ( mated to males ) release eggs in the hepatic portal vasculature and are the principal cause of morbidity . Comparative separate transcriptomes of female and male adult worms were previously assessed with using microarrays and Serial Analysis of Gene Expression ( SAGE ) , thus limiting the possibility of finding novel genes . Moreover , the egg transcriptome was analyzed only once with limited bacterially cloned cDNA libraries . To compare the gene expression of S . mansoni eggs , females , and males , we performed RNA-Seq on these three parasite forms using 454/Roche technology and reconstructed the transcriptome using Trinity de novo assembly . The resulting contigs were mapped to the genome and were cross-referenced with predicted Smp genes and H3K4me3 ChIP-Seq public data . For the first time , we obtained separate , unbiased gene expression profiles for S . mansoni eggs and female and male adult worms , identifying enriched biological processes and specific enriched functions for each of the three parasite forms . Transcripts with no match to predicted genes were analyzed for their protein-coding potential and the presence of an encoded conserved protein domain . A set of 232 novel protein-coding genes with putative functions related to reproduction , metabolism , and cell biogenesis was detected , which contributes to the understanding of parasite biology . Large-scale RNA-Seq analysis using de novo assembly associated with genome-wide information for histone marks in the vicinity of gene models constitutes a new approach to transcriptome analysis that has not yet been explored in schistosomes . Importantly , all data have been consolidated into a UCSC Genome Browser search- and download-tool ( http://schistosoma . usp . br/ ) . This database provides new ways to explore the schistosome genome and transcriptome and will facilitate molecular research on this important parasite .
Schistosomiasis is a parasitic disease caused by blood-dwelling worms of the genus Schistosoma . It is an important public health problem , with high morbidity and mortality in endemic countries . Over 230 million people worldwide are infected by Schistosoma spp . , comprising three main species [1]; Schistosoma mansoni is the species responsible for infecting people in the Americas , Middle East and Africa [1] . The parasite has a complex life cycle that includes several morphological phenotypes in the intermediate Biomphalaria spp . snail host and in the human definitive host , with adult worms having separate sexes , and the mated female worms releasing hundreds of eggs daily in the mesenteric circulation of the human host [2] . In the past decade , genomic tools have helped to reveal relevant molecular players in parasite biology . Thus , the S . mansoni genome was sequenced [3] , followed by a systematically improved high-quality version of the genome [4]; however , the latter still includes over 800 genome sequence fragments and a number of incomplete gene annotations . The first S . mansoni transcriptome analyses from six different life cycle stages ( cercaria , schistosomulum , adult worm , egg , miracidium and germ ball ) were performed using first-generation EST bacterial cloning and sequencing technology [5] , with limited sequencing depth . Later , with the new second-generation , cloning-independent RNA-Seq techniques , mixed-sex adult worms [4] or male adult worms [6] were studied; however , no separate male and female gene expression was assessed by RNA-Seq . The comparison of S . mansoni male and female adult worm transcriptomes was performed only using oligonucleotide microarrays , which use short ( 60 nt ) probes to detect known genes [7] , and using Serial Analysis of Gene Expression ( SAGE ) , which sequences short ( 10–14 bp ) SAGE tags [8] that need to be matched to previously known full-length gene sequences to unequivocally assign their identity . In addition , the egg transcriptome was analyzed only once using Sanger sequencing [5] , and no additional studies have been performed on this life-cycle stage using unbiased , large-scale sequencing . In this study we compared egg , female and male transcriptomes of S . mansoni using large-scale RNA-Seq , which enabled the identification of genes functionally related to these specific developmental forms of the parasite . In addition , we documented the extension of the 5’-ends of hundreds of transcripts over the previous S . mansoni sequence predictions [4] . We then cross-referenced the new genomic coordinates of these transcript start sites ( TSSs ) with the genomic coordinates of the publicly available dataset of promoter-associated H3K4me3 histone marks obtained by ChIP-Seq [9] , and the coincident coordinates for the TSS and the H3K4me3 provided genome-wide evidence of a possible involvement of this histone modification in the transcriptional regulation of these genes . Moreover , an in silico analysis of the de novo-assembled transcriptome revealed new protein coding genes with conserved domains not previously predicted in the parasite genome .
Approximately 200 S . mansoni ( BH strain ) adult worm pairs were freshly obtained through the periportal perfusion [10] of hamsters infected 6–8 weeks earlier with 200–300 mixed-sex cercariae . After perfusion , the adult males were separated from the adult females by keeping the parasites in 150 mm culture dishes at room temperature for a period of 15 to 30 min in Advanced RPMI Medium 1640 ( Gibco , #12633–012 , Thermo Scientific , USA ) supplemented with 10% heat-inactivated calf serum ( freshly added ) , 12 mm HEPES ( 4- ( 2-hydroxyethyl ) piperazine-1-ethanesulfonic acid ) pH 7 . 4 , and 1% antibiotic/antimycotic solution ( Gibco , #15240–096 ) . Eggs of S . mansoni were isolated from the livers of five hamsters , each infected 6–8 weeks earlier with 200–300 cercariae , using the method described by the group of Brindley and collaborators [11] . Immediately before the hamster liver extraction , the S . mansoni adult worms were collected by periportal perfusion [10] . PolyA+ RNA was extracted from eggs , from 200 adult male worms and from 200 adult female worms using two rounds of the FastTrack MAG Maxi mRNA Isolation Kit ( Invitrogen , Thermo Scientific , USA ) , as described [6] , with the following modifications: 1 , 400 units of RNase Out ( Invitrogen ) and 20 mM of Vanadyl Ribonucleoside Complexes ( VRC ) were added to Lysis Buffer L4 at the first step of the sample preparation; at the final binding step , an additional 1 , 400 units of RNase Out ( Invitrogen ) was added to the sample while the tube remained in the rotator; treatment with 30 units of DNase I Amplification Grade ( Invitrogen ) was performed for 45 min at room temperature; and six washings were performed at the final washing step before elution . These modifications resulted in PolyA+ RNA samples with small percentages of rRNA contamination ( 3 to 7% ) , as estimated with a Bioanalyzer using the RNA 6000 Pico Kit ( Agilent Technologies , Santa Clara , CA , USA ) . Because the Roche 454 sequencing platform did not provide a protocol for the construction of strand-oriented RNA-Seq libraries , we previously developed a method for generating strand-oriented cDNA libraries for 454 sequencing [6] , and here , we have improved the protocol , aiming to correct the tendency toward preferential sampling of the 3′-end of the transcripts verified in the previously described method [6] . In the new strand-oriented 5′-end-first cDNA libraries method , the design of the primers was modified so that sequencing initiates from the 5′-end of the transcripts . The resulting single-stranded cDNA ( sscDNA ) libraries were PCR amplified ( 10 cycles of PCR for male and female libraries , and 17 cycles of PCR for egg libraries ) , and the resulting directional double-stranded cDNA ( dscDNA ) libraries were purified using AMPure beads ( Agencourt , Beckman Coulter , Indianapolis , IN , USA ) and quantified using PicoGreen ( Invitrogen ) . A detailed description of the library construction is given in the Supplementary Methods in S1 Text , along with a summary scheme of the procedure ( see Fig A in S1 Text ) . Directional dscDNA libraries generated as described above were sequenced from the 5′-end using the emPCR Titanium Kit and the Titanium Sequencing Kit on a Roche 454 Genome Sequencer FLX instrument , following the manufacturer’s instructions [12] . Data processing used standard 454 software procedures to generate nucleotide sequences and quality scores for all reads . High-quality ( minimum Phred score 20 ) trimmed reads were generated , and the sequence complexity filtering criteria were applied using PRINSEQ [13] . Sequencing data were deposited at the NCBI Sequence Read Archive ( SRA ) under the accession number SRP063353 . To exclude reads of ribosomal , mitochondrial or transposable element origin , we performed an alignment of reads against rRNA , mitochondrial sequences and 29 S . mansoni transposon sequences for which there were published curated full-length sequences [14–16] utilizing BLASTn [17] , and reads displaying an alignment with an e-value ≤ 10−15 were deleted . To evaluate the coverage of the reads , the filtered reads were mapped to the genome with Tophat2 [18] , and the distribution along the 5′ to 3′ gene body was assessed using the RSeQC package [19] . The S . mansoni transcriptome was reconstructed without a reference genome , using all reads from egg , female and male RNA-Seq with a Trinity assembler ( Version from April 2014 ) [20] and a k-mer size of 25 . This project has been deposited at the Transcriptome Shotgun Assembly ( TSA ) division of DDBJ/EMBL/GenBank under the accession numbers GDQY00000000 ( putative novel genes ) and GDUI00000000 ( all other contigs ) . The version described in this paper is the first version , GDQY01000000 and GDUI01000000 . For annotation , the contigs resulting from the Trinity assembly were aligned to the predicted Smp gene sequences from the S . mansoni genome version 5 . 2 , utilizing BLASTn [17] and applying as threshold an e-value = 10−5 , coverage ≥ 20% and strand specificity . In the case of the alignment of a contig to multiple Smps , the hit with higher identity was considered as the correct Smp alignment . For each Smp with multiple contigs aligned to it , we considered the contig with the highest coverage of the Smp as the representative of evidence of expression of that Smp . All Trinity-assembled contigs were aligned to the genome sequence using BLAT [21] . Contigs with genomic coordinates that did not intersect the genomic coordinates of any Smp-predicted gene were analyzed with InterProScan [22] using the Pfam database [23] , Batch CD-search tool [24] and Coding Potential Calculator ( CPC ) [25] to assess the protein-coding potential of these transcripts . Contigs with a Pfam conserved domain were clustered into categories using domain-centric Gene Ontology ( dcGO ) [26] and the online tool CateGOrizer ( http://www . animalgenome . org/tools/catego/ ) with the GO_slim2 option . Multiple alignments of Lifeguard proteins from S . mansoni , from five other schistosome species ( obtained from ftp://ftp . sanger . ac . uk/pub/pathogens/HGI/ ) and from various invertebrate and vertebrate species were restricted to the BAX1-I domain and were performed utilizing the ClustalX2 program [27] . Complete sequences for the orthologs with high identity to SmDLFG1 and 2 proteins were obtained from the five other schistosome species by analyzing their preliminary genome sequence with the program Spaln [28] . Except for the latter Schistosoma sequences , the accession numbers of the sequences used in the analysis are indicated in Fig G in S1 Text . Phylogenetic analyses were performed using Bayesian inference methodology using MrBayes program v3 . 2 . 2 x64 [29] . The analysis was performed using default parameters , except for the use of the command “prset aamodelpr = mixed , ” which enables sampling across all fixed amino acid rate matrices ( models for amino acid evolution ) implemented in the program . Analyses were stopped after 1 , 000 , 000 generations , with samplings every 100th generation . Tree information was summarized utilizing “sumt burnin = 2500” , which discards the first 250 , 000 generations . A measured potential scale reduction factor ( PSRF ) parameter equal to 1 was obtained using the “sump burnin = 2500” command , indicating a convergence of the analysis . The resulting tree was visualized using the TreeView program [30] . Transcript abundance in eggs , females and males was quantified using the Trinity assembly output and the reads from each form as input for the Sailfish tool [31] . The number of reads was normalized using the upper quartile , correcting for the different sequencing depths of the libraries . Significant differential expression between two conditions ( egg versus female; egg versus male; female versus male ) was computed using the NOISeq program [32] with the NOISeq-sim option and the following parameters: nss = 5 , to simulate five technical replicates , each comprising 20% of the reads in the dataset ( pnr = 0 . 2 ) , allowing a small variability ( v = 0 . 02 ) . To identify contigs with significant differential expression , a probability P ≥ 95% was used as the cutoff . Next , with the list of contigs representative of each Smp , we searched for the most highly expressed genes in the egg , the female or the male . For this purpose , we identified genes that simultaneously had a significantly higher expression in eggs in the NOISeq comparison with both males and females , and we repeated the procedure , identifying the genes that simultaneously had a significantly higher expression in females than in both eggs and males , as well as the genes that simultaneously had a significantly higher expression in males than in both eggs and females; these genes were flagged on the full list of significantly differentially expressed genes as Egg_High , Female_High or Male_High , respectively . The most highly expressed genes in eggs , females or males were categorized using Gene Ontology ( GO ) terms and the Ontologizer tool [33] , with all genes detected in the transcriptome as background . GO terms for S . mansoni genes were obtained from the Metazoa Mart database ( http://metazoa . ensembl . org/biomart/martview/ ) , and p-values were calculated by the parent-child union method . To refine the S . mansoni predicted gene model structures , we mapped our Trinity-assembled contigs to the genome using BLAT [21] . We then used Bedtools utilities [34] to cross-reference the mapped coordinates of our RNA-Seq transcripts with the coordinates of the coding sequences of the Smp predicted genes [4] and flagged the Smp predicted genes that were extended at the 3′-end and/or the 5′-end . Moreover , we flagged genes in the vicinity of one another that were merged by our RNA-Seq data . Next , we compared our list of extended genes obtained above with the 5′- and 3′-UTR annotations available for 2 , 160 Smp genes [4] ( excluding the UTR annotations of another 617 Smp genes for which the UTR coordinates are inconsistent with the coordinates of the coding sequence ) , and we flagged the Smps for which our assembled contigs predicted a different or a longer UTR . To assess further evidence for the transcription start site ( TSS ) of S . mansoni genes , we downloaded the Chromatin Immunoprecipitation Sequencing ( ChIP-Seq ) dataset of adult worm Histone H3K4me3 ( SRR1107840 ) [9] , and we used the HOMER pipeline [35] to map the ChIP-Seq sequences to the genome and to find the genomic coordinates of H3K4me3 enriched peaks . HOMER found enriched peaks by calculating the density of the reads at the peaks that should be at least 4-fold higher than the peaks in the surrounding 10 kb region [35] . We then used Bedtools with a window of ± 500 bp to search for an H3K4me3 peak around the Smp gene 5′-end or around the RNA-Seq transcript 5′-end . We flagged the RNA-Seq transcripts that extended the 5′-end of Smp genes and had an H3K4me3 peak around this new 5′-end . Using the Bedtools closest function , we compared all expressed Smp genes of males and females having the H3K4me3 mark near ( within +/- 500 bp of ) the 5´-end against the dataset of expressed Smp genes without the histone mark to evaluate the prediction of gene TSS . Our gene models ( Trinity-assembled contigs ) and all the histone mark data [9] mapped to the genome are accessible through a local installation of the UCSC Genome Browser in a Box ( GBiB ) [36] at http://schistosoma . usp . br/ . Total RNA was extracted from eggs using Trizol reagent ( Invitrogen ) , according to the manufacturer's instructions , and treated with DNAse I using the RNeasy Micro Kit ( QIAGEN , Germantown , MD , USA ) . RNA from adult female and male worms was extracted and treated with DNase I using the RNeasy Mini Kit ( QIAGEN ) . Three biological replicates were assessed for each life cycle stage and form . The Reverse Transcriptase ( RT ) reaction was performed with 1 . 0 μg of each total RNA sample using the SuperScript III First-Strand Synthesis SuperMix ( Invitrogen ) . Each real-time qPCR was run in three technical replicates with Sybr Green PCR Master Mix ( Applied Biosystems , Thermo Scientific , USA ) , 160 nmol of each primer ( Forward and Reverse ) and cDNA from the reverse transcription , using the 7500 Real-Time PCR System ( Applied Biosystems ) with the default cycling parameters . Specific pairs of primers ( S1 Table ) for selected genes were designed using the Primer 3 tool ( http://biotools . umassmed . edu/bioapps/primer3_www . cgi ) with default parameters , anchoring each primer of a pair on a different exon . The housekeeping gene PAI1 ( Smp_009310 ) was chosen from nine genes that showed no differential expression in the RNA-Seq data , and the real-time qPCR data for the nine genes is shown in Fig B in S1 Text . Data were analyzed using the RefFinder tool [37] to determine a geometric mean of ranking values for each gene among the three stages and to choose the most stable gene for qPCR normalization ( lowest ranking gene ) . Real-time data were normalized in relation to the level of expression of the PAI1 gene , and p-values were determined by one-way analysis of variance ( ANOVA ) and Tukey post-hoc tests . The statistical significance of the correlation between the RNA-Seq and qPCR data was calculated using the Spearman test . Infected hamsters were maintained at the Instituto Adolfo Lutz , and the Comissão de Ética no Uso de Animais do Instituto Adolfo Lutz ( CEUA-IAL ) reviewed and approved the animal care and use protocol , license number 07/2013 . The experimental procedures were conducted according to the Brazilian national ethical guidelines for animal husbandry ( Lei 11794/2008 ) .
We performed separate sequencing of dscDNA strand-oriented libraries generated using RNA transcripts of seven-week-old S . mansoni adult male or female worms recovered from hamster portal vein perfusions and of S . mansoni eggs recovered from hamster livers . A total of ~2 . 6 million RNA-Seq reads was obtained , and reads matching transposon , mitochondrial and ribosomal genes were filtered out , resulting in ~1 . 5 million high-quality , strand-oriented , long reads , with an average length of 278 nt ( ranging from 40 to 1 , 026 nt ) ( Table 1 ) . Each of the three parasite forms was sampled ( 250 to 730 thousand reads each , Table 1 ) ; however , the egg sequencing depth was hampered because the RNA yield , purity and stability were lower than those of the adult worms . Using the Trinity de novo assembler [20] without a reference genome , we obtained 23 , 967 contigs representing the S . mansoni transcriptome from eggs and adult worms , with an average contig length of 669 nt ( ranging from 201 to 6 , 508 nt ) ( Table 2 ) . Of this total , 3 , 799 contigs represented different isoforms ( Table 2 ) assembled by Trinity that belonged to 1 , 676 putative alternatively spliced transcript fragments , i . e . , an average of 2 . 3 isoforms per contig . The remaining 20 , 168 contigs corresponded to unique transcript fragments , with no evidence of alternative splicing ( Table 2 ) . We then mapped the contigs to the genome to cross-reference the genomic coordinates with the coordinates of predicted Smp genes [4] and found that 13 , 268 contigs ( 55% of total ) matched 6 , 760 known predicted Smp genes ( Table 2 ) , an average of two contigs per targeted Smp gene . Notably , a large number of contigs ( 10 , 472 , 44% of total ) had no overlap with any predicted Smp gene ( Table 2 ) , and these contigs were found to map to intergenic regions of the genome , to intronic gene regions or to the antisense strands of Smp genes . Less than 1% of contigs ( 227 contigs ) did not map to the genome ( Table 2 ) . Some of these contigs might belong to genome sections that have not been sequenced yet . We subsequently cross-referenced the genomic coordinates of known S . mansoni retrotransposons [38] with the coordinates of the contigs that mapped outside of Smp genes ( intergenic and intronic antisense regions ) , searching for the intersection of coordinates between them . We found that despite having filtered out individual transposon reads from the input file , the assembly still contained a few contigs ( 98 out of the 10 , 472 , i . e . , 0 . 9% ) that overlapped with retrotransposon regions . We detected evidence of expression for 6 , 760 predicted Smp genes ( Table 3 ) by cross-referencing the Smp gene coordinates with the coordinates of contigs that matched those Smps ( see S2 Table for the list of genes and their expression levels ) . Of this total , 4 , 610 genes were expressed in the egg stage , 6 , 288 were expressed in females and 4 , 947 genes were expressed in males ( Fig C in S1 Text ) . Interestingly , 3 , 443 genes were expressed in all three parasite forms ( Fig C in S1 Text ) . The quantitative expression level of Smp genes for each parasite form ( egg , female or male ) was assessed by counting the number of reads from the respective library that matched to each contig , using normalized expression data to correct for the differences in sequencing depth for each dataset . Subsequently , pairwise comparisons between the stages identified the contigs with significant ( P ≥ 95% ) differential expression among the three conditions; a total of 4 , 364 contigs were detected as significantly differentially expressed in one stage compared with at least one other stage , and these differentially expressed genes are shown in Fig 1A ( see S2 Table for the list of genes and their corresponding differential expression significance ) . The number of Smp genes that were most highly expressed in one given form compared with both the other two forms was identified ( Table 3 ) , and these genes were flagged as Male_High , Female_High or Egg_High in S2 Table . By this approach , we detected 510 genes most highly expressed in males , 672 genes most highly expressed in females , and 615 genes most highly expressed in eggs ( S2 Table ) . This set of genes corresponds to a representative differential gene expression profile for each parasite form . Interestingly , we found five micro-exon genes ( MEGs ) most highly expressed in males , namely two MEG-4 genes ( Smp_085840 and Smp_163630 ) , MEG-8 ( Smp_171190 ) , MEG-11 ( Smp_176020 ) and MEG-14 ( Smp_124000 ) ( S2 Table ) . In addition , MEG-1 ( Smp_122630 ) was most highly expressed in females , while MEG-5 ( Smp_152580 ) was detected as highly expressed both in females and males . Three MEGs were most highly expressed in eggs , namely two MEG-2 genes ( Smp_159810 and Smp_180310 ) and MEG-3 ( Smp_138080 ) . To identify significantly enriched gene categories among the genes most highly expressed in eggs , females or males , we performed GO analyses . Significantly enriched GO categories ( p-value ≤ 0 . 01 ) identified gene groups related to a number of different parasite development and maintenance biological processes ( Fig 1B and S3 Table ) . Some categories were present in more than one parasite form but with a significant enrichment p-value in only one . Among the Molecular Function and Biological Process ontologies , the three most significantly enriched GO categories in eggs were carbohydrate phosphatase activity ( GO:0019203 ) , lipid transport ( GO:0006869 ) and response to stress ( GO:0006950 ) . In females , the enriched categories were cellular protein modification process ( GO:0006464 ) , DNA metabolic process ( GO:0006259 ) and catalytic activity ( GO:0003824 ) . In males , the three most significantly enriched GO categories were calcium ion binding ( GO:0005509 ) , potassium ion transport ( GO:0006813 ) and protein tyrosine kinase activity ( GO:0004713 ) . Taken together , these results point , for the first time , to relevant biological processes enriched in the S . mansoni egg stage compared with adult worms . In addition , these results indicate a set of genes involved in biological processes enriched in either S . mansoni male or female worms . In the literature , there is an absence of qPCR data comparing egg , female and male gene expression; similarly , a control housekeeping gene for S . mansoni has been previously evaluated in the literature only for mixed sex adult worms [7 , 39] . Additionally , the α-tubulin gene has been used in many studies as a housekeeping gene for normalization among all life cycle stages; however , in our RNA-Seq and qPCR data , α-tubulin is highly differentially expressed in eggs , with approximately 25 times higher expression in this stage when compared with its expression in male and female adults , as detected by qPCR ( Fig 2A ) . In this context , we chose the Plasminogen Activator Inhibitor PAI1 gene ( Smp_009310 . 1 ) as a housekeeping gene from 9 possible candidates identified by searching the set of genes expressed by all three parasite stages in the RNA-Seq data for the genes that were not differentially expressed in the three RNA-Seq pairwise comparisons of this study . We confirmed PAI1 by qPCR as an adequate housekeeping gene for use in the normalizations by performing qPCR for all 9 candidate normalizer genes and analyzing the data with RefFinder [37] , as shown in Fig B in S1 Text . Using PAI1 as a housekeeping gene , we selected and tested by qPCR six differentially expressed genes for each parasite form to provide for an independent validation of the RNA-Seq data with different biological samples . These genes were selected for their association with important biological functions in each of the parasite forms , as noted later in the Discussion . We selected glycolipid transfer protein ( GLTP ) , tubulin , translocase of outer membrane 70 ( TOM70 ) , RNA polymerase I ( PolI ) , DEAD box RNA helicase ( DDX ) and nuclear receptor SmE78 as genes highly expressed in eggs . All genes displayed higher expression in eggs when assayed by qPCR , but only four genes were statistically significant when compared with their expression in males and females; in one additional case , only the difference between eggs and males was statistically significant ( Fig 2A ) . The egg RNA-Seq data for the selected genes correlated well with the qPCR data , with a Spearman’s correlation of 0 . 77 and p-value < 0 . 001 . We selected the tyrosinases Tyr1 and Tyr2 , p14 , eggshell protein , trematode eggshell protein and ATP-binding cassette transporter as genes highly expressed in females , based on the RNA-Seq data . The qPCR confirmed the higher expression in females compared with that in eggs and males for all six selected genes ( Fig 2B ) , with a Spearman’s correlation between RNA-Seq and qPCR of 0 . 80 and p-value < 0 . 0001 . We selected Na/K ATPase , calcium binding protein ( CaBP ) , Discoidin domain receptor ( DDR ) , serotonin receptor ( 5HTR ) , Wnt5 and scavenger receptor CD36 as genes highly expressed in males , based on the RNA-Seq data . All genes displayed higher expression in males , but this difference was statistically significant when compared with the expression in eggs and females in only four of the genes . In the two remaining cases , the difference was statistically significant only when compared with either females or eggs ( Fig 2C ) . The Spearman’s correlation between the qPCR and RNA-Seq data for males was 0 . 50 , and p-value = 0 . 034 . RNA-Seq contigs were used to improve the Smp gene model predictions from the 5 . 2 version of the genome [4] , which includes a total of 10 , 852 Smp gene models . RNA-Seq Trinity-assembled contigs that mapped to the genome with genomic coordinates intersecting the coordinates of gene model exons of any predicted Smp transcribed in the same strand were considered as evidence of an mRNA transcribed from that gene . Using this approach , it was possible to document the extension of Smp genes , and these extensions could occur either at the 5′- or the 3′-end of genes . Our transcriptome data extend the 5′-ends of 3 , 337 genes and the 3′-ends of 2 , 417 genes , of which 747 were extended at both ends ( S4 Table ) . It is evident that the majority of genes were extended at the 5′-end , and this bias resulted from using RNA-Seq libraries that predominantly covered the 5′ region of RNAs , as documented by mapping the individual reads along the S . mansoni complement of genes ( Fig D in S1 Text ) . Protasio and collaborators [4] had previously improved the original Smp predicted gene models [3] by updating 731 predicted genes that were merged or split . Our transcriptome data showed that despite this gene annotation improvement , there are still an additional 589 gene models that should be altered by merging each predicted sequence with the sequence of the neighboring predicted gene ( S5 Table ) ; in each case , we found a single transcript contig that overlaps both neighboring gene predictions . A typical case involves Smp_101670 and Smp_124310 , which map to adjacent regions in the genome ( Chr_1:5 , 968 , 156–6 , 004 , 959 ) and are transcribed in the same strand; with our transcriptome data ( contig c7177_g7_i1 ) , it was possible to determine that the two predicted genes are actually part of the same transcript ( S5 Table ) . These new gene models contribute to an improved annotation of the S . mansoni gene complement . In eukaryotes , it is known that the histone H3 lysine 4 trimethylation ( H3K4me3 ) mark is present in the promoter region of expressed genes , in the vicinity of their Transcription Start Sites ( TSSs ) , and in the 5′-UTR region of genes [40] . Roquis and collaborators performed H3K4me3 chromatin immunoprecipitation followed by sequencing ( ChIP-Seq ) of mixed-sex adult worms [9] , and we used this dataset to assess the presence of enriched peaks of the H3K4me3 mark that could indicate the TSS regions in the S . mansoni genome . The H3K4me3 peaks were found near ( within +/- 500 bp of ) 3 , 084 Smp genes , showing that these gene models have their predicted 5′-end close to the promoter region ( S4 Table ) . Among these genes , 2 , 454 genes have a contig from our dataset confirming the 5′-end of the Smp gene . With the extension of the 5′-end of another 3 , 337 Smp genes ( that were re-structured using our RNA-Seq data ) , the H3K4me3 TSS histone mark was identified for a further 673 re-structured Smp genes ( S4 Table ) . Using all expressed genes found in our RNA-Seq data , we searched for the closest H3K4me3 histone mark around the predicted Smp annotation TSS regions or the extended Smp TSS regions . Unsurprisingly , for those expressed genes for which we found an H3K4me3 histone mark near the TSS ( +/- 500 bp from the TSS ) , we detected a peak centralized at the midpoint of the TSS , but also a tail extending downstream from the peak ( Fig E in S1 Text , blue line ) . For those expressed genes without the histone mark at the gene TSS ( no H3K4me3 within +/- 500 bp from the TSS ) , the closest H3K4me3 was found approximately 1–2 kbp upstream of the TSS or 1 kbp downstream of the TSS ( Fig E in S1 Text , red line ) . These distance distribution patterns are different from the distribution pattern expected by chance , obtained in a random-position-generated TSS set ( Fig E in S1 Text , gray line ) . To exemplify the scenario of improved Smp gene models , we selected a specific S . mansoni locus ( Chr_ZW: 2 , 310 , 000–2 , 350 , 000 ) , which hosts the Smp_142960 gene ( Fig 3 , blue boxes and arrowheads ) and two additional short mono-exonic genes , namely Smp_177650 and Smp_188430 ( Fig 3 , short blue boxes within intron 6 of Smp_142960 ) . Our RNA-Seq assembly generated three contigs that map to the locus with a completely different architecture ( Fig 3 , red boxes and arrowheads ) ; all three contigs new 5’-ends coincided with H3K4me3 histone marks ( Fig 3 , green peaks ) , confirming that they represent new genes with novel TSSs . The new gene on the upstream side of the locus ( contig c3145_g1_i1 ) encodes an isoprenoid biosynthesis enzyme , Class 1 domain ( cl00210 ) ( e-value = 1 . 92∙10−6 ) , and is not part of the glutathione synthase gene originally annotated as Smp_142960 . In fact , the full domain of glutathione synthase is encoded by the new gene ( contig c807_g1_i1 ) on the downstream side of the locus . It should be noted that the 3′-end of both these new genes was not covered by our RNA-Seq data , which is consistent with the fact that our new RNA-Seq method covered predominantly the 5′-ends of genes ( Fig D in S1 Text ) . This analysis cross-references ChIP-Seq information on TSS histone marks with RNA-Seq contig data and looks for evidence of an extended Smp gene prediction , where the 5′-end of the gene overlaps the TSS histone mark , and is shown here for the first time for S . mansoni , adding important information regarding the parasite’s gene regulatory regions . Using these data , it was possible to observe that the histone H3K4me3 deposition in S . mansoni frequently overlaps the first exon of the gene and extends into the first intron ( Fig E in S1 Text ) . The large number of contigs ( 10 , 472 contigs ) with no match to Smp genes raised the hypothesis of potential new genes not yet described in S . mansoni , and transcripts with open reading frames ( ORFs ) longer than 150 nt were considered as candidates for further analyses . Subsequently , using the Conserved Domains Database ( CDD ) , we investigated the similarity of the proteins encoded by these ORFs with protein-conserved domains present in proteins from other species , which identified 232 contigs encoding protein-conserved domains , thus indicating potential new genes ( S6 Table ) , among which 159 contain full-length ORFs . These contigs were deposited at NCBI TSA under accession number GDQY01000000 . For the entire set of potentially new protein-coding genes , we assessed the presence of an H3K4me3 TSS histone mark close to the 5′-end of the contigs and found 79 with evidence of this mark in the gene promoter region ( S6 Table ) . In this set of potential new S . mansoni genes , we searched for contigs encoding protein-conserved domains not yet reported among the Smp genes . Closer inspection revealed contigs encoding significantly conserved protein domains ( score < 1 x 10−5 , covering approximately 70 to 100% of the domain ) , none of which were present in any Smp gene ( S6 Table ) . Specifically , our RNA-Seq data identified an interesting new putative gene with a meiosis expressed gene ( MEIG ) protein-conserved domain , which was not predicted among Smp genes . This gene plays an essential role in the regulation of spermiogenesis in mammals [41] and was also detected in ovaries of mouse embryos when the oocytes reached the prophase I meiotic stage [42] . The SmMEIG gene encodes a protein with 85 amino acids and is orthologous to the S . japonicum MEIG gene ( accession number CAX73271 . 1 , identity = 91% and e-value = 6∙10−50 , covering 100% of the target gene ) , as determined using BLASTx with the nr NCBI protein database . This putative gene was only detected in male RNA-Seq data , but the qPCR results ( Fig 4 ) showed that this new gene is expressed in all three parasite stages; the expression in eggs is two-fold higher than that in males . This gene could play an important role in schistosome female oocyte and male sperm production , and the higher expression in eggs suggests a function not yet characterized in this parasite stage . Another new putative gene encodes the enzyme VKOR , which is responsible for the recycling of vitamin K cofactor in eukaryotes , reducing vitamin K that is oxidized upon the carboxylation of glutamic acid residues of proteins in apoptosis , signal transduction and growth control pathways [43] . This new putative gene is expressed in eggs , females and males , as detected by RNA-Seq data and qPCR ( Fig 4 ) , showing that the parasite might use this cofactor in the metabolic pathways in all three parasite forms . This putative VKOR gene is found in many mammal species , and using the BLASTx tool , we found that the S . mansoni VKOR gene is orthologous to the H . sapiens gene with an identity of 30% and e-value of 10−12 , covering 41% of the target gene . We selected other new putative genes detected in S . mansoni by RNA-Seq for testing with qPCR , such as Reticulon and BolA . These new putative genes were detected in all three parasite forms ( Fig 4 ) , confirming their expression . Reticulon proteins in eukaryotes are localized to the endoplasmic reticulum , and there is evidence that they influence endoplasmic reticulum-Golgi trafficking , vesicle formation and membrane morphogenesis [44] . The BolA gene is widely conserved from prokaryotes to eukaryotes and seems to be involved in cell proliferation or cell-cycle regulation [45] . We confirmed the presence of these putative new genes in the three parasite forms . We also detected a new S . mansoni gene from the transmembrane BAX inhibitor motif ( TMBIM ) family containing the Bax Inhibitor 1 domain ( BI-1 ) [46] . This new protein-coding gene ( contig c17331_g1_i1 ) encodes a 259-amino-acid-long protein that shares similarity with the Lifeguard members of this family , with an expected value of 0 . 003 ( 27% identity and 43% similarity ) for the alignment with a Drosophila wilistoni protein containing a BAX1_i domain . Several other hits with similar proteins containing this domain are obtained , confirming the consistency of this result . Indeed , we constructed a PSSM matrix with the new protein utilizing PSI-blast and considering the proteins aligned with e-value cutoff of 0 . 01; after a single round of iteration using this PSSM matrix , at least one hundred different Lifeguard proteins were aligned with e-values lower than 10−30 . This result indicates that although this new protein displays a relatively divergent sequence from known Lifeguard proteins , it retains residues that are conserved among members of this family . Moreover , a comparison between the transmembrane helix profile of this protein and of a known Lifeguard protein revealed a very high similarity , providing further evidence of homology between the sequences ( Fig F in S1 Text ) . Therefore , we named the new gene SmDLFG1 ( Schistosoma mansoni Divergent Lifeguard 1 ) . The RNA-Seq contig c17331_g1_i1 encoding the SmDLFG1 gene does map to two different and adjacent loci on Chromosome 1 , in a region where no gene had been predicted . The first higher-score match ( BLAT score = 888 ) is at the Chr_1:623 , 131–636 , 930 locus . The second match has some mismatches at the 5′-end of the contig and a lower matching score ( BLAT score = 790 ) . Using the genome sequence at the locus of the SmDLFG1 second match ( Chr_1:606 , 203–612 , 641 ) and RNA-Seq data from NCBI SRA , it was possible to detect reads with a 100% match , which confirmed the expression of an isoform of SmDLFG1 that we named SmDLFG2 , mapping to Chr_1:605 , 936–612 , 714 . The five Smp paralog genes from this protein family with complete BAX1_ domain ( Smp_044000 or G4VD38 , Smp_181470 or G4VEN1 , Smp_150500 or G4VGF8 , Smp_026160 . 1 or G4VKQ6 , Smp_210790 or G4V7Q6 ) and the newly identified SmDLFG1 and SmDLFG2 protein-coding genes were used in a phylogenetic analysis that included representative proteins of this family from five other schistosome species as well as from various invertebrate and vertebrate species ( Fig G in S1 Text ) . RNA-Seq detected the expression , in the three parasite forms , of all six Smp predicted gene paralogs of the TMBIM family , but apart from the SmDLFGs , only Smp_026160 ( annotated as Putative growth hormone inducible transmembrane protein ) exhibited a significantly higher expression in eggs compared with males or females . The new SmDLFGs are highly expressed in females compared with males and eggs ( Fig 4 ) ( the qPCR primers did not distinguish between the two isoforms ) . GO categorization of 198 out of the 232 protein-coding putative new genes , encoding conserved-protein domains , identified GO biological processes related to development , metabolism , cell organization and biogenesis ( Fig 5 ) . This dataset of putative new genes encoding conserved-protein domains provides an opportunity of identifying new important genes in metabolic pathways where steps remain missing . The remaining 10 , 444 contigs not encoding conserved protein domains were classified according to their protein-coding potential with the CPC tool , which identified only 703 contigs with protein-coding potential ( however with no identifiable conserved domain ) , whereas 9 , 741 contigs were classified as non-coding and therefore represented putatively expressed long non-coding RNAs ( lncRNAs ) ( S6 Table ) .
The S . mansoni genome and transcriptome have been explored for the past decade [47] , providing information related to the gene expression profiling and transcription regulation of certain life-cycle stages of the parasite . In this study we have obtained , for the first time , large-scale RNA-Seq separate profiles of S . mansoni females and males , as well as the egg-derived expression profile . We also used , for the first time , a combination of de novo transcriptome assembly with existing genome coordinates of predicted genes , along with newly mapped public ChIP-Seq data , which permitted the identification of novel putative S . mansoni genes . Because we have , for the first time , generated an individual gene profile for each of these three parasite forms , we explored this additional information by searching the literature for the possible functions of a selected set of genes most highly expressed in each form , as described below . A set of 6 genes for each of the three parasite forms was selected for validation by RT-qPCR based on the fact that most of these genes were highlighted by the functional analyses mentioned above . Twenty-nine of the thirty-six comparisons ( 81% ) performed confirmed , using RT-qPCR , the significant expression enrichment previously determined by RNA-Seq ( Fig 2 ) , a fraction of the confirmation similarly found in the literature [48] . The Spearman correlations between RNA-Seq data and RT-qPCR were in the range 0 . 50 to 0 . 80 with p-values in the range 0 . 034 to < 0 . 0001 . We consider this result to be a successful validation , especially if we note that none of the directions of enrichment in the RNA-Seq and RT-qPCR data showed a conflicting opposite result; it is already known that all transcriptome techniques , including microarray , RNA-Seq and qPCR , have inherent pitfalls that affect quantification and cannot be fully controlled [49] . First , we analyzed the differentially expressed genes in the egg stage in light of the known biological characteristics of the eggs . The glycolipid transfer protein ( GLTP ) ( Smp_076390 ) from the lipid transport GO category is highly expressed in eggs compared with males and females . GLTP is responsible for the intermembrane transfer of lipids linked to sugars , such as glycosphingolipids [50] . As schistosomes are unable to synthesize fatty acids , the uptake of these compounds from the host is essential . We suggest that the enrichment of lipid transport genes in the egg stage might be related to an important uptake of lipids from the host , possibly related to embryo development . In fact , it is known that the schistosome egg uptakes cholesteryl ester from HDL vesicles , mediated by the CD36 receptor , and that this uptake is important for egg maturation [51] . In this context , we speculate that because the CD36 receptor has such an important function in egg development , the enrichment of lipid transport genes among the highly expressed egg genes could be related to the transport function . Thus , GLTP could be an important gene for egg development and maintenance inside the host circulation until the eggs reach the host intestine lumen , where they are released . Egg formation is dependent on structural components such as microtubules , an enriched GO category . An important gene from the microtubules category is tubulin ( Smp_027920 ) , whose expression was 17-fold greater in eggs compared with that in females and 34-fold greater than that in males . Microtubule genes enriched in eggs will integrate the layer between the eggshell and the developing miracidium , known as Reynolds’ layer , comprising microfibrils in a granular matrix [52] . Genes related to stress response were also increased in eggs , and in this group , we highlight the mitochondrial genes from the translocase of outer membrane ( TOM ) machinery that serve as the main entry gateway of preprotein into the mitochondria [53] . The mRNA expression of TOM70 ( Smp_010930 ) and of HSP70 ( Smp_065980 ) was detected in eggs , and the proteins encoded by these genes are key to the mitochondrial import pathway [54] and contribute to the folding and refolding of proteins after stress denaturation [55] . The mRNA level of Pol I ( Smp_129500 ) , which encodes the enzyme responsible for the transcription of ribosomal DNA ( rDNA ) [56] , is significantly higher in eggs than in males , suggesting ribosome production and cellular proliferation inside the egg . Another gene with a high expression level in eggs encodes a protein related to transcription , namely the DEAD box RNA helicase ( Smp_013790 ) . DEAD box proteins unwind short duplex RNA and remodel RNA-protein complexes , and they are important players in RNA metabolism from transcription and translation to mRNA decay [57] . The nuclear receptor SmE78 gene ( Smp_000340 ) also showed an increased expression in eggs . Wu et al . had previously shown that nuclear receptors are important transcriptional modulators , and in S . mansoni , SmE78 may be involved in growth and vitellogenesis [39] . To further validate the gene enrichment analysis of the egg stage , which has the lowest RNA-Seq coverage , we have manually inspected three genes known to be highly expressed in eggs , namely omega-1 [58] , IPSE/alpha-1 [59] and kappa-5 [60] . Only the omega-1 and IPSE/alpha-1 genes are annotated as Smp genes , namely Smp_193860 and Smp_112110 , respectively . Indeed , they are listed in S2 Table as highly expressed genes in the egg stage with the flag Egg_High , as these two genes were detected in our RNA-Seq data as significantly differentially expressed in eggs compared with males and females . Interestingly , the omega-1 gene ( Smp_193860 ) , which encodes a protein with 127 amino acids , appears in our RNA-Seq data as contig c7716_g1_i1 with an additional five new exons at the 5’-end of the gene , and the new longer ORF encodes an omega-1 protein with 236 amino acids , which is now compatible with its described molecular weight of 31 kDa [58] . The kappa-5 gene with accession AY903301 . 1 [60] is not yet annotated as an Smp gene; only a paralog , Smp_150240 is annotated in the genome , with an identity of 89% and query coverage of 74% compared to the kappa-5 gene . Because the kappa-5 gene is not present in the Sanger Genome annotation , we did not detect it in the automated analysis , but by a curated manual investigation , using the NCBI accession AY903301 . 1 , it was possible to align the sequence to the genome with genomic coordinates Chr_3:13873256–13879312 minus strand; we found that contig c6481_g1_i1 is the transcript corresponding to the kappa-5 gene . This contig is highly expressed in the egg stage , with normalized egg read counts of 1 , 909 . 35 and no reads detected in the adult male and female forms . Subsequently , we analyzed the genes differentially expressed in females in light of the known biological characteristics of the females . Their transcriptome profile is highly linked with egg production , as the S . mansoni female produces approximately 350 eggs daily , and proteins that are important to the eggshell structure should be expressed . The hardened and tanned structure of the eggshell is derived from tyrosinase activity , which catalyzes the cross-linking of proteins known as quinone tanning [61] . The mRNA expression levels of Tyrosinase 1 ( Smp_050270 ) and Tyrosinase 2 ( Smp_013540 ) in females are higher than in eggs and males . Interestingly , Tyr1 has a much higher expression level than Tyr2 , although the two genes encode highly similar proteins . The higher abundance in females is consistent with the fact that tyrosinase originates from the female vitelline cells inside the vitellaria , acting on eggshell formation . The most studied S . mansoni eggshell proteins are p14 and p48 . Chen et al . showed that the p14 gene is expressed only in mature female vitelline cells and is undetectable in the RNA obtained from eggs [62] . Our RNA-Seq results show that the p14 ( Smp_131110 ) transcript was detected at low levels in eggs extracted from hamster liver , and the low abundance of this transcript in eggs was confirmed by qPCR . A high expression level of p14 was detected in females compared with males or eggs , as expected . Two other proteins that were not yet shown as part of eggshell synthesis are eggshell protein chorion ( Smp_000430 ) and Trematode eggshell protein ( Smp_000390 ) , both with the Trematode eggshell domain ( Pfam:08034 ) . We found that the genes encoding these two proteins are highly expressed in females , and compared with p14 , the expression of eggshell protein chorion gene is twice as high , pointing to a new protein for possible exploration as a candidate antigen in liver granuloma formation . Female reproduction requires a high bioenergetic supply for the production of hundreds of eggs per day . Consistently , genes from the ATPase activity category ( Fig 1B ) are enriched in our analyses of female highly expressed genes , including genes related to the production of ATP through mitochondrial oxidative phosphorylation . Additionally , the ATP-binding cassette transporter gene ( Smp_040540 ) exhibited higher expression in females than in males , but interestingly , the expression level in eggs was almost the same as that in females . In addition to the energy supply for egg production , it was recently proposed that enzymes present in the schistosome tegument could act similarly to human cell-surface ATPases , inhibiting platelet activation and modulating the host coagulation mechanism [63 , 64] . Finally , we analyzed the differentially expressed genes in adult males in light of the known biological characteristics of the males . Their transcriptome was already individually explored by RNA-Seq [6]; however , the different transcriptome profiles of the two sexes that reflect their distinct biological characteristics were not previously compared . We found that a number of genes most highly expressed in males are involved in the regulation of transmembrane transport ( GO:0034762 ) , membrane ( GO:0016020 ) and heparin sulfate proteoglycan binding ( GO:0043395 ) , probably belonging to the male tegument . This result is consistent with the fact that the male body is more highly exposed to the host immune cells in the circulation than the female , and consequently , the tegument renewal rate is higher in the male than in the female [65] . Among the most highly expressed genes in males , we found an enrichment of the potassium ion transport category ( GO:0006813 ) , and the high expression of the sodium-potassium ATPase gene ( Smp_124240 ) in males compared with in females was confirmed by qPCR . The surface plasma membrane is involved in nutrient uptake , involving several amino acid and sugar transporters , aquaporins , anion selective channels and Na+/K+ and Ca2+ ATPases [66 , 67] . We detected enriched GO categories among male genes associated with the schistosome muscle layer , such as the troponin complex and the calcium binding protein Sm20 gene ( Smp_005350 ) , whose expressions were higher in males than in females , according to both the RNA-Seq data and qPCR . Our RNA-Seq data showed that Smp_005350 is fully transcribed; this gene encodes a 58-kDa calcium binding protein with distinct Ca2+ binding motifs . It should be noted that Mohamed et al . studied a distinct 20 . 8 kDa antigenic S . mansoni calcium binding protein encoded by the U91941 cDNA clone and named Sm20 . 8 [68] , subsequently renamed Smp_086530 , which is distinct from the Sm20 gene ( Smp_005350 ) detected here as abundant in the male . Receptor activity and protein tyrosine kinase activity were GO categories enriched in male-expressed genes , such as Discoidin domain receptor DDR ( Smp_133250 ) , Serotonin receptor 5HTR ( Smp_126730 ) and Wnt5 ( Smp_145140 ) , encoding a signaling molecule . This finding suggests that a characteristic cell signaling process might operate in the male . The scavenger receptor CD36 antigen transcript ( Smp_011680 ) , which is a mediator of cholesteryl ester uptake by adult worms [51] , was detected by qPCR in all three forms , with males and females having the same level of CD36 expression , in disagreement with Fitzpatrick et al . , who detected the CD36 antigen transcript only in female worms [7] . The divergent result from Fitzpatrick et al . could be explained by the different parasite strain ( Puerto Rican strain in the Fitzpatrick et al . study , BH strain in this study ) . Micro-exon genes ( MEGs ) constitute a large family characterized in the parasite genome , each gene with multiple symmetrical exons , arranged in tandem , with lengths that are a multiple of three nucleotides ( from 6 to 36 nucleotides for each exon ) [3 , 69] . This arrangement leads to protein variation through alternative splicing [69] , and may have an impact on the escape of parasites from host defenses . Differential expression of MEGs between males and females has only been observed with microarrays in a comparison of male esophagus with female gastrodermis [70] . Here we compared eggs , female and male adult worms and identified stage-specific MEGs , most highly expressed in each of the three forms , supporting the idea that MEGs might be specifically modulated in response to defenses of the host . Interestingly , MEG-4 and MEG-14 , most highly expressed in males , have been previously described using whole mount in situ hybridization to be specifically expressed in the esophagus of adult worms from both sexes [71] . Male and female highly expressed MEG-5 had its protein product previously detected in tegumental preparations [69] . Egg highly expressed MEG-2 and MEG-3 protein products were detected in egg secretions and only the latter was also detected in schistosomula secretions , being produced in the head gland [69] . These observations suggest that either the expression of esophageal MEGs is more robust in males than in females or that in the males this organ contributes with a higher amount of mRNA for the total pool . We used the transcriptome data as a guide to improve Smp gene predictions , adding UTRs or new coding exons to hundreds of Smp genes . Moreover , we cross-referenced the genomic coordinates of the genes with the genome-wide coordinates that we obtained by genome-mapping the publicly available H3K4me3 adult worms ChIP-Seq dataset [9] , and we confirmed the recently described presence in adult schistosomes of the H3K4me3 mark at the TSSs of thousands of Smp predicted genes [72] . More importantly , we identified the presence of the H3K4me3 mark at the TSSs of 79 out of 232 putative novel protein-coding genes identified by our RNA-Seq , which provided additional evidence of the regulation of these novel putative genes by histone modification . We also found the H3K4me3 TSS histone mark for another 525 re-structured Smp gene models ( out of 2 , 083 Smp gene models with their 5′-ends extended by our RNA-Seq data ) , thus providing additional confirmation that the original gene model predictions for those 525 genes were improved by using this new 5′-end information . For those genes whose predicted TSS genome positions do not intersect with H3K4me3 peaks , the distance to the closest H3K4me3 peak could indicate the most probable TSS positions [40]; for the Smp predictions that are distant from a H3K4me3 peak , we found that this distance was between 1–2 kbp upstream from the predicted TSS . Finally , the novel protein-coding genes detected in our transcriptome data are complementary to the existing S . mansoni gene predictions and annotations , allowing the discovery of genes with possible biological relevance to the parasite . We suggest that among other genes , the SmMEIG gene should be investigated as a potential regulator of parasite sexual reproduction and egg laying . Attention could also be given to the divergent Lifeguard genes SmDLFG1 and 2 , two possible inhibitor regulators of cell apoptosis . Interestingly , examining the genomes and gene predictions of other platyhelminths did not provide any evidence of close orthologs , suggesting that this protein would be specific to schistosome species among the platyhelminths . By examining the phylogenetic tree ( Fig G in S1 Text ) , it is possible to observe that the branch that contains the two new SmDLFGs from S . mansoni contains no other protein from species outside the Schistosoma genus , while its branch length is very long . Therefore , it appears to be an isoform specific to this genus that was probably subjected to a positive selection process at some point during its evolution . A possible explanation for this phenomenon would be the co-optation of this protein for some process related to host-parasite interaction , as the new SmDLFGs are transmembrane proteins ( see Fig F in S1 Text ) ; in fact , it was previously demonstrated in the Rickettsiaceae model that genes coding for parasite membrane proteins tend to display positive selection in relation to genes for membrane proteins of non-parasites [73] . The finding that the new SmDLFGs are highly expressed in females compared with males and eggs ( Fig 4 ) is interesting because the genes from this family are described in humans as encoding an apoptosis inhibitor protein; females are known to be more resistant to drug treatments than males , and the high expression of a possible apoptosis inhibitor protein should be considered when repurposing apoptosis-inducing cancer drugs to treat schistosomiasis [74] . Other putative new genes and gene fragments detected here are likely involved in different biological processes , such as metabolism , development , morphogenesis and cell communication , which raises the possibility of finding and confirming missing genes in the parasite’s molecular pathways . We also confirmed in this study that thousands of lncRNAs ( > 200 nt ) are transcribed in S . mansoni [6 , 75] , and many of these non-coding RNAs may possibly exert regulatory functions that have yet to be explored [75] . In fact , Guttman and co-workers [76] showed that in mammals , there is a class of conserved lncRNAs ( i . e . , the large intervening non-coding RNAs , lincRNAs ) whose transcription and processing appear to be similar to protein-coding genes , with Pol II transcription , 5´-capping and poly-adenylation . Since then , a number of other studies have confirmed that most lncRNAs are poly-adenylated , although there are also non-poly-adenylated ones . Our RNA-Seq cDNA libraries have captured polyA+ RNAs , and the informatics analysis of the sequenced RNAs has identified thousands of long transcripts with non-coding potential , thus providing additional evidence for the existence of such poly-adenylated long noncoding transcripts in S . mansoni . Cross-reference with known transposon sequences showed that these transcripts do not originate from transcribed repeats . In fact , these lncRNAs are part of the polyA+ RNA pool of a normal eukaryotic cell , and in human cancer cells one of the best characterized consequences of an altered expression of lncRNAs is an important change in deposition of regulatory histone marks at the promoter regions of oncogenes and tumor-suppressor genes [77] . The lncRNA sub-population of a cell , which has long been overlooked , is now being studied by a wide variety of methods [78] that are revealing lncRNAs as an important layer of information that in many instances controls transcription , translation , imprinting , histone modifications , among other functions in the eukaryotic cells . Additionally , the data discussed in this work are available in a new S . mansoni Genome Browser Database ( http://schistosoma . usp . br/ ) , which uses the UCSC Genome Browser in a Box platform [36] and provides a S . mansoni public data-searching tool ( using the option “Tools > Blat” with a query sequence , or entering a gene name in the enter position or search term window ) , a convenient data-downloading tool ( using the option “Tools > Table Browser” ) and a visualization tool in a user-friendly format .
|
Schistosomiasis is a public health problem caused by parasites of the genus Schistosoma , of which S . mansoni is the primary causative agent . The parasite has a complex life cycle; their sexual reproductive stage is dependent on female and male adult worms mating inside the mesenteric circulation of the human host , with the female releasing hundreds of eggs daily . This phase of the life cycle is responsible for the development of pathology , which is proportional to the number of eggs accumulating in the liver and intestine of the human host . Genome and transcriptome sequencing of this parasite represent important advances in schistosome research , but there is still a need for integrated analyses to better understand the biology of this parasite . In this study , we describe the first large-scale transcriptomes of eggs , and female and male adult worms , the parasite forms that are mainly responsible for the pathology of schistosomiasis . We were able to cross-reference the gene transcription regions with promoter regions , thus improving the gene annotations . Moreover , we identified the expression of novel protein-coding genes not yet described in the current genome annotation , advancing the biological knowledge regarding this parasite .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Schistosoma mansoni Egg, Adult Male and Female Comparative Gene Expression Analysis and Identification of Novel Genes by RNA-Seq
|
While the majority of multiexonic human genes show some evidence of alternative splicing , it is unclear what fraction of observed splice forms is functionally relevant . In this study , we examine the extent of alternative splicing in human cells using deep RNA sequencing and de novo identification of splice junctions . We demonstrate the existence of a large class of low abundance isoforms , encompassing approximately 150 , 000 previously unannotated splice junctions in our data . Newly-identified splice sites show little evidence of evolutionary conservation , suggesting that the majority are due to erroneous splice site choice . We show that sequence motifs involved in the recognition of exons are enriched in the vicinity of unconserved splice sites . We estimate that the average intron has a splicing error rate of approximately 0 . 7% and show that introns in highly expressed genes are spliced more accurately , likely due to their shorter length . These results implicate noisy splicing as an important property of genome evolution .
Most mammalian mRNAs are processed from much longer precursors in a series of splicing reactions . Regulation of these splicing reactions can lead to alternatively spliced forms of mRNA from the same pre-mRNA [1] , and there is considerable interest in cataloguing the functionally important transcripts of all mammalian genes . Towards this end , transcript diversity has been examined using data from full mRNA sequences , expressed sequence tags ( ESTs ) , or high-throughput sequencing of cDNA libraries ( RNA-Seq ) [2]–[6] . In particular , recent RNA-Seq studies have established that nearly all multiexonic human genes have multiple detectable isoforms [2] , [5] . The observation of extensive alternative splicing could indicate that most genes have many functionally-relevant isoforms; alternatively , many transcripts could be nonfunctional “noise” [7]–[10] . The latter explanation is supported by a few pieces of evidence from analyses of EST databases . In particular , a large fraction of exon-skipping events in human genes is not observed in mice ( ie . is not conserved ) [11]–[13] , and the number of observed isoforms of a gene correlates with the number of exons it has ( and thus the theoretical number of potential transcripts it could produce ) [8] . Additionally , it is hypothesized that short introns in humans ( as well as in other eukaryotes ) have evolved to preferentially trigger degradation via nonsense-mediated decay ( NMD ) mechanisms when the spliceosome fails to remove them , suggesting that such errors are common enough to exert a detectable selective pressure [14] . There are also theoretical reasons to expect splicing to be error-prone . First , the binding sites for proteins important in exon recognition comprise a large mutational target; it has been estimated that approximately 30 bases are necessary for fully efficient splicing of an intron [15] . The potential for mutational disruption of these binding sites has been referred to as part of the “intrinsic cost of introns” [16] . Further , the large size of introns relative to exons gives ample opportunity for the mutational creation of new ( weak ) binding sites . Although mutations which create or disrupt binding sites may be slightly deleterious , the large number of possible such mutations makes it inevitable that some will reach fixation in a population . This is particularly relevant in species , such as humans , with relatively small long-term effective population sizes . It is plausible , then , that the human genome carries a substantial load of suboptimal sequences which cause the generation of aberrant transcript isoforms . In this study , we present direct evidence that this is indeed the case .
We compared the junctions we identified to gene models from the UCSC , Ensembl , Vega , and RefSeq databases , and to spliced ESTs from Genbank [23]–[27] . Of the 306 , 606 splice junctions , 154 , 927 ( 50 . 5% ) are not annotated as parts of known gene models , and 136 , 313 ( 44 . 5% ) are not present in Genbank . For splice junctions not present in gene models , we estimate an FDR of about 2% ( Methods ) . The extensive unannotated splicing we observe is largely due to junctions that are rarely seen in our data ( Figure 1B ) : while 50% of all observed junctions are not present in gene models , these account for only 1 . 7% of all junction-spanning sequencing reads ( Table 1 ) . For example , 21 of the 32 splice junctions observed in the gene HERPUD1 are unannotated , but only around 0 . 5% of the reads from this gene are derived from these 21 unannotated junctions ( Figure 2 ) . We see no sign that our identification of isoforms is near saturation ( Figure S1 ) ; thus deeper sequencing of transcriptomes will likely continue to identify additional low-abundance isoforms . Next , we quantified overall levels of alternative splicing . To do this , we considered a set of splice sites covered by at least 50 reads in our data ( there are 77 , 754 such 5′ splice sites and 77 , 733 such 3′ splice sites; of these are annotated in gene model databases ) . We then counted both the number of different places in the genome to which each site is spliced , as well as the proportions of reads covering each junction . We estimate that the “major” splice form accounts for 98 . 4% of reactions involving each splice site; in our data the average splice site is involved in 1 . 8 different splicing reactions . 5′ splice sites in untranslated regions ( UTRs ) are involved in a mean of 2 . 6 splicing reactions , versus 1 . 8 for 5′ splice sites in protein-coding regions; the corresponding numbers are 3 . 2 and 1 . 7 for 3′ splice sites . We then evaluated how the splice junctions correspond to known gene models . We split the junctions into five classes: ( i ) those where the junction is annotated; ( ii ) those between splice sites that are both annotated ( but not annotated as being spliced together ) ; ( iii ) those where the 5′ splice site ( but not the 3′ splice site ) is annotated; ( iv ) those where the 3′ splice site ( but not the 5′ splice site ) is annotated; and ( v ) those where neither splice site is annotated . 80% of the unannotated junctions involve at least one annotated splice site , and though many newly-identified splice sites fall near annotated sites , the majority do not ( Table 1 ) . This indicates that a large fraction of the low-abundance isoforms are not modifications of known exons , but instead contain entirely new exons . We next asked whether we could confirm these observations in other cell lines and primary tissues . We first performed the same analysis on RNA-Seq data generated on a different set of human LCLs [28] . We identified 219 , 322 splice junctions at an FDR of 1 . 5% , 82 , 658 of which are unannotated ( Figure S1 ) . We then analyzed RNA-Seq data from primary human liver samples ( George Perry , unpublished data ) ; we identified 156 , 905 splice junctions at an FDR of 0 . 8% , 29 , 655 of which are unannotated ( Figure S1 ) . Finally , we analyzed RNA-Seq data from a set of several primary tissues [2]; we identified 136 , 499 splice junctions at an FDR of 1 . 8% , 21 , 743 of which are unannotated ( Figure S1 ) . The numbers of unannotated splice junctions in these studies are roughly consistent with the observed numbers found in the Yoruba data , given the lower sequencing depths of those other studies and hence better sampling of common junctions relative to rare junctions ( Figure S1 ) . This confirms that the observation of extensive unannotated splicing is broadly generalizable; for the rest of the paper we focus on the original set of Yoruban LCL data since this is the largest RNA-Seq dataset currently available . We also considered how much splicing shows evidence of being restricted to particular individuals , rather than shared across the entire population ( due to , for example , sequence polymorphisms which influence splicing [17] , [28]–[30] ) . Though it is difficult to estimate this precisely , several analyses suggest that only a couple percent of splice junctions , at maximum , show evidence of being restricted to certain individuals ( Text S1 ) . We hypothesized that unannotated , rarely-used splice sites are the result of evolutionarily-neutral ( or perhaps slightly deleterious ) splicing errors . To test this , we compared the sequence conservation across placental mammals ( using the phyloP score [31] ) between the unannotated and annotated splice sites ( we assume that current gene databases are highly enriched for truly functional exons ) . If the unannotated splice sites are functionally relevant , their sequence conservation should be comparable to that of annotated splice sites . For this analysis , we used the set of splice junctions where one end is an annotated splice site and the other is more than 50 bases away from an annotated splice site . Annotated splice sites show a striking pattern of sequence conservation–the splice sites themselves are highly conserved , as are the regions exonic of the splice sites ( Figure 3A ) . In contrast , the conservation scores around unannotated splice sites show little , if any , signal of evolutionary constraint ( Figure 3B ) . This same pattern holds in an independent sample of LCLs and in primary human liver samples ( Figures S2 and S3 ) . To exclude the possibility that our analysis overlooked splice sites that are conserved in only a subset of placental mammals , we repeated this analysis using phyloP scores calculated using only primates , and saw the same pattern ( Figure S4 ) . Further , humans have reduced polymorphism in the annotated splice sites , but no such reduction in the unannotated splice sites ( Figure S5 ) . The most parsimonious explanation of these observations is that the majority of rarely-used , unannotated splice sites are simply due to mis-spliced transcripts . To estimate the fraction of mature mRNAs resulting from mis-splicing , we identified a set of splice junctions where both the 5′ and 3′ splice sites are highly conserved ( those with a phyloP score ) . We then asked how often the conserved splice sites are spliced to unconserved splice sites ( to provide a conservative lower bound on the amount of mis-splicing , we used a relaxed threshold of a phyloP score 0 . 5 to call a splice site as unconserved for this analysis ) . Approximately 0 . 7% of reads involving either end of a conserved junction are to unconserved splice sites . Given that the median gene in the human genome has four exons ( and thus three splicing reactions ) , this suggests that approximately 2% of transcripts from the average gene are mis-spliced . Because mis-spliced transcripts are preferentially removed by NMD mechanisms , this is likely a conservative estimate . We next tried to identify factors that are predictive of an intron's level of splicing error . The two factors we considered were the intron's length and the expression level of the gene in which the intron falls . Longer introns show higher levels of mis-splicing ( Figure 4 ) , while highly expressed genes show somewhat lower levels ( Figure S6 ) . These associations may be confounded , however , by the fact that highly expressed genes tend to have shorter introns [32] . Indeed , the association between splicing error rate and gene expression level disappears after correction for intron length ( Figure S5 ) , indicating that this association is largely driven by the lower splicing error rate of small introns . The different sequence composition of introns in highly expressed genes [33] may also influence their lower rate of splicing error ( Text S1; Figure S7 ) . Finally , we considered the mechanism which results in the generation of mis-spliced transcripts . To do so , we looked for hexamers enriched in the vicinity of unconserved , rarely-used splice sites , as compared to nearby “decoy” splice sites ( matching GT or AG ) which we never observed to be used in our data . 574 hexamers show significant enrichment or depletion exonic of the 5′ SS , and 728 exonic of the 3′ splice site ( Figure 5A; ; test ) . Although the relative enrichments of hexamers near the 5′ and 3′ splice sites are similar , a set of hexamers matching the binding site for the U1 snRNP ( which recognizes the 5′ splice site ) is strongly depleted in the vicinity of the 5′ splice sites , but shows only limited or no depletion in the vicinity of 3′ splice sites ( Figure 5A ) . This is consistent with the observation that competitive binding of the core splice factors plays an important role in splice site choice [34] . There is a smaller , but still substantial , number of hexamers enriched or depleted intronic of the splice sites ( 295 and 282 for 5′ and 3′ splice sites , respectively ) . We compared our list of hexamers to the list of exonic splicing enhancers ( ESEs ) identified by Fairbrother et al . [35] , some of which were validated experimentally , and found that 79% of the ESEs are enriched exonic of both 3′ and 5′ splice sites in our data ( , test ) . This is evidence that noisy splicing is a result of the binding of the same splice factors used to identify exons more generally . Further support for this possibility comes from the observation that the hexamers identified near noise splice sites demarcate the boundaries of constitutive exons in our data ( Figure 5B , 5C ) .
Two main observations support the contention that the majority of low-abundance isoforms are due to splicing errors . First , rarely-used splice sites are enriched near often-used splice sites , and additionally show a periodic pattern around those sites ( [20] , [21] , Figure 1 ) . Our interpretation of these observations ( like that of Dou et al . [20] and Chern et al . [21] ) is that the splicing machinery occasionally misses its “intended” splice site , and that the resultant isoforms which disrupt the protein-coding reading frame are preferentially degraded by the NMD machinery . Second , rarely-used , unannotated splice sites ( we interpret the fact that a splice site is annotated as evidence that it is relatively highly expressed in at least one tissue ) show little evidence of evolutionary conservation across placental mammals ( Figure 3 ) or primates ( Figure S3 ) , or of constraint within humans ( Figure S4 ) . We conclude , then , that the majority of low-abundance splice forms are indeed “noise” . In fact , because our identification of splice forms is not yet at saturation ( Figure S1 ) , we extrapolate that the majority of different mRNA isoforms present in a cell are not functionally relevant , though most copies of a pre-mRNA produce truly functional isoforms . We speculate that this conclusion will hold at the protein level as well . One goal in understanding the mechanism of splicing is the generation of a “splicing code”–a set of rules which the cell uses to convert the information present in a pre-mRNA sequence to a properly spliced mRNA [36]–[38] . This code presumably involves the binding sites for various splice factors [35] , [39] , [40] , as well as local chromatin structure [41]–[46] . The results presented here suggest that , instead of being a deterministic function that maps a pre-mRNA sequence to a spliced transcript , the “splicing code” instead defines a probability distribution for each pre-mRNA on a large number of possible splice forms . That is , the same pre-mRNA sequence will stochastically result in a large number of isoforms , presumably even within the same cell . We suggest that low-probability events may be informative about the parameters of this distribution . The level of splicing error observed in a gene reflects a balance between the continuous input of mutations that disrupt splicing and the ability of selection to remove them [16] . Indeed , selection for proper discrimination between introns and exons has affected genome evolution in a number of ways , by constraining the composition of amino acids coded near splice sites [47] and influencing the sequence composition of introns [14] , [48] . We have shown that longer introns are more prone to splicing errors . This is consistent with the increased rate of birth of new , alternatively spliced exons in long introns [49] , [50] , and supports the contention that long introns are more deleterious than short introns [51] , [52] . This may contribute to selection for short introns in highly expressed genes [32] . One implication of the above reasoning is that the level of splicing error observed in an organism should depend on the complexity of the splicing machinery in the species ( ie . , the number of potential mutations that could affect splicing ) and the effective population size of the species ( and hence the effectiveness of natural selection in removing those mutations ) . This is consistent with the observation that levels of alternative splicing vary considerably across eukaryotes [53] . A prediction , then , is that species with larger effective population sizes or simpler splicing mechanisms should have lower rates of splicing error . Progress in RNA-Seq technology will soon allow relatively unbiased exploration of the evolution of splicing noise in a wide variety of species , and allow propositions such as these to be tested .
For the main analysis of LCLs , we used RNA-Seq data generated on 75 HapMap cell lines derived from Yoruban individuals . Data from 69 of these were reported in Pickrell et al . [17] . We also generated RNA-Seq data on six additional HapMap Yoruban cell lines , using the same protocol as in Pickrell et al . [17] . Each cell line was sequenced in two lanes on the Illumina GA2 platform , one lane at the Yale sequencing center using 35 base pair sequencing reads , and one at the Argonne sequencing center using 46 base pair sequencing reads . The cell line identifiers and basic quality control metrics are presented in Table 1 in Text S1 . All data are available at http://eqtl . uchicago . edu . For the analysis of an independent set of human LCLs , we obtained RNA-Seq data on a set of 60 HapMap cell lines derived from individuals of European descent [28] . These data consist of paired-end 37 base pair reads generated on the Illumina GA2 platform , obtained from http://jungle . unige . ch/rnaseq_CEU60/ . We treated each end of a paired-end read independently , and used the protocol described below to identify splice junctions . For the analysis of primary human liver cell , we used RNA-Seq data from four human liver samples ( G . Perry , unpublished ) . These data consist of paired-end 76 base pair reads generated on the Illumina GA2 platform . RNA-Seq data from multiple other tissues were obtained from Wang et al . [2] and Wu et al . [54] . All data were processed in the same manner , described below . Here , we describe our approach for de novo identification of splice junctions , which is similar to previous approaches [17] , . Software is available at http://eqtl . uchicago . edu . By doing the mapping in this way , we expect to recover approximately 55% of all the 46 base pair reads that span splice junctions ( outside repetitive regions of the genome ) , and approximately 30% of all the 35 base pair reads that span splice junctions ( there are 10 allowable breakpoints in a 35 base pair read , and 26 allowable breakpoints in a 46 base pair read ) . We note that we are also limiting ourselves to identifying introns with a maximum length of 20 kb; this is sufficient for the majority of introns in humans . In Figure 1 , we show the density of splice sites near known protein-coding sites . In this analysis , we used only splice sites annotated as being protein-coding in all the transcripts of the gene in Ensembl and Refseq . To generate this figure for the 3′ splice sites , we identified all the 5′ splice sites covered by at least 20 reads , spliced to at least two 3′ splice sites , and where one of those 3′ splice sites contributed % of all the reads from that 5′ splice site . We will call that 3′ splice site the “major” splice site , and the other the “minor” splice site . We then recorded all the positions of matches to AG in the region surrounding the “major” splice site . For each distance from the “major” splice site , we can then count the number of “minor” splice forms at that distance , as well as the number of AG dinucleotides that would lead to a splice site at that distance . The ratio of those two numbers is plotted in Figure 1 . We excluded the positions from −2 to +2 from the splice site due to ambiguity in the read mapping . Analysis for the 5′ splice sites is analogous . We downloaded the Ensembl , UCSC , Vega , and RefSeq gene models from the annotation of hg18 in the UCSC Genome Browser on Dec . 31st , 2009 . We downloaded the spliced EST track on March 1st , 2010 . Throughout the paper , when we refer to an “annotated” splice junction , we mean one present in at least one of the Ensembl , UCSC , Vega , or RefSeq databases as of that date . In several places in the paper , we use presence of a splice site in these databases as a proxy for function ( rather than , for example , read depth in our data ) . This is supported by our analysis of conservation ( Figure 3A ) ; even rarely-used splice sites in LCLs which are present in gene databases show high levels of sequence conservation ( Figure S8 ) . This is likely due to that fact that some fraction of rarely-used splice sites in LCLs are abundantly used in some other tissue , and thus are both functionally relevant and annotated in current databases . We can estimate the FDR for the junctions we identified by considering how often each junction is consistent with a GT-AG intron versus control pairs of dinucleotides ( recall that each junction read is usually consistent with several pairs of potential splice sites ) . Of the 392 , 612 junctions initially identified , 306 , 606 are consistent with GT-AG or GC-AG ( 298 , 346 are consistent with the former , and 8 , 260 with the latter ) . In contrast , 4 , 230 are consistent with control dinucleotides GT-TC or GC-TC ( note that the controls simply contain the complement of the 3′ splice site consensus dinucleotide ) . If we assume that all of the controls are false positives , this gives an FDR of 1 . 4% ( 4 , 230/306 , 606 ) . If we restrict ourselves only to the 240 , 644 splice junctions that have not been previously observed , 154 , 927 are consistent with GT-AG or GC-AG , and 2 , 985 are consistent with the control pairs of dinucleotides . This gives an FDR for the set of unannotated junctions which contain intronic matches to GT-AG or GC-AG as 1 . 9% ( 2 , 985/154 , 927 ) . We saw no evidence of enrichment for AT-AC introns , and so did not consider them . In the analysis of sequence conservation , we used phylop scores generated on the 44-way vertebrate alignment , and downloaded from the UCSC Genome Browser [31] . There are three sets of scores available ( representing scores for constraint in all vertebrates , all placental mammals , and all primates ) ; in the analysis presented in the main text , we used scores generated using the placental mammals . For each intron with highly conserved splice sites , we counted the fraction of reads from either splice site to an unconserved splice site , as described in the main text . To estimate the expression level of each gene , we used both our RNA-Seq data and that from the different tissues assayed by Wang et al . [2] . For each tissue , we divided the number of reads mapping to exons of each gene by the length of the exons of the gene . We then took the maximum expression level across tissues as the expression level of the gene for this analysis . To look for hexamers enriched around “noise” splice sites , we used a set of splice junctions where one end of the splice junction is to an annotated splice site , and the other is to an unannotated , unconserved ( phyloP score ) splice site . To be conservative , we removed all the unconserved splice sites within 50 bases of an annotated splice site . Then we identified a set of control sites–for each unconserved splice site , we found an unused GT ( or AG ) dinucloetide between the used splice site and the nearest annotated one . We then extracted 100 bases around both the used splice sites and the control sites , and counted the frequencies of hexamers in each class ( for the 3′ splice site , we excluded the 20 bases intronic and 2 bases exonic of the splice site from this analysis; for the 5′ splice site , we excluded 5 bases intronic and 5 bases exonic ) . We did this separately for 3′ and 5′ splice sites , and for the intronic and exonic regions of both types of site . Significance was assessed by a test .
|
Most human genes are split into pieces , such that the protein-coding parts ( exons ) are separated in the genome by large tracts of non-coding DNA ( introns ) that must be transcribed and spliced out to create a functional transcript . Variation in splicing reactions can create multiple transcripts from the same gene , yet the function for many of these alternative transcripts is unknown . In this study , we show that many of these transcripts are due to splicing errors which are not preserved over evolutionary time . We estimate that the error rate in the splicing of an intron is about 0 . 7% and demonstrate that there are two major types of splicing error: errors in the recognition of exons and errors in the precise choice of splice site . These results raise the possibility that variation in levels of alternative splicing across species may in part be to variation in splicing error rate .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"genetics",
"and",
"genomics",
"evolutionary",
"biology/genomics"
] |
2010
|
Noisy Splicing Drives mRNA Isoform Diversity in Human Cells
|
Control of tsetse flies using insecticide-treated targets is often hampered by vegetation re-growth and encroachment which obscures a target and renders it less effective . Potentially this is of particular concern for the newly developed small targets ( 0 . 25 high × 0 . 5 m wide ) which show promise for cost-efficient control of Palpalis group tsetse flies . Consequently the performance of a small target was investigated for Glossina fuscipes fuscipes in Kenya , when the target was obscured following the placement of vegetation to simulate various degrees of natural bush encroachment . Catches decreased significantly only when the target was obscured by more than 80% . Even if a small target is underneath a very low overhanging bush ( 0 . 5 m above ground ) , the numbers of G . f . fuscipes decreased by only about 30% compared to a target in the open . We show that the efficiency of the small targets , even in small ( 1 m diameter ) clearings , is largely uncompromised by vegetation re-growth because G . f . fuscipes readily enter between and under vegetation . The essential characteristic is that there should be some openings between vegetation . This implies that for this important vector of HAT , and possibly other Palpalis group flies , a smaller initial clearance zone around targets can be made and longer interval between site maintenance visits is possible both of which will result in cost savings for large scale operations . We also investigated and discuss other site features e . g . large solid objects and position in relation to the water's edge in terms of the efficacy of the small targets .
The major vectors of Human African Trypanosomiasis ( HAT ) are in the Palpalis group tsetse flies , especially the G . fuscipes subspecies , which are responsible for transmission of >90% of reported HAT cases [1] , [2] . In the present situation with limited drug and no vaccine availability , vector control remains an important addition to current efforts against HAT . Tsetse control with insecticide-treated blue/black cloth panels ( c . 1–2 m wide ×1 m high ) , called targets [3] , have been used successfully for several Morsitans group tsetse fly species , but only to a limited extent for Palpalis group tsetse [4] . Control of Palpalis group flies is costly and requires high densities of 10–30+ targets to be deployed per km2 . In contrast , Morsitans group tsetse can be controlled with odour-baited targets at densities as low as 4per km2 [5] , [6] , [7] . It is clear from published studies that factors such as the vegetation , the coverage of the habitat achieved with deployed targets and the correct siting and maintenance of targets play a very important role in efficient control [8] . Targets or traps have to be deployed in sites which allow for the maximum number of tsetse flies available in the range of attraction to locate them . If an odour is used with the device for control of Morsitans group flies , this range is about 5–150 m plus , while an unscented target or trap has a range of about 5–30 m [9] . Limited artificial odours exist at present for Palpalis group flies [10] , [11] so the trap or target's efficacy relies heavily on its visibility . The accepted principle for identifying a suitable site for a trap or target for tsetse species is that the site has open access and visibility in most directions with no large bushes nearby and no low overhanging canopy . For example , optimal sites for the Morsitans group flies G . m . morsitans and G . pallidipes are open and well away from trees and bushes [8] . For G . austeni ( also a Morsitans group fly ) sites inside the shaded forest , but still ‘open’ due to a high tree canopy and little undergrowth , is best [12] . The optimal trapping sites reported for the Palpalis group fly , G . f . fuscipes , are open sites close to the water's edge [13] , or an open site outside the forest but not more than 5 m away from the forest edge [14] . Optimal sites for G . tachinoides and G . p . gambiensis are on the river's edge in direct sunshine [15] . In practice the best available site in the chosen control area , or the next best potential site , will be selected and improved by cutting back vegetation and clearing undergrowth to increase visibility of the target or trap . However , the majority of sites will also include some other features such as large tree trunks , thick bushes , large rocks etc . This immediate arrangement of vegetation and solid objects around the site , i . e . the site morphology , can significantly affect tsetse catches [8] . For example , if a leafy bush with overhanging canopy grows within 1 m of a target catches of G . m . morsitans and G . pallidipes decreased by 70–80% , while if encroaching vegetation reduced the site clearing to 2 m diameter and covered about 66% of the perimeter catches also decreased by 70% [8] . Despite the importance of the Palpalis group tsetse in disease transmission there is limited information available on the effects of site morphology on target or trap efficiency for these flies , apart from the general description of what is believed to be a good site mentioned above . Understanding the impact of site morphology , especially vegetation encroachment , is imperative following the newly developed cost-efficient small targets ( c . 0 . 125 m2 ) for control of five major HAT vectors namely , G . fuscipes fuscipes , G . f . quanzensis , G . f . martinii , G . palpalis gambiensis and G . tachinoides [16] , [17] , [18] . These small targets , as much as 8× smaller than the standard 1×1 m target and using 24× less material than the biconical trap , show great potential for economic savings in control of Palpalis group tsetse . However , the effectiveness of such small targets might be severely and rapidly compromised in the field if vegetation re-growth is as serious a problem as it is with Morsitans group flies as described above . Potentially this factor could rapidly negate the economic savings of using small targets . To address these concerns we have evaluated the performance of small targets for G . f . fuscipes in different scenarios of site morphology and vegetation encroachment as may be typically encountered in the tropical environment . The better understanding of the behaviour of G . f . fuscipes in relation to site features will contribute to effective and efficient deployment of control and monitoring devices in large scale control of G . f . fuscipes .
Studies were performed from May to December 2010 on two small islands ( each c . 0 . 5 km2 ) , called Big and Small Chamaunga ( 0° 25′ S , 34°13′ E ) , off Mbita point in Lake Victoria , Kenya . See [10] , [16] for detailed description . The standard sampling device was a 25×25 cm target made from blue cotton cloth with an adjacent flanking net ( 25×25 cm ) of fine black netting . Henceforth , the term ‘target’ refers to this combination of cloth and netting . Electrocuting grids fitted in a frame covered both the cloth and netting and killed flies on impact , which then fell into trays of water below the grids . See [16] for detailed description . Experiments ran for 12 days each during the peak activity time of G . f . fuscipes , from 09:00–12:00 hours . The standard experimental design was a series of Latin-squares of treatments x days x sites , with sites at least 50 m apart . Analysis of variance was performed after transforming the daily catches ( n ) to log ( n+1 ) . Only detransformed catches are discussed in the text , while the transformed standard errors of the difference ( SED ) are provided in Tables 1 and 2 . The term ‘significant’ denotes that means are different at the P<0 . 05 level of probability or less . We investigated the following four aspects of site morphology and scenarios for vegetation encroachment; diagrams of the arrangements of targets and surrounding vegetation and other objects are shown in Fig . 1 . All treatments were compared to a standard target without any surrounding bushes or other objects in a clearing c . 5 m in diameter . 1 . Vegetation encroachment from the sides , for example when a target site is not maintained and vegetation re-growth results in: a ) obstruction of the perimeter and b ) decreasing the diameter of the site's clearing . In situations such as these the visibility of the small target and access for tsetse to it and around it , becomes restricted . To simulate bushes , we fixed leafy branches to stick frameworks to form hedges ( Fig . 2 ) which we placed in various arrangements around the target ( Figs . 1A–C ) , as described below . Similar hedges used to simulate site effects for the Morsitans group tsetse G . m . morsitans and G . pallidipes showed that there was no significant difference in the responses of tsetse to artificial bushes and real ones [8] . The first experiments studied the effect of percentage obstruction of the perimeter of a target site . The target was either completely unobstructed ( 100% visibility , control treatment ) or ( A ) bushes ( 1 . 5 m long , 1 m from the target ) were placed on all four sides ( 0% visibility , ) or on two sides ( 50% visibility ) with the hedges being placed either ( B ) orthogonally or ( C ) in parallel to the long axis of the target . The next experiments looked at the effect of surrounding the targets with an incomplete ring of bushes as follows: 2 . Vegetation encroachment from above; e . g . when a target is deployed under a tree or shrub with overhanging branches . Metal poles of appropriate length were used to support a framework of green sticks with interwoven leafy branches which formed a canopy above the target ( Fig 3 ) . Canopies were 1 . 5×1 . 5 m in diameter and 2 m , 1 m or 0 . 5 m above ground level ( Fig . 1E , with overhead vegetation only ) . A subsequent experiment then investigated a combination of a canopy above a target and a bush next to it , for example when a large bush grew next to as well as over the target . The canopy was 1 m above the target and either ( A ) one or ( B ) hedges were placed orthogonally c . 0 . 75 from the target ( Fig 1E ) . 3 . Proximity to solid objects; e . g . large rocks which may obscure a target , or a thick tree trunk next to the target . Due to the great variety in size , colours , shapes and combinations of site morphology in nature , it is not possible to duplicate these exhaustively or change these features between sites . A partial simulation of large rocks could be achieved by placing drums horizontally on the ground , or vertically on top of each other to simulate these large objects ( Fig 1 , diagrams D&F ) . The drums were made of plastic ( 50 cm diameter × 80 cm high , volume = 160 L ) covered with matt black cotton cloth and placed either next to , or in front of a target . In addition , we also looked at the responses of G . f . fuscipes to a small target next to a real tree bole ( a paw-paw tree bole 30 cm diameter , 1 . 8 m high ) and whether the orientation of the target to the tree was of importance , i . e . with the blue cloth or the black netting panel closest to the bole ( Fig 1G ) . 4 . Catches of G . f . fuscipes at different distances from the water's edge . This was done because standard field procedure is to place the device close to the water's edge [19] , [20] partly to increase visibility , but also because casual field observations show flies apparently move along the water's edge . A standard small target was deployed in a randomized block design in four sites . The control site was the water's edge , with the other three sites at 2 m or 4 m inland or 2 m into the water . For the latter , the target and collection tray were fixed to a floating platform of sticks and the electric cables lengthened to reach the power supply on the shore . The same set-up was repeated but using a standard Biconical trap as collection device .
Following the vegetation encroachment experiments , we looked at the effect on catches of large solid objects next to a small target . As described in Material and Methods , we used drums as artificial rocks and tree trunks for this study , due to the difficulty in otherwise simulating these objects in the field . Our data showed that when either the ‘rock’ or ‘tree trunk’ was placed next to the target there was no significant difference compared to catches from the control target ( Table 1 , experiment 7 ) . In fact , the catches of female G . f . fuscipes increased in both cases , by 1 . 2× when the rock ( Treatment A ) was used ( 6 . 7 tsetse/day , s . e . d . = 0 . 1 ) and by 1 . 1× when the tree was used ( Treatment B , 6 . 1 tsetse/day ) . Tsetse flies are attracted to large black objects and black drums and flat black cloth panels are used routinely in experiments to increase visual attraction [8] , [16] . Therefore the observed increase in catches may be expected , but the more interesting question is what happens when such large black objects obscure the visibility of the small target , e . g . when a large rock is directly obscuring a small target . We found ( Experiment 8 ) that the unobscured target ( 9 . 9 flies/day , s . e . d . = 0 . 08 ) caught 80% ( P<0 . 001 ) more females than the target with one drum in front ( Treatment A , 2 . 0 flies/day ) and 98% more females than with a drum on each side ( Treatment B , 0 . 03 flies/day ) . Catches of male G . f . fuscipes showed no significant difference ( P = 0 . 2 ) between the target in the open and either of the treatments , although 20% less flies were caught with one drum in front of the target ( Treatment A , 3 . 2 flies/day ) and 80% less with a drum on each side of the target ( Treatment B , 0 . 2 flies/day ) , completely obscuring the frontal views . When placing a target next a real tree trunk ( Table 1 , exp . 9 ) there was a doubling in female catches with both the blue cloth closest to the trunk ( Treatment B , 2 . 2 flies/day ) and with the netting closest to trunk ( Treatment C , 2 . 8 flies/day ) although this was not significant ( P = 0 . 26 ) . Finally , we investigated the effect of the position of a small target and biconical trap in relation to the water's edge ( Table 2 , experiment 1 & 2 ) . For G . f . fuscipes , the trapping sites usually used in control campaigns are open and close to , or right on the water's edge [13] . Casual field observations indicate that flies may use the water's edge as a movement ‘corridor’ , perhaps due to more abundant green vegetation for shelter , higher humidity and higher chance of finding a host , particularly monitor lizards which inhabit these aquatic margins . However , our results show that a small target placed on the water's edge did not catch significantly more female flies than targets placed 2 m ( 7 . 2 flies/day , s . e . d . = 0 . 15 ) , 4 m inland ( 6 . 0 flies/day ) , or floating 2 m into the water ( 5 . 8 flies/day ) . When a biconical trap was used as collection device ( Table 2 , experiment 2 ) , female catches on the water's edge were slightly better ( 5 . 7 flies/day , s . e . d . = 0 . 2 ) than that at 2 m inland ( 2 . 5 flies/day ) , 4 m inland ( 4 . 8 flies/day ) and 2 m in the water ( 1 . 6 flies/day ) . However , the differences were not significant . Although deployment at the water's edge may be desirable , it appears not to be essential because target efficiency does not decrease significantly over a few meters at least . This is important as it means targets can be sited to minimise losses due to flooding .
Vegetation encroachment around a small target , from the sides and above , does not significantly affect its killing efficiency for G . f . fuscipes as long as there are some openings between adjacent bushes , wider than 30 cm . These results are intriguing because the rapid re-growth potential of the tropical vegetation in the habitat of G . f . fuscipes and other Palpalis group tsetse , combined with the small size of the targets , could make it seem improbable that these targets will remain effective . Indeed , our results show that only one such scenario , grass regrowth very close to the target , poses a serious threat to their performance . Our simulation of grass height corresponds roughly to between c . 15 ( 15 cm high ) and 60 days ( 60 cm high ) as observed in the rainy season in the field . As expected , the small diameter clearings ( 0 . 75 m ) created by the proximity of surrounding grass significantly decreased target catches . However , this represents severe and complete grass regrowth around a target , something which does not happen frequently in nature because grass rarely grows uniformly and there always remain some openings between clumps of grass to allow visibility and access to a target . In addition and as matter of routine , this scenario is easily prevented by the proper initial clearing of target sites . In some circumstances this can aided by the subsequent use of systemic herbicides such as glyphosate which can inhibit grass regrowth for several months afterwards . For example , application of glyphosate maintained reduced grass cover for up to 26 weeks on a rainforest edge [21] . Limited studies have been done on the effect of vegetation encroachment on the efficiency of a target or trap for Palpalis group tsetse species . The most relevant studies are from Morsitans group flies [8] where the effect of vegetation close to a trap dramatically and significantly reduced catches of G . m . morsitans and G . pallidipes . For example , one bush with an overhanging canopy next to a trap , decreased catches of both Morsitans group species by more than 80% , while a decrease in diameter of clearing size from 12 m to 2 m led to about 65% decrease in catches . In contrast , our data for G . f . fuscipes showed no significant difference between control and treatment catches in both scenarios , even with only 1 m diameter clearings . The presence of a few bushes surrounding a target site , not obscuring more than about 70% visibility , may in fact be slightly beneficial . An apparently similar situation was evident with G . m . morsitans and G . pallidipes , where trap catches increased if 2–6 bushes were within 2–12 m from the target [8] . However , the smallest clearing size used for the G . m . morsitans and G . pallidipes experiments was 2 m radius ( 4 m diameter ) , at which catches of both species were 65% less than the open trap . As the clearing diameter was increased to 12 m , the catches increased . For G . f . fuscipes a remarkably small clearing of even 1 m diameter remained effective . The importance of an opening about 50 cm wide between adjacent bushes around a target was evident as catches reduced significantly ( by 68% for females ) if this opening was 30 cm and less . This was also found for G . m . morsitans and G . pallidipes , with catches of both species increasing significantly when the opening size is widened from 25 cm to 50 cm and more [8] . We showed that G . f . fuscipes readily enters between and through leafy vegetation to locate a small target . This behaviour corresponds with the habitat along the islands and shore of Lake Victoria , where their main hosts are monitor lizards . G . f . fuscipes have to locate these medium to small-sized reptiles between the leafy vegetation and rocks . Other site features such as large rocks or tree boles close to the target also affect catches of G . f . fuscipes , e . g . a single large solid object to the side of a small target , whether this was an artificial rock or tree bole , or a natural tree bole , actually increased catches . On the other hand , if one or more such objects obscured the frontal view of the target , catches decrease significantly . In addition , it would seem that the waters edge is not a required trap or target site for G . f . fuscipes . This is important as changes in water height can easily sweep away control devices with much cost to control programmes . The priority should be given to visibility rather than proximity to waters edge ( at least within the 4 m investigated here ) , because target efficiency does not decrease significantly over just a few meters between the water's edge and inland . As illustrated in this work , the small targets retain their killing efficacy in several situations of vegetation encroachment , even in small clearings of 1 m diameter and with leafy bushes close-by and above . Nevertheless , in practice , we recommend that sites be cleared to at least 2–3 m in diameter during initial deployment and that overhanging or intruding vegetation be cut back . This will allow for maximum visibility of the target during the first months after deployment . Maintenance intervals will vary between locations depending on vegetation regrowth rates , but under conditions in the study area we expect the small targets to remain efficient for 3–6 months after initial deployment , with no maintenance visits required in-between . If possible the use of a systemic herbicide applied on the site will prevent the regrowth of grass and other vegetation . The possible herbicides available for use next to watercourses are very limited; for example glyphosate is the only product registered for such use in the U . K . The data presented here demonstrates the potential for less frequent maintenance visits to cut back and control vegetation , which is a major financial constraint in tsetse control operations [22] where targets have to be serviced regularly to maintain efficiency . Another reason for maintenance visits is to ensure that the target is still in its correct position , is upright , the cloth is in good condition and that the moving parts are free . When using large targets , this maintenance has to be carried out regularly and irrespective of whether the vegetation needs clearing . This will be largely unnecessary when using the small targets because they will be more stable and not blown over or bent by strong winds as frequently as large targets . Clearly , there is potential for low-cost , low-maintenance control of G . f . fuscipes , and there is a necessity of these types of studies on other Palpalis group tsetse species in other tropical environments , to allow for better understanding and control of these major vectors of HAT .
|
Sleeping Sickness ( Human African Trypanosomiasis ) is a serious threat to health and development in sub-Saharan Africa . Due to lack of vaccines and prophylactic drugs , vector control is the only method of disease prevention . Small ( 0 . 25×0 . 5 m ) insecticide-treated targets have been shown to be cost-efficient for several Palpalis group tsetse flies , but there are concerns that they may become obscured by vegetation with a subsequent reduction in efficiency . We showed that the efficiency of the small targets was largely uncompromised by vegetation encroachment because G . f . fuscipes readily enter between and under vegetation to locate a small target , e . g . into small ( 1 m diameter ) site clearings and underneath a very low ( 0 . 5 m ) canopy . This implies that the dense vegetation , typical of the riverine habitats of Palpalis group tsetse , will not compromise the performance of tiny targets , as long as there are adequate openings of >30 cm between vegetation . Moreover , the maintanence of cleared areas around targets seems less important for the control of G . f . fuscipes with consequent savings in costs for control operations .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"african",
"trypanosomiasis",
"parasitic",
"diseases"
] |
2011
|
Vegetation and the Importance of Insecticide-Treated Target Siting for Control of Glossina fuscipes fuscipes
|
On a global scale scabies is one of the most common dermatological conditions , imposing a considerable economic burden on individuals , communities and health systems . There is substantial epidemiological evidence that in tropical regions scabies is often causing pyoderma and subsequently serious illness due to invasion by opportunistic bacteria . The health burden due to complicated scabies causing cellulitis , bacteraemia and sepsis , heart and kidney diseases in resource-poor communities is extreme . Co-infections of group A streptococcus ( GAS ) and scabies mites is a common phenomenon in the tropics . Both pathogens produce multiple complement inhibitors to overcome the host innate defence . We investigated the relative role of classical ( CP ) , lectin ( LP ) and alternative pathways ( AP ) towards a pyodermic GAS isolate 88/30 in the presence of a scabies mite complement inhibitor , SMSB4 . Opsonophagocytosis assays in fresh blood showed baseline immunity towards GAS . The role of innate immunity was investigated by deposition of the first complement components of each pathway , specifically C1q , FB and MBL from normal human serum on GAS . C1q deposition was the highest followed by FB deposition while MBL deposition was undetectable , suggesting that CP and AP may be mainly activated by GAS . We confirmed this result using sera depleted of either C1q or FB , and serum deficient in MBL . Recombinant SMSB4 was produced and purified from Pichia pastoris . SMSB4 reduced the baseline immunity against GAS by decreasing the formation of CP- and AP-C3 convertases , subsequently affecting opsonisation and the release of anaphylatoxin . Our results indicate that the complement-inhibitory function of SMSB4 promotes the survival of GAS in vitro and inferably in the microenvironment of the mite-infested skin . Understanding the tripartite interactions between host , parasite and microbial pathogens at a molecular level may serve as a basis to develop improved intervention strategies targeting scabies and associated bacterial infections .
Streptococcus pyogenes or group A streptococcus ( GAS ) is a human specific pathogen , which can cause a wide variety of diseases that typically originate from localised infections of skin ( impetigo ) or throat ( pharyngitis ) . Multiplication and lateral spread of GAS invading the skin can result in erysipelas and cellulitis in the deep layers of the skin or in necrotising fasciitis . Disease progression from here can cause severe systemic infections such as streptococcal toxic shock syndrome ( STSS ) and life-threatening sepsis . Autoimmune-mediated complications , in particular , rheumatic heart disease ( RHD ) and post-streptococcal glomerulonephritis ( PSGN ) can develop after the initial infection has resolved . To date , GAS remains in the top ten global causes of mortality with at least approximately 500 , 000 deaths a year [1 , 2] . Scabies , caused by infection with Sarcoptes scabiei , is an important risk factor for impetigo resulting from GAS and Staphylococcus aureus infections [2–6] . Inhibition of innate defences including the complement system is a prerequisite for successful establishment of bacterial infections . GAS and S . aureus have evolved mechanisms to prevent activation of the complement cascades [7–16] . Recently we have shown that scabies mites may offer further congenial conditions for infections by these bacteria by flooding their immediate surroundings with a multitude of complement inhibitors [17–20] . In particular the scabies mite serpin B4 ( SMSB4 ) , a 54 kDa serine protease inhibitor , inhibits complement activation [20] and promotes the growth of GAS [19] and S . aureus [21] . SMSB4 is secreted into the mite digestive system , where it co-localises with ingested host complement factors [20] and it is excreted with the mite faeces into the epidermal mite burrows [20] . Bacteria , in particular cocci , have been found in great abundance in the epidermal mite burrows [22] . In multiple clinical reports the colonisation of mite-infected skin with GAS [23] , S . aureus [22 , 24] , and other pathogens [25 , 26] has been thought to be the main cause of systemic infection and detrimental disease outcomes for patients with severe scabies . The complement system , an immediate host defence against invading pathogens , consists of more than 30 soluble plasma proteins that constitute a series of enzymatic cascades [27] . Complement can be activated via three different pathways , namely classical pathway ( CP ) , lectin pathway ( LP ) and alternative pathway ( AP ) . The CP is antibody-dependent and initiated by binding of C1q , a pattern recognition molecule ( PRM ) to the bacterial bound immune complexes such as IgG , natural IgM or direct binding to surface microbial sugars [28–30] . The LP is initiated when microbial surface sugars are recognised by the PRMs , mannose binding lectin ( MBL ) or M- , L- and H-ficolins . These two pathways form the enzyme complex CP/LP-C3 convertase ( C4b2a ) [31–33] . In the AP , C3 naturally breaks down to C3H2O at a low level to which factor B ( FB ) binds , and this assembly is cleaved by factor D , forming an AP-C3 convertase ( C3bBb ) [34] . This enzyme complex generally requires stabilisation by properdin [35 , 36] . The C3 convertase is the key enzyme resulting from the complement activation , and it cleaves C3 to release an important opsonin , C3b . Deposition of C3b on the microbial surface is crucial as it marks the microbes for an efficient uptake and subsequent killing by phagocytes . Furthermore , at a high local concentration C3b binds to C3 convertase , thereby turning into C5 convertase ( C4b2a3b/C3bBb3b ) . C5 convertase cleaves C5 into C5a and C5b . C5a is a potent chemoattractant , which recruits neutrophils , monocytes and macrophages to the site of infection . C5b and other complement components ( C6 , C7 , C8 and C9 ) form the membrane attack complex ( MAC/C5b-9 ) on the cell surface , causing direct cell lysis in sensitive cells , such as gram-negative bacteria [37 , 38] . To date , studies on interactions between complement and GAS were only focused on the CP and AP [39–42] . Here we investigate the role of all three complement pathways innately controlling establishment of GAS infection . We found that CP plays a major role followed by AP , while the role of MBL-dependent LP was insignificant . Furthermore , we analysed the role of the scabies mite complement inhibitor SMSB4 in the survival of GAS in fresh blood to better understand the mechanisms underlying the link between GAS and scabies when co-infecting the human host . Our data showed that SMSB4 promoted the growth of GAS in blood by inhibiting the activation of the CP and the AP , which presumably caused the reduction of opsonisation and anaphylatoxin release . This is the first study analysing molecular interactions that may govern the initial events of overcoming human complement defence during co-infection of the skin by scabies mites and GAS .
Normal human serum ( NHS ) for complement activation assays and fresh blood samples for bactericidal assays were prepared from blood donated by healthy volunteers . Informed written consent was obtained from all blood donors . Blood from one donor was used in all further assays requiring fresh whole blood . The protocols for sourcing blood for complement assays were approved by the Human Research Ethics Committee of the QIMR Berghofer Medical Research Institute ( P443 ) . Ten ml of venous blood collected into a Vacutainer ( Becton Dickinson ) was obtained from at least 7 healthy volunteers . Tubes containing the blood samples were allowed to clot at room temperature ( RT ) for 30 min . Samples were centrifuged at 2000 ×g for 10 min at 4°C and the clotted blood was removed . Samples were centrifuged again at 2000 ×g for 10 min at 4°C . Sera were pooled , aliquoted into 500 μl volumes and stored at -80°C until use . Depleted sera ( C1q- and FB- ) were purchased from Quidel ( San Deigo , USA ) . These sera were prepared from pooled human sera from healthy donors , which were specifically depleted of either C1q or FB . MBL deficient serum ( MBLd ) was purchased from the Statens Serum Institut ( Copenhagen , Denmark ) . It was prepared from pooled sera from blood collected from otherwise healthy donors with the MBL genotype B/B . IgG was depleted from NHS using Albumin and IgG depletion SpinTrap columns prepacked with Protein G Sepharose ( GE Healthcare ) , following the manufacturer’s instructions . GAS isolates were obtained from the culture collection from the scabies and bacterial pathogenesis laboratory at QIMR Berghofer MRI . Strains used here were GAS 88/30 ( emm 97 ) [43 , 44] , PRS30 ( emm 83 ) [45] , both emm-cluster D , PRS8 ( emm 12 ) [45] , 5448 ( emm 1 ) [46] , both emm-cluster A-C , PRS55 ( emm 9 ) , PRS15 ( emm 48 ) , both emm-cluster E [45] . All strains were cultured at 37°C and 5% CO2 either on Columbia Blood Agar supplemented with 0 . 1% CaCO3 ( w/v ) and 4% defibrinated horse Blood ( Equicell products , Australia ) ( CBAC ) or in Tryptic Soy Broth ( Thermo Fisher Scientific Pty . Ltd . , Australia ) ( TSB ) . GAS cell suspensions were prepared from mid-log growth phase cultures ( OD600 = 0 . 35 ) . Cells were harvested by centrifugation ( 4000 ×g , 10 min , 4°C ) , washed twice in phosphate buffered saline ( PBS ) and re-suspended to a final OD600 = 0 . 03 in the same buffer . This cell suspension corresponds to approximately 1x 105 colony forming units ( cfu ) /ml . Bacteria were enumerated by plate count of cfu/ml on CBAC agar at 37°C and 5% CO2 overnight . DNA encoding SMSB4 was cloned and expressed in Escherichia coli BL21 ( Qiagen ) , purified under denaturing condition and refolded into active serpin as described previously [21] . Briefly , SMSB4 cDNA ( Yv5004A04 , GenBank accession no . JF317222 ) of the human scabies mite S . scabiei cloned into the pQE9 expression vector ( Qiagen ) was transformed into E . coli BL21 . E . coli cells were cultivated overnight at 37°C in Luria broth ( Becton Dickinson ) containing 100 μg/mL ampicillin . After inoculation in 2YT medium ( Becton Dickinson ) containing 100 μg/mL ampicillin , the cells were grown at 37°C , shaking at 200 rpm until an OD600 of 0 . 6–0 . 7 was reached . Expression of recombinant SMSB4 was induced by addition of 0 . 5 mM IPTG and continued shaking at 200 rpm for a further 4 h . Cells were collected by centrifugation at 6000 ×g at 4°C for 20 min , re-suspended in serpin buffer ( 50 mM Tris , pH 8 . 0 , 100 mM NaCl , 10 mM EDTA , 1 mM PMSF ) and lysed in 250 μg/ml lysozyme and 10 μg/ml DNase at room temperature ( RT ) under continuous rotation for 1 h . All of the following purification steps were performed at 4°C . After sonication of the spheroplasts by a Sonifier 250 ( Branson ) , inclusion bodies were washed five times using serpinX buffer ( 50 mM Tris , pH 8 . 0 , 100 mM NaCl , 10 mM EDTA , 0 . 5% ( v/v ) Triton X-100 ) and retrieved by centrifugation ( 16 , 000 ×g for 20 min at 4°C ) . The resulting pellet was dissolved in solubilisation buffer ( 6 M guanidine hydrochloride , 50 mM Tris , pH 7 . 8 , 1 mM DTT ) for 1 h . Proteins were further purified by nickel affinity chromatography . Solubilised protein was diluted 1:1 with bind buffer ( 6 M urea , 100 mM NaH2PO4 , 10 mM Tris , pH 8 . 0 , 5 mM imidazole , 150 mM NaCl , 1% ( v/v ) glycerol , 1 mM DTT ) and bound overnight to a pre-equilibrated 1 ml Ni-NTA matrix ( Qiagen ) in a PolyPrep column ( BioRad ) on a rotating shaker . The column was washed twice with 5 ml of wash buffer ( 6 M urea , 100 mM NaH2PO4 , 10 mM Tris , pH 6 . 3 , 5 mM imidazole , 150 mM NaCl , 1% ( v/v ) glycerol , 1 mM DTT ) . Bound proteins were eluted twice using 3 ml of elution buffer ( 6 M urea , 100 mM NaH2PO4 , 10 mM Tris , pH 8 . 0 , 250 mM imidazole , 150 mM NaCl , 1% ( v/v ) glycerol and 1 mM DTT ) . Purified recombinant proteins were refolded overnight by drop wise addition of the protein elution into refolding buffer ( 300 mM L-arginine , 50 mM Tris , 50 mM NaCl and 5 mM DTT , pH 10 . 5 ) using a Minipuls 3 pump ( Gilson ) at a flow rate of 20 μl/min under gentle stirring . Refolded proteins were concentrated using an Ultrasette Lab Tangential Flow Device ( 10 kDa MWCO , PALL Life Sciences ) , followed by further concentration in centrifugal filters ( 10 kDa MWCO , Amicon Ultra , Millipore ) . Protein concentrations were determined by Bradford protein assay ( Bio-Rad ) with bovine serum albumin ( BSA ) ( Invitrogen ) as a standard according to the manufacturer’s instructions . Molecular mass and purity were confirmed using SDS-PAGE analysis with Coomassie blue R-250 staining . For all assays , SMSB4 was buffer exchanged into the corresponding assay buffers using 0 . 5 ml centrifugal filters ( 10 kDa MWCO , Amicon Ultra , Millipore ) . Bactericidal assays were performed with fresh human blood collected in standard vacutainers containing hirudin as anticoagulant at a concentration of 25 μg/ml ( Dynabyte Informationssysteme GmbH , Munich , Germany ) . Hirudin ( lepirudin ) generally preserves the complement reactivity , making it the most suited anticoagulant for complement in vitro studies [47 , 48] . The assays were performed as described previously [21] with minor modifications . Bacteria were grown overnight at 37°C and 5% CO2 in 5 ml TSB . The overnight culture was diluted to an initial OD600 of 0 . 05 in a fresh aliquot of 5 ml TSB and the GAS culture was grown to mid-log growth phase ( OD600 0 . 35 ) at 37°C and 5% CO2 . This culture was diluted in PBS to obtain an approximately 1×103 cfu/ml challenge dose . To 100 μl of human venous blood , either of the following compounds were added in a volume of 27 . 5 μl: purified recombinant SMSB4 in the experimental samples , BSA or GVB2+ buffer ( 5 mM veronal buffer , 140 mM NaCl , 0 . 1% ( w/v ) gelatin , 1 mM MgCl2 , 0 . 15 mM CaCl2 , pH 7 . 35 ) in the negative controls . Finally 12 . 5 μl of the GAS suspension were mixed into a total volume of 140 μl . Samples were placed on a rotisserie and incubated with end over end mixing for 3 h at 37°C . Subsequently 50 μl aliquots from each appropriately diluted tube were plated in duplicate on CBAC agar plates . The plates were incubated overnight at 37°C and 5% CO2 and bacterial numbers were enumerated as cfu/ml . Bacterial recovery was calculated as a percentage of the number of bacteria recovered from samples treated with various test compounds in reference to the GAS challenge dose in PBS without addition of blood . To coat a 96-well assay plate ( Maxisorp Immuno Plate , Nunc , Denmark ) with GAS cells , 100 μl of approximately 1×105 cfu/ml of GAS cell suspension was added to the wells , incubated first at 37°C for 1 h and subsequently kept at 4°C overnight . Wells were washed 4 times with 200 μl PBS and 0 . 05% Tween-20 in between each step of the assay . The cells were incubated with blocking buffer ( 4% BSA in PBS and 0 . 05% Tween-20 ) for 2 h at RT . Meanwhile , aliquots of 35 μl of 10% pooled human serum diluted in GVB2+ buffer were incubated with 35 μl of SMSB4 or BSA of varying concentrations at 37°C , 200 rpm for 1 h in a V-shaped bottom 96-well plate ( Nunc ) . Sixty μl of these mixtures were then transferred to the wells of GAS coated plate , which was further incubated at 37°C for 1 h . Bound complement proteins were detected by incubation with 60 μl of primary antibodies against human complement factors for 1 h at RT . For immunodetection , antibodies against C1q , C3d , C4c ( Dako , Denmark ) , properdin ( R&D systems ) , sC5b-9 neoantigen-specific antibody recognising the MAC complex ( Complement Technology Inc . , USA ) , IgG ( Sigma ) were used at a dilution of 1:4000 and antibodies against FB ( Complement Technology Inc . , USA ) , MBL , Ficolin H ( R&D system ) , Ficolin M and L ( Thermo Scientific ) were used at dilution of 1:1000 . The wells were subsequently incubated with 60 μl of horseradish peroxidase ( HRP ) -conjugated goat anti-rabbit , HRP-conjugated rabbit anti-goat , HRP-conjugated goat anti-mouse secondary antibodies ( Dako , Denmark ) at dilutions of 1:1000–1:4000 in blocking buffer at RT for 30 min to 1 h , depending on the primary antibody specificity and signal intensity . Sixty μl of OPD reagent ( Dako , Denmark ) containing 0 . 01% hydrogen peroxide was added to each well and incubated at RT until the ‘serum only’ positive control turned yellow . Reactions were stopped by addition of 50 μl of 0 . 5 N H2SO4 and absorbances were measured at OD490 with a POLARstar Optima fluorescent microtiter plate reader ( BMG Labtech , Melbourne , Australia ) . Statistical significance was determined using one way or two way ANOVA , with Tukey’s , Dunnett’s or Sidak’s multiple comparisons tests ( GraphPad Prism software , version 6 . 0; GraphPad Software Inc . USA ) . Values of p<0 . 05 were considered significant .
GAS strains belong to one of three clusters based on the arrangement of genes for cell surface M- or M-like proteins ( emm clusters ) . Members of these clusters exhibit tissue tropisms; cluster A-C GAS strains are generally found in the throat while cluster D is skin tropic and cluster E is referred as generalists with no specific tropism [43] . For this study we needed to define the baseline effect of complement on GAS strains belonging to these clusters . Hence , we selected six clinical strains ( two from each cluster ) and assessed their ability to survive in fresh human blood containing active complement and phagocytes . GAS strains 88/30 [43 , 49] and PRS30 [45] belong to emm cluster D and are predominantly recovered from skin . PRS8 [45] and 5448 [46] are of emm cluster A-C and exhibit throat tropism . PRS55 and PRS15 [45] belong to emm cluster E and are generalists ( either skin and/or throat tropism ) . All these strains survived in blood during the 3 h incubation period showing between 1 . 5- and 5- fold growth relative to the control , which had PBS in place of blood ( Fig 1A ) . These results suggest that irrespective of the differences in preferential tissue tropism , all strains had a similar level of resistance to whole blood . Furthermore , as the strains survived in the blood , the blood samples had little , if any , opsonic antibodies . Nonetheless , we did not find a growth increase of >32 fold , which is expected in these assays with a 3h incubation in blood in the absence of type-specific antibodies [50–53] . Hence , there seemed to be a growth-attenuation in blood , presumably due to the presence of generalised IgGs and active complement . To further characterise this baseline immunity in subsequent experiments , we used the skin-tropic strain 88/30 , which showed the highest survival in blood ( ~5 fold growth ) . We investigated the deposition of complement components on the GAS surface in the initial activation step of the complement cascades , specifically C1q for CP , FB for AP and MBL for LP . The rationale for this experiment was to determine the contribution of each complement pathway by assessing which components are deposited on the surface of GAS . Ninety six-well Maxisorp plates coated with GAS cells were exposed to an increasing concentration of NHS , and deposition of each complement component on the bacteria cell surface was detected by ELISA using complement factor -specific antibodies . While the amount of both C1q and FB deposited on the GAS cell surface was proportional to the concentrations of serum used , MBL did not appear to bind to the GAS cell surface at any of the serum concentrations tested ( Fig 1B ) . This result suggested that CP and AP , but not MBL-dependent LP are activated by GAS . Since the LP could be triggered by other lectin pathway PRMs such as M- , L- and H-ficolins , we investigated the deposition of these molecules onto the cell surface of 88/30 . None of the ficolins deposited ( supplementary data , S1 Fig ) , indicating that the LP may not be important for controlling GAS . To confirm this result by an independent experiment , we investigated deposition of the opsonin C3b on the bacterial surface upon activation of the complement pathway with GAS using three commercially sourced human sera , C1 depleted ( C1- ) , FB depleted ( FB- ) and MBL deficient ( MBLd ) . We expected that the absence of either classical or alternative pathway activation would result in lowered C3b deposition . Accordingly , we found that C3b deposition decreased by 40% in C1- and 20% in FB- sera relative to the NHS control ( Fig 1C ) . This data suggested that both CP , and to a lesser extent AP are responsible for the deposition of C3b on the surface of GAS in this assay . We also observed a 60% reduction in C3b deposition with the MBLd serum ( Fig 1C ) . To address this result obtained from the MBLd serum , we compared all three sera in a C1q deposition assay ( Fig 2A ) . As expected , the C1q- and FB- sera showed respectively little and normal levels of C1q deposition . By contrast , the MBLd serum showed 50–75% more C1q deposition compared to NHS ( p<0 . 0001 ) . From these results we inferred that the baseline level anti-GAS antibodies in MBLd serum may be low . Indeed addition of anti-GAS antibodies to the assay rescued C3b deposition by the MBLd serum ( Fig 2B ) . Thus the reduction observed in the C3b deposition in MBLd serum ( Fig 1C ) was due to the absence of anti-GAS antibodies required for the activation of CP in this serum . To confirm that the baseline immunity to GAS was due to the presence of general IgG , we depleted IgG from NHS by protein G sepharose column based affinity chromatography . The IgG depleted serum ( 99% ) reduced the deposition of C1q ( 99% ) , C4b ( a part of C4b2a , i . e . the CP-C3 convertase ) ( 86% ) and C3b ( 78% ) on GAS 88/30 ( Fig 2C ) . Reduction in the deposition of these complement components were directly attributable to the absence of antibody and thus related to the classical pathway . These results concur with the phagocytic killing assays described above ( Fig 1A ) . Taken together , the data supports the conclusion that CP and AP are the main complement pathways in controlling the growth of GAS and that the NHS used had a baseline immunity to GAS , impedimental to the establishment of infection . Since we earlier showed that scabies mite SMSB4 is a complement inhibitor and promoted growth of GAS [19] , we sought to ascertain whether this protein is able to impact on the growth attenuation of the six diverse strains owing to baseline immunity . We treated blood with 2 μM of recombinant SMSB4 prior to the addition of GAS . This resulted in a significant increase in the numbers of cfu of all strains tested ( 7–11 fold rise ) compared to that of the challenge dose ( Fig 3A ) . We found that SMSB4 promoted growth of 88/30 in a concentration dependent manner ( Fig 3B ) , suggesting that complement could play a significant role in delaying the onset of GAS skin infection . To understand SMSB4 mediated inhibition of opsonophagocytosis in blood we investigated the effect of this protein on the formation of CP-C3 convertase ( C4b2a ) , which was measured by deposition of C4b on the 88/30 cells . We demonstrated that SMSB4 caused near complete inhibition of C4b depositions on GAS surface at 0 . 4 μM concentration ( Fig 4A ) . Likewise , the C3 convertase specifically formed via AP was assayed by probing for the depositions of FB ( C3bBb ) and properdin ( a stabiliser of the AP-C3 convertase ) . SMSB4 at 0 . 4 μM concentration caused approximately 25% reduction of FB deposition and near complete reduction of properdin deposition on the GAS surface ( Fig 4B and 4C ) . These results suggest that SMSB4 greatly decreased the formation of CP-C3 convertase , and moderately decreased the formation of AP-C3 convertase . Furthermore we tested the effect of SMSB4 on the integral parts of the activated complement network , namely C3b deposition and release of anaphylatoxin C5a . Approximately 85% decrease in C3b deposition was observed when NHS was treated with 0 . 3 μM SMSB4 ( Fig 5A ) . SMSB4 reduced the deposition of C5b-9 complex by almost 40% when a concentration range of 0 . 1–0 . 4 μM was tested ( Fig 5B ) . This was an indirect indication that a proportional reduction in the release of anaphylatoxin C5a occurred in the presence of the mite complement inhibitor SMSB4 , as previously observed in a related experiment [54] . Taken together these results illustrated that SMSB4 offers an advantage to GAS by preventing activation of the complement cascade .
The majority of studies regarding host-pathogen interactions of GAS have focused on GAS virulence . Research addressing the relative importance of the host complement response to GAS infection has lagged behind , mainly due to a lack of suitable systems to study such complex interactions . Despite this , it has been established based on a limited number of publications , that CP and AP are major players in controlling GAS infections [39 , 41 , 42 , 55 , 56] . However , to our knowledge , a potential role of the LP during GAS infections has not been investigated . There are multiple animal models for analysis of S . pyogenes pathogenesis [57] . A mouse model genetically engineered to be deficient in specific complement components has been utilised to address the host complement response towards GAS infections [42] . Such models are useful because immune naive serum can be generated from mice and complex interactions between the three complement pathways can be dissected out . However , GAS is not a natural pathogen of mice and considerable differences exist between the murine and human complement systems [16 , 58] . Human complement is commonly investigated using human sera , either from donors with normal complement from which specific complement components have been artificially removed , or from patients with a natural deficiency in a specific complement component , e . g . MBL . However , it can be difficult to directly compare results between these commercial sera because the precise composition of individual complement samples cannot be standardised . Our data is in agreement with previous reports that GAS activates complement via CP and AP [39 , 41 , 42 , 55 , 56] . The lack of depositions of MBL and other PRMs of LP on GAS from this study is also in agreement with a previous report by Nordenfelt et al . that GAS in human serum were coated with complement proteins of the CP and AP , but not LP [59] . However , the question of whether CP or AP is more important for the activation of complement by GAS remains enigmatic . Initial reports in 1979 using human sera stated that the AP is the primary complement pathway activated in the absence of type-specific IgG [39 , 60] . Later work by Carlsson et al . in 2003 and 2005 reported that the CP is the main complement pathway in human serum when activated by an M-protein deficient GAS strain [41 , 56] . The level of CP-dependent opsonisation may thus depend on the absence or presence of M-protein , as most M-proteins recruit host C4BP to degrade CP-C3 convertase while some M-proteins recruit host factor H ( FH ) , which degrades AP-C3 convertase . More recently in 2006 , Yuste et al . investigated the host complement response towards four clinical isolates of GAS in mice , genetically engineered to lack either C1q or FB [42] . They compared mice sera with commercial human sera , which were depleted in either C1q or FB . The authors reported that the AP was the main complement pathway activated , mainly because opsonisation and mice survival was most reduced in FB- mice when infected with GAS . It is apparent that there were differences in the complement responses between mice and human sera , and that further differences may be GAS strain-related . In the particular context of GAS and scabies mite co-infection and the host complement response , the IgG-dependent CP appears to be the predominant complement pathway , in conjunction with a lesser effect of the AP . The GAS strain 88/30 is a skin isolate from a scabies patient ( as recorded in the culture collection at the Menzies School of Health Research , Darwin ) . It carries emm97 , M protein , which is currently M protein non-serotypable [49] . It is not known whether surface M proteins of 88/30 recruit host regulators such as C4BP and FH [40 , 61] , which affect the type of complement response [16 , 42 , 55 , 56 , 62] . CP-dependent opsonisation would be more important if the GAS surface binds FH , whereas the AP effect may be more pronounced for a strain that recruits C4BP . Overcoming host innate immunity by GAS is a prerequisite for successful infection . This fine balance between the host and the pathogen may often be influenced by a co-infecting pathogen . In this regard we reported earlier that scabies mites aid GAS infection , presumably by secreting into their immediate surroundings proteins that inhibit complement function [19] . The mechanism or the complement pathways involved were not known . In this study , we described that both CP and AP play a role in the initial immunity against GAS . Furthermore we demonstrate that the scabies mite protein SMSB4 inhibits the formation of C3 convertase mainly via the CP , as indicated by its potent inhibition on C4b deposition . This leads to a reduced opsonisation , which has several downstream effects: a moderate reduction in the AP-C3 convertase , as C3b is a component of this C3 convertase , and a reduced amount of CP- and AP-C3 convertases in combination with decreased C3b deposition , which reduces the amount of C5 convertase formed , and hence impacts on the release of anaphylatoxin C5a . In summary , the scabies mite protein SMSB4 inhibits ( i ) the formation of C3 convertases via CP and AP , ( ii ) the deposition of C3b and ( iii ) the formation of anaphylatoxin C5a . C3b deposition is crucial for the eradication of microbes as it marks the microbes for an efficient uptake and subsequent killing by phagocytes . C5a is a potent chemoattractant , which recruits neutrophils , monocytes and macrophages to the site of infection . We have recently shown that SMSB4 interferes with phagocytosis of S . aureus by neutrophils [21] and this may also apply to the uptake of GAS . In this light the protection mediated by a recent and highly promising combinatorial synthetic peptide vaccine strategy against GAS [44] may be compromised in scabies patients , as this vaccine critically depends on the presence and action of neutrophils . Complement is a superabundant system , involving a large number of components in a complex network and among individuals variations in the composition are often observed . Pathogens susceptible to complement have accordingly evolved multiple evasion strategies in order to ensure an anti-complement milieu in their immediate environment . Many pathogens have numerous different complement inhibitors acting on different targets within the complement system . Staphylococci , for example , have evolved an arsenal of molecules to counteract the complement system encoded by about 2% of their total genome [63] . Redundancy is very common and seems to be needed . In this light , it is a captivating thought that pathogens may ‘join forces’ against the onslaught of the host complement defense . Other studies have proposed similar hypotheses , so for example for cysteine proteinases from Porphyromonas gingivalis which have been shown to provide an advantage to other periodontal pathogens residing in the same location [64] . Scabies mites have evolved an astonishing repertoire of complement inhibitors , comprised of at least two classes—Scabies Mite Serpins ( SMSs ) [20] and Scabies Mite Inactivated Serine Proteases ( SMIPP-Ss ) [65] . These are represented in the genome as multi-copy families . Thirty-three SMIPP-Ss and six serpins have been identified and it is possible that the ongoing mite genome project will identify more . The apparent range of specific mechanisms preventing complement function indicates that mites inhibit the complement system at many points and additive effects of mite complement inhibitors have been demonstrated [20] . Notably , complement factors are ingested by mites [18 , 66] but MAC formation is not detected in the gut [66] , suggesting that the anti-complement defence system generated by the mite may be indeed very efficient in vivo . The mite serpin SMSB4 is only one of many mite complement inhibitors that scabies mites release simultaneously into the epidermis . We propose that as a whole , mite complement inhibitors accumulate to high anti-complement activities in the confined space of the gut and epidermal burrows , allowing the parasites and associated bacteria to evade the adverse effects of complement activation . This promotes the growth of bacteria present in the confined microenvironment of the epidermal burrows . We do not think that the circulating blood of scabies patients will contain physiologically relevant amounts of SMSB4 or that mite complement inhibitors have a systemic effect in the dermis or in the body . A clear limitation of the in vitro work presented here is that artificial assay systems were utilised and these test conditions mimic the in vivo conditions only partially . To provide a clear and detailed picture we tested one single complement inhibitor on its own , which again does not reflect the in vivo situation but is the only way to dissect the particular function of this protein . Finally , SMSB4 was produced as a recombinant protein in E . coli and underwent a lengthy refolding procedure . Hence only a portion of serpin molecules in the preparation was likely biologically active , which may explain why the relatively high concentrations of SMSB4s in a micromolar range were needed to observe a significant complement inhibition in some of the assays presented . Future research should aim to investigate the synergism between scabies mites and pathogenic bacteria in complement inhibition in an in vivo setting; potentially suitable animal models have been established [67] . It is intriguing to consider that the collective complement-inhibitory function of multiple mite excretory proteins in combination with complement inhibitors produced by GAS and other bacteria present [14] promotes the survival of bacterial pathogens in the microenvironment of the epidermal burrows produced by the mites . This molecular link between complement inhibition by mite proteins and bacterial survival is a novel aspect of pyoderma pathogenesis that may have important implications for the development of alternative therapies . Improving the treatment and management of scabies requires foremost a better understanding of the interactions between scabies mites , the bacteria subsequently infecting the scabies lesions and the host immune system .
|
The molecular mechanisms that underpin the link between scabies and bacterial pathogens were unknown . We proposed that scabies mites play a role in the establishment , proliferation and transmission of opportunistic pathogens . We investigated here the synergy between mites and one of the most recognised mite associated pathogens , Streptococcus pyogenes . As part of the innate immune response mammals have a pre-programmed ability to recognise and immediately act against substances derived from fungal and bacterial microorganisms . This is mediated through a sequential biochemical cascade involving over 30 different proteins ( complement system ) which as a result of signal amplification triggers a rapid killing response . The complement cascade produces peptides that attract immune cells , increases vascular permeability , coats ( opsonises ) the surfaces of a pathogen , marking it for destruction , and directly disrupts foreign plasma membranes . To prevent complement mediated damage of their gut cells , scabies mites secrete several classes of complement inhibiting proteins into the mite gut and excrete them into the epidermal mite burrows . Furthermore , these inhibitors also provide protection for S . pyogenes . We verified here specifically the impact of the mite complement inhibitor SMSB4 , to identify the molecular mechanisms behind the long recognised tendency of S . pyogenes to infect mite-induced skin lesions .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"complement",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"body",
"fluids",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"tropical",
"diseases",
"microbiology",
"parasitic",
"diseases",
"animals",
"ectoparasitic",
"infections",
"sexually",
"transmitted",
"diseases",
"neglected",
"tropical",
"diseases",
"immunologic",
"techniques",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"complement",
"inhibitors",
"proteins",
"medical",
"microbiology",
"scabies",
"immunoassays",
"microbial",
"pathogens",
"mites",
"immune",
"system",
"arthropoda",
"complement",
"activation",
"biochemistry",
"blood",
"anatomy",
"physiology",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2017
|
Complement inhibition by Sarcoptes scabiei protects Streptococcus pyogenes - An in vitro study to unravel the molecular mechanisms behind the poorly understood predilection of S. pyogenes to infect mite-induced skin lesions
|
Viral infection is a stimulus for apoptosis , and in order to sustain viral replication many viruses are known to carry genes encoding apoptosis inhibitors . F1L , encoded by the orthopoxvirus modified vaccinia virus Ankara ( MVA ) has a Bcl-2-like structure . An MVA mutant lacking F1L ( MVAΔF1L ) induces apoptosis , indicating that MVA infection activates and F1L functions to inhibit the apoptotic pathway . In this study we investigated the events leading to apoptosis upon infection by MVAΔF1L . Apoptosis largely proceeded through the pro-apoptotic Bcl-2 family protein Bak with some contribution from Bax . Of the family of pro-apoptotic BH3-only proteins , only the loss of Noxa provided substantial protection , while the loss of Bim had a minor effect . In mice , MVA preferentially infected macrophages and DCs in vivo . In both cell types wt MVA induced apoptosis albeit more weakly than MVAΔF1L . The loss of Noxa had a significant protective effect in macrophages , DC and primary lymphocytes , and the combined loss of Bim and Noxa provided strong protection . Noxa protein was induced during infection , and the induction of Noxa protein and apoptosis induction required transcription factor IRF3 and type I interferon signalling . We further observed that helicases RIG-I and MDA5 and their signalling adapter MAVS contribute to Noxa induction and apoptosis in response to MVA infection . RNA isolated from MVA-infected cells induced Noxa expression and apoptosis when transfected in the absence of viral infection . We thus here describe a pathway leading from the detection of viral RNA during MVA infection by the cytosolic helicase-pathway , to the up-regulation of Noxa and apoptosis via IRF3 and type I IFN signalling .
Cell death by apoptosis can protect a multicellular organism against viral infection: if the first infected cell dies in time the virus will not have the chance of producing new viral particles . Many viruses counter this cellular response by carrying genes coding for inhibitors of the host cell's apoptosis machinery . This interrelationship has first been shown clearly in a baculovirus where the loss of the caspase-inhibitor p35 leads to cell death and reduction in viral productivity [1] . Numerous viral anti-apoptotic genes have by now been found , whose products often have structural resemblance to mammalian anti-apoptotic proteins [2] , [3] . A number of poxviruses carry , among other anti-apoptotic genes , genes coding for proteins that resemble the inhibitory members of the mammalian Bcl-2-family of proteins . These genes have very little ( if any ) primary sequence homology with the mammalian proteins but crystallization studies have shown that they adopt a virtually identical structure . One critical feature of anti-apoptotic function of Bcl-2-proteins is the existence of a hydrophobic groove that accommodates the BH3-domain of pro-apoptotic Bcl-2-family members and thereby inhibits apoptosis [4] . Intriguingly , there is considerable variation at this critical structural position in poxviral Bcl-2-related proteins . The M11L protein from myxoma virus can bind BH3-domains with high affinity and cell biological studies indicate that the binding to pro-apoptotic Bax and Bak ( via their BH3-domains ) is necessary for the anti-apoptotic function of M11L [5] . Other poxviral proteins with a Bcl-2-like fold ( as determined in the crystal structure ) lack the structural hydrophobic groove and hence the ability to bind BH3-domains , despite having the other structural features of Bcl-2 [6] . These proteins ( and other , related poxvirus-encoded proteins ) seem to have functions in manipulating the host immune response rather than in apoptosis prevention [7] . How these structures and diverging functions have evolved is intriguing but unclear . The F1L protein has been found to inhibit apoptosis in vaccinia virus ( VACV ) as well as in the highly attenuated VACV strain modified vaccinia virus Ankara ( MVA ) [8] , [9] . The crystal structure of MVA F1L surprisingly showed that F1L protein dimerizes through a helical domain swap and that it , although it can bind some BH3-domain-peptides , does so with low affinity . Even the binding to the BH3-domain of Bim , the pairing of F1L with the highest affinity among all BH3-domains , occurs at a much lower affinity than the binding of any BH3-domain to mammalian Bcl-2 proteins thought to be functionally relevant [10] , [11] . F1L protein has further been identified as an inhibitor of caspase-9 [12] . Mammalian anti-apoptotic Bcl-2-like proteins ( Bcl-2 , Bcl-XL , Bcl-W , Mcl-1 , A1 ) can inhibit most forms of apoptosis , namely those using the release of cytochrome c from mitochondria as a signalling step . Cytochrome c is normally localised in the mitochondrial intermembrane space . Once released to the cytosol it initiates the formation of a big protein complex known as the apoptosome; in the apoptosome caspases are activated [13] . The release of cytochrome c is achieved by the action of the two pro-apoptotic effectors of the Bcl-2-family , Bax and Bak , either one alone or both together . Bax/Bak in turn are activated by one or several of the Bcl-2-family subclass of BH3-only proteins ( Bim , Bid , Puma , Noxa , Bmf , Bad , Bik , Hrk ) . The anti-apoptotic Bcl-2-like proteins can bind ( although with varying affinity ) both BH3-only proteins and Bax/Bak and probably inhibit apoptosis by this direct binding; whether they inhibit apoptosis preferentially by binding to BH3-only proteins or to Bax/Bak is under dispute . The other BH3-only proteins very likely others operate indirectly by inhibiting the suppressors ( Bcl-2 and Bcl-2-like inhibitors ) [4] , [14] . Given its anti-apoptotic function and Bcl-2-like structural fold it is very likely that F1L also inhibits apoptosis by binding of pro-apoptotic Bcl-2 family proteins . ( VACV ) F1L binding to Bak ( but not Bax ) has been demonstrated in immuno-precipitation experiments from infected cells [15] , [16] and this interaction has been mapped within the protein [17] . It has also been suggested that F1L directly binds Bim [18] . However , even the combined inhibition of Bim and Bak is unlikely to explain the profound inhibition of apoptosis in VACV- or MVA-infected cells [8] , [9] . F1L-deficient VACV and MVA induce Bax/Bak-dependent apoptosis in their host cells [9] , [17] suggesting that viral infection activates BH3-only proteins , which then activate Bax/Bak . Bim appears to play some role as Bim-deficient cells undergo slightly less apoptosis than wt cells when infected with VACV-ΔF1L [18] . Bad has also been proposed to play a role in VACV detection as MEK induced activation of Bad has been shown to be inhibited by VACV secreted VGF protein [19] . The receptors that an infected cell uses to detect the viral infection and that then activate the apoptotic apparatus , and which components of the apoptotic machinery ( besides Bim ) are involved in transmitting this signal are unknown . In this study we analysed these upstream signals . We found that both Bim and ( more importantly ) Noxa were involved in the induction of apoptosis by MVAΔF1L . Analysis of the upstream signals suggested a contribution from the type 1 interferon ( IFN ) signalling pathway in MVA-induced activation of the apoptotic pathway .
For these studies we used infection of various cell types and cells from genetically modified mice for analysis with either wt MVA or with the deletion mutant virus MVAΔF1L , which lacks the F1L gene . Expression of the F1L protein in infected cells was detected from 8 through to 48 h ( Supplementary Figure S1 ) . Upon infection of mouse embryonic fibroblasts ( MEFs ) , cell death was first detectable around 8–12 h post-infection ( Supplementary Figure S2A ) . As reported earlier [9] , infection with MVAΔF1L induced apoptosis via the mitochondrial pathway . Mitochondrial apoptosis depends on the activities of either Bax or Bak . MVAΔF1L induced apoptosis was reduced in Bak-deficient MEFs and abolished in Bax/Bak-double-deficient MEFs ( Figure 1A , Supplementary Figure S2B ) much like it has previously been reported for UV mediated apoptotic stress [20] . These results suggest that infection with MVA mainly triggers Bak but in its absence also Bax , whereas F1L blocks this process ( similar results have been reported previously for MVA by us [9] and for VACV by others [15] , [16] ) . Bax/Bak are activated by the action of BH3-only proteins . To test the contribution of individual BH3-only proteins we tested MEFs from mice deficient in individual BH3-only proteins for differences in apoptosis when infected with MVAΔF1L . As shown in Figure1B and Supplementary Figure S2C , the loss of Bim conferred a small degree of protection; a similar effect has previously been reported for VACVΔF1L [18] , [19] . No protection was seen by individual loss of Puma , Bid , Bad or Bmf ( Figure 1D , Supplementary Figure S2D ) unlike previously reported for VACVΔF1L , where Bad-specific RNAi strongly inhibited apoptosis induced by infection of HeLa cells [19] . However , MEFs deficient in Noxa were clearly protected against apoptosis induced by infection with MVAΔF1L , and this effect was stronger than the one seen in Bim-deficient cells ( Figure 1C , D , Supplementary Figure S2C ) . In earlier work we had found that there was no effect of the loss of individual BH3-only proteins during infection with MVAΔF1L ( results mentioned as data not shown in [9] ) . We are unable to retrace what it was that produced this incorrect result at the time . Noxa protein is typically expressed only at low levels but is induced by pro-apoptotic stimuli . We were unable to detect Noxa protein in MEFs ( which is most likely due to the low sensitivity of antibodies against mouse Noxa by Western blot ) and therefore used the human epitheloid cell line HeLa to test for Bim and Noxa-expression during MVA-infection . Bim protein levels decreased within 20 h . p . i , perhaps explaining the observation that Bim only plays a minor role during MVA or MVAΔF1L infection ( Figure 2A ) . Loss of Bim protein was also observed in MEFs upon infection ( data not shown ) . Infection with either MVA or MVAΔF1L led to the induction of Noxa protein in HeLa cells , beginning at about 4 h p . i . ( Figure 2B ) . Expression of F1L occurred with a similar time course ( Supplementary Figure S1 ) . This suggests that the infection by MVA induces the expression of Noxa , which is required for full induction of apoptosis , while F1L inhibits Noxa-dependent apoptosis . Noxa and , to a lesser extent , Bim are therefore triggers of apoptosis induced by MVAΔF1L . Although these data indicate that F1L blocks Bim- and Noxa-dependent apoptosis , it is not clear how this works on a molecular level . It has been suggested that F1L directly binds to and inhibits Bim [10] , [18] . We tested this hypothesis in two separate experimental systems . In the first of those , membrane fractions including intact mitochondria from uninfected control cells and from cells infected with MVA ( the latter containing F1L protein ) were incubated with a synthetic peptide comprising the Bim BH3-domain ( the active domain in BH3-only proteins ) . This leads to the activation of Bax and Bak and the release of cytochrome c [21] , [22] , [23] . In mitochondria isolated from control cells , Bim-peptide caused substantial release of cytochrome c at concentrations starting around 30 µM ( similar to results from other studies , Figure 2C ) . The release of cytochrome c from mitochondria from MVA-infected cells , and hence expressing F1L , was achieved at slightly higher concentrations of Bim peptide ( less than a 2-fold ) ( Figure 2C ) . In the second experimental setting , we used a MEF cell line where BimS ( the most active splice form of Bim ) is placed under the control of a tetracycline-inducible promoter . Addition of tetracycline to the culture medium induces Bim in these cells and causes rapid apoptosis ( Supplementary Figure S3A , B ) . Prior infection with MVA for 8 h , to ensure F1L expression , again had only a very small protective effect in this system ( Figure 2D , Supplementary Figure S3A , C ) . While overexpression of Bims is not reduced during MVA infection , endogenous Bim ( BimEL ) levels appear to be lower in MVA infected cells , as observed in Helas ( Supplementary Figure S3B , Figure 2A ) . These results suggest that direct binding of F1L to Bim , which can occur with low affinity [10] , plays only a minor role in the protection against virus induced apoptosis by F1L in these cells . Cell lines and MEFs are versatile tools for the analysis of MVA-infection in vitro but are not the natural host cells of MVA-infection in vivo . Infection of mice with a recombinant MVA driving the expression of GFP in infected cells showed that upon intravenous infection especially macrophages and dendritic cells ( DCs ) and to a lesser extent lymphocytes were infected in the spleen ( Figure 3A; Supplementary Figure S4 shows results at a 10-fold higher dosage of virus ) . We therefore undertook an analysis of the apoptotic response of these host cells to infection with MVA or MVAΔF1L . Mouse myeloid macrophages and DCs were differentiated from bone marrow . DCs differ from epithelial cells and fibroblasts in that they undergo apoptosis upon infection with wt MVA [24] , [25] . We found substantial induction of apoptosis in both macrophages and DCs upon infection with MVA ( Figure 3B , C ) . However , apoptosis induced by MVAΔF1L was much faster and reached higher levels over 48 h ( Figure 3B , C ) . As we had found for fibroblasts , the loss of Bim gave small protection against apoptosis induced by MVA or by MVAΔF1L in both cell types ( Figure 3B , C ) . We then tested apoptosis induction in macrophages , DCs , activated T-cells and B cells isolated from wt , Bim- , Noxa- or Bim/Noxa-deficient mice . MVAΔF1L induced substantial apoptosis in all these cell types while MVA-induced apoptosis reached higher levels only in the two myeloid cell types ( DC , macrophages ) but not in lymphocytes ( Figure 4 , Supplementary Figure S5 and S6 ) . There were small differences in the pattern of apoptosis-sensitivity in the various cell types . The protection against MVAΔF1L induced apoptosis by the loss of Bim was minimal in myeloid but more clearly detectable in lymphoid cells . Single loss of Noxa provided moderate protection in all cell types , and the combined loss of Bim and Noxa protected not completely but very strongly against apoptosis induced by MVAΔF1L infection in all cells analysed ( Figure 4 , S5 , S6 ) . To ensure that the decrease in apoptosis observed due to loss of Bim and to a greater extent loss of Noxa was not specific for mouse , we tested the role of BH3-only proteins in HeLa cells by RNAi . siRNAs directed against Bim , Puma or Noxa were transfected into HeLa cells prior to infection , alone or in combination . As shown in Figure S6 , there was no detectable protection by knock-down of Bim , Puma or the combination of Bim and Puma . However , Noxa-specific RNAi strongly reduced apoptosis induced by MVAΔF1L infection and inhibited the strong induction of Noxa induced by MVAΔF1L infection in these cells ( Supplementary Figure S7B , lanes 3 , 5 and 8 ) . The combination of Bim- and Noxa-knock-down failed to give more protection than RNAi against Noxa alone , suggesting that Bim played only a very minor role in these cells , as previously reported for VACVΔF1L [19] . These data suggested that the induction of Noxa protein upon infection with MVAΔF1L was a major triggering event of apoptosis . Noxa was originally identified as a p53 transcriptional target . However , macrophages lacking expression of p53 did not show a significant decrease in apoptosis in response to MVA or MVAΔF1L ( Supplementary Figure S8A ) . Subsequent work has shown that the Noxa promoter is also regulated by a CreB binding site and an interferon-stimulated responsive element ( ISRE ) [26] , and that Noxa is induced by IFN I . INFβ was clearly induced during MVA and MVAΔF1L infection ( Supplementary Figure S8B ) , as was IL-18 , a cytokine whose secretion is regulated through the inflammasome ( Supplementary Figure S8C ) . However , loss of the inflammasome activity by ASC deficiency had no effect on the apoptotic response of macrophages to MVA or MVAΔF1L ( Supplementary Figure S8D ) . Since IFN I are induced during MVA and MVAΔF1L infection within 8 h we focussed on the investigation of this system and its role in Noxa-induction and apoptosis during infection with MVAΔF1L . We first isolated and tested macrophages from mice deficient in various interferon responsive factors ( IRFs; transcription factors involved in the regulation of IFN and IFN-responsive genes ) . There was significant protection against apoptosis induced by either MVA or MVAΔF1L by the combined loss of IRF-1 and -3 or IRF-1 , -3 and -7 but not the loss of IRF-1 or IRF-1 and -7 ( Figure 5A , S6B; mice deficient only in IRF-3 were not available to us ) . IRF-3 is therefore required for the full induction of apoptosis by infection with MVA/MVAΔF1L . IRF-3 is activated by signalling pathways that can be triggered by the recognition of viral molecules . In particular , viral nucleic acids can be recognized by a number of receptors including protein kinase R ( PKR; recognizing double stranded RNA ) , cytosolic helicases ( RIG-I , MDA5; recognizing free 5′-triphosphates and double stranded RNA , respectively ) and Toll-like receptors ( TLR ) . We first tested MEFs from mice deficient in protein kinase R ( PKR ) but found no protection while MEFs double deficient in IRF-3 and -7 were significantly protected ( Figure 5B ) . We also tested macrophages deficient in MyD88 ( a major adaptor molecule in the Toll-like receptor signalling pathway ) and observed that there was no protection in response to MVA or MVAΔF1L induced apoptosis ( Supplementary Figure S8D ) . We then turned to the analysis of the signalling pathways that induce Noxa-promoter activity . A luciferase reporter construct containing either 1 . 2 kb Noxa promoter-sequence ( carrying an ISRE and bindings sites for p53 and creB ) , or the same sequence lacking the binding sites for creB and IRFs [26] were transfected into MEFs from mice with various gene deletions . These cells were then either treated with etoposide ( known to induce Noxa via p53 ) or infected with MVA or MVAΔF1L . Etoposide induced Noxa-promoter activity in cells from all genotypes tested and this induction was similar in the cases of the intact and mutant promoters ( which still contains the p53 recognition site ) , consistent with the p53-dependent induction of Noxa by etoposide ( Figure 5C ) . The induction of Noxa-promoter activity by infection with MVA or MVAΔF1L , however , was almost abrogated by the promoter deletions ( Figure 5C ) . This indicates that the Noxa induction is almost completely achieved via the IFN I system . The activity of the intact promoter upon infection was normal in IRF-1-deficient cells but strongly diminished in cells deficient in IRF-3 and -7 , again consistent with an important role of IRF-3 in virus-induced up-regulation of Noxa . When we tested cells deficient in the mitochondrial helicase adapter MAVS ( also known as IPS-1 , VISA or CARDIF ) we also found a significant reduction in Noxa-promoter induction upon infection with MVA or MVAΔF1L ( Figure 5C ) . These data suggested that the pathway to apoptosis induction involved the signalling axis from cytosolic helicases , their mitochondrial adapter MAVS and the transcription factor IRF3 . MEFs from mice lacking MAVS were significantly protected against MVAΔF1L induced apoptosis , consistent with the MAVS-dependent induction of Noxa-promoter activity observed in the experiments described above ( Figure 6A ) . Mouse Noxa protein is difficult to detect with the antibodies available . We were able to detect Noxa-induction upon infection with MVA in the presence of the proteasome inhibitor MG-132 . No Noxa was detectable upon infection of MAVS-deficient MEFs ( Figure 6B ) . A MAVS-dependent signal is thus induced upon infection with MVA orMVAΔF1L , which causes the IRF-3-dependent up-regulation of Noxa and apoptosis . One of the primary functions of this pathway is the induction of type 1 interferons [27] . INF-β secretion can be detected in the supernatant of MVA or MVAΔF1L infected macrophages ( Supplementary Figure S8B ) ( this has been reported for MVA previously [28] , [29] . IFN-β often serves to amplify an IRF-dependent response . We therefore tested the effect of IFNβ signalling on apoptosis induced by MVA or MVAΔF1L . INFβ-secretion was induced to a similar extent by infection with MVA or MVAΔF1L ( Supplementary Figure S8B ) . When IFN-β-signalling was disrupted by the use of macrophages from mice deficient in the IFN I receptor ( IFNAR ) there was again protection against apoptosis induced by viral infection to a similar extent as by the loss of Noxa or IRF-3 ( Figure 6C and D ) . However , disruption of IFN-β-signalling failed to inhibit IRF-3 translocation to the nucleus in response to MVA infection in HeLa cells ( transfected with siRNA against IFNAR ) or macrophages ( from IFNAR-deficient mice ) ( Supplementary Figure S9A , S9D and S9E ) . IRF-3 translocation to the nucleus was slightly slower in HeLa cells transfected with IFNAR-specific siRNA ( Supplementary Figure S9A , and S9B ) , concurrent with reduced apoptosis of HeLa cells treated with siRNA to IFNAR ( Supplementary Figure S9C ) . No differences were detectable by analysis of IFNAR-translocation by Western blot of cytosolic and nuclear fractions in either cell type ( Supplementary Figure S9D , E ) . The activation of IRF-3 is therefore probably insufficient for induction of Noxa; this latter step likely requires IFN-signalling through Stat proteins . NF-κB , which is activated by MAVS and can contribute to IFN I-signalling did not appear to play a role in the induction of Noxa as no NF-κB activity was induced by viral infection ( Figure 6E ) . Viral infection also failed to increase the NF-κB response to extrinsically added IL-1 ( there was even some reduction , which may be due to the expression of a functional soluble IL-1-receptor by MVA [30] , Figure 6E ) . These results show that MVA- and MVAΔF1L-infection induces apoptosis dependent on the signalling adapter MAVS , the transcription factor IRF-3 , IFN-β and the BH3-only protein Noxa . We then turned to HeLa cells , on one hand to confirm these results in another cell type and on the other hand to test the role of the helicases upstream of MAVS by RNAi . The individual genes were knocked down by transfection of siRNA , and the cells were infected with MVA or MVAΔF1L . No apoptosis induction by MVA was seen as expected ( Figure 7A ) . RNAi directed against IFNAR , IRF3 or MAVS all substantially reduced apoptosis induced by MVAΔF1L as well as the induction of Noxa by MVA ( Figure 7A ) . Mixed RNAi against both IRF3 and IFNAR showed a slightly enhanced effect . Recognition and apoptosis pathways therefore appear to be similar in mouse and human cells . MAVS is the signalling adapter of two helicases , RIG-I and MDA5 , which recognise viral RNA-species [27] . This signalling system has been very well characterised for RNA-viruses and although it is generally assumed that similar receptors exist for viral DNA the only receptor universally accepted drives the activation of the inflammasome rather than of the type 1 interferon signalling [31] . However , viral RNA produced during infection with a DNA-virus like MVA may still be recognised by the helicase machinery , and evidence has recently been reported that MVA can be recognised by MDA5 [28] . It has also been described that MVA induces the formation of viral dsRNA in the cytosol post RNA secretion from the viral particle [32] . Indeed , RNAi against MDA5 and to a slightly lesser extent against RIG-I provided protection against apoptosis induced by MVAΔF1L , concomitant with the diminished induction of Noxa protein ( Figure 7A ) . This suggested that indeed viral RNA , generated during MVA infection , is recognised by these helicases to signal through MAVS for the induction of Noxa . Although we saw no protection by transfection of a RIG-I helicase fragment ( RIG-IC; constructed as a dominant negative inhibitor of RIG-I-signalling [33] ) , the inhibitory helicase-like molecule LGP2 [34] reduced apoptosis induced by MVAΔF1L ( Supplementary Figure S10A ) . More importantly , overexpression of LGP2 but not RIG-I or RIG-IC led to an increase of FLAG expressing attached ( viable ) cells after infection with MVAΔF1L indicating protection of transfected cells against apoptosis induction by MVAΔF1L ( Supplementary Figure S10B ) . This suggested that it is the viral RNA produced during infection with MVA that is recognised and that triggers the activation of the apoptotic pathway . To test this hypothesis we performed experiments with virus treated with UV-light . Complete cross-linking of the viral genome by UV-treatment results in VACV or MVA that can still infect cells but cannot efficiently generate viral RNA [35] . UV-treated MVA or MVAΔF1L failed to induce apoptosis in HeLa cells ( Figure 7B ) . Lack of apoptosis by UV-treated MVAΔF1L correlated with lack of Noxa induction ( data not shown ) . We then infected cells with MVA or MVAΔF1L and isolated total RNA; this RNA contains host cell and viral RNA in the case of untreated virus but only host cell RNA when isolated from cells that remained un-infected or had been infected with UV-treated virus . When this RNA was transfected into uninfected HeLa cells , it induced substantial apoptosis ( to about the same extent as transfected poly I∶C or 3′-triphosphate RNA ) while RNA extracted from cells infected with UV-treated virus ( containing no viral RNA ) had no pro-apoptotic effect ( Figure 7C ) . Viral RNA mediated induction of apoptosis correlated with the induction of Noxa protein ( Figure 7D ) and was inhibited in HeLa cells carrying Noxa-specific shRNA ( Supplementary Figure S11B ) . HeLa cells transfected with siRNAspecific for MAVS , IRF-3 , MDA-5 , or IFNAR also showed decreased apoptosis when transfected with RNA from infected cells ( Supplementary Figure S11B ) . This supports the hypothesis that viral RNA , produced during infection with MVA , is the active component in terms of apoptosis induction upon virus recognition .
In this study we provide evidence that MVA triggers apoptosis in host cells upon the recognition of viral RNA by host cell helicases . This signalling utilises the known pathway encompassing cytosolic helicases , the mitochondrial adapter MAVS , the type 1 interferon signalling machinery and the induction of the BH3-only protein Noxa . Bim makes a small contribution , and F1L can inhibit apoptosis pathways triggered by MVA . How Bcl-2-family members interact to induce or to inhibit apoptosis is still not entirely clear . Bim-induced apoptosis is often linked to transcriptional induction of Bim , especially in the immune system [4] . There was no induction but a small decrease of Bim protein during MVA-infection . The role of Bim may therefore in this case lie more in the steady state inhibition of anti-apoptotic Bcl-2 proteins , which may lower the general threshold of apoptosis induction and account for the relatively mild protective effect of Bim-loss on apoptosis induction by MVAΔF1L . This interpretation is also in accordance with our finding that F1L ( as expressed during infection ) provides very little protection against direct expression of Bim or the effect of the Bim BH3-peptide . Other BH3-only proteins did not seem to make major contributions , as suggested by the lack of protection by their individual loss and by the strong protection by combined loss of Bim and Noxa . This protection was however incomplete while the combined loss of Bax and Bak provided complete protection . Therefore , other BH3-only proteins may make minor contributions that are too small to be apparent when the BH3-only protein is deleted on its own . Alternatively , there may be BH3-only independent activation of Bax/Bak . How this could occur is unclear but it has been demonstrated for heat shock [36] , and it is conceivable that viral infection causes a sufficiently strong disturbance of cellular systems to bring about such a ( still relatively minor ) effect . Upon infection of mice MVA preferentially infects macrophages and DC although T and B cells are also infected . Lymphocyte apoptosis is to a large extent regulated by Bim [37] and it is therefore not surprising that loss of Bim had the strongest protective effect in lymphocytes . The combined role of Bim and Noxa was apparent in all cell types tested . The main difference was that MVA induced substantial apoptosis in macrophages and DC but little in lymphocytes and none in fibroblasts and epithelial cells . The simplest explanation for this would be different expression levels of pro-apoptotic proteins in different cell types . It has to be pointed out that it is still uncertain how F1L functions molecularly . However relevant the detectable binding to Bak and to Bim may be ( shown for VACV F1L ) , this is not sufficient to explain the complete protection observed during infection . It has recently been suggested that VACV F1L acts like Mcl-1 [38] . Although this may be the case in the circumstance of viral infection , Mcl-1 has a much broader activity in terms of binding specificity for pro-apoptotic Bcl-2 family proteins . Noxa is required for full apoptosis in cells infected with MVAΔF1L , and F1L therefore very likely interferes with Noxa-dependent apoptosis . Since the Noxa BH3-domain does not bind to F1L with significant avidity [10] it seems unlikely that F1L binds Noxa directly . Noxa may inactivate Mcl-1 , releasing Bak , which is then inhibited by F1L . An additional function of F1L has recently been reported , namely the direct inhibition of caspase-9 [12] , and this mechanism may add to the anti-apoptotic effect of MVA-infection . The earliest components of the apoptotic pathway are therefore the BH3-only proteins Bim and Noxa . The activity of Noxa is linked to its transcriptional induction [39] . Although viral replication blocks cellular transcription , the infected cell still appears to have the time to induce Noxa protein . The induction of Noxa follows a pathway that , with some variations , has been described in other situations . RIG-I/MDA5 signal through MAVS to activate IRFs , which induce IFN-β [27] . IRF-3 may directly induce Noxa-expression , and the amplifying effect of IFN-β may be through increased IRF-3 expression [28] . Alternatively , IFN-β may induce Noxa; this is suggested by the requirement of IFN I-signalling for full apoptosis induction in cells infected with MVAΔF1L . In melanoma cells , the activation of RIG-I or MDA5 by transfection of 3′-triphosphate RNA or poly I∶C , respectively , caused Noxa-induction and Noxa-dependent apoptosis without additional type 1 interferon signalling [40] while during MVA-infection this feedback , probably signalling through Stat-proteins , appears to be necessary . Not only ( RNA-recognising ) helicases can induce apoptosis but also TLR3 ( using a mitochondria-independent pathway ) [41] , which also recognises viral ( double-stranded ) RNA . Intriguingly , the DNA-recognising TLR , TLR9 does not have this potential ( nor do the RNA-receptors TLR7/8 ) . It is intriguing to observe that even a DNA virus is recognised by RNA-sensors to induce apoptosis . It has been demonstrated recently that MVA-infection of macrophages leads to MDA5-dependent production of IFN-β while RNAi to RIG-I had no effect on IFN-β-production [28] . Our results suggest that RIG-I also makes some contribution; perhaps there is some cell-to-cell variation , or a different time course of involvement of the two helicases that makes apoptosis-induction more dependent on RIG-I than IFN-β-induction . Recent work shows that RIG-I can recognise RNA transcribed from DNA by RNA polymerase III [42] , [43] , and such a mechanism could operate for DNA-viruses . MDA5 probably mainly recognises double-stranded RNA [44] , and the recognition by MDA5 may be limited to some DNA-viruses that , like Poxviruses , generate dsRNA-intermediates due to overlapping reading frames [32] . But since all viruses at some stage will synthesise RNA , this is clearly the more general infection-associated molecule than DNA , and to focus on RNA as recognition molecule for viruses makes sense . Apoptosis is one of the defence mechanisms against viruses . Viruses pre-date the evolution of metazoan apoptosis , and it is certainly conceivable that viral infection was one of the driving forces in the evolution of the apoptotic apparatus . As an essentially cell-autonomous defence system , apoptosis does not require cellular specialisation . In a complex organism like the human , apoptosis is very likely one of several defence mechanisms but may well contribute to host defence against microbial infection .
Animal experiments were conducted according to the legal framework in Germany ( as set by federal law in the ‘German Animal Protection Act’ ( Tierschutzgesetz ) ) and had been approved by the regional authorities ( Regierung von Oberbayern; permit number:211-2531-6-8/99 ) following the regular procedures of application and approval . wt ( 3T3 ) , wt ( SV40 ) , Bak−/− ( SV40 ) , Bax−/− ( SV40 ) , Bak−/−Bax−/− ( SV40 ) , Bim−/− ( 3T3 ) , Puma−/− ( SV40 ) , Noxa−/− ( 3T3 ) , Bid−/− ( 3T3 ) , Bad−/− ( 3T3 ) , PKR−/− ( 3T3 ) , IRF-1−/− ( 3T3 ) , IRF-3−/−7−/− ( 3T3 ) , MAVS−/− ( SV40 ) , MEFs were grown in Dulbecco's Modified Eagle Medium ( DMEM ) , supplemented with 10% fetal calf serum , 1% Penicillin /Streptomycin , and 55 µM 2-mercaptoethanol . ( MAVS−/− were kindly provided by Zhijian J . Chen , Dallas ) . HeLa cells were grown in DMEM , 10%FCS , 1%Penicillin /Streptomycin . wt TetBims MEFs and T-REx 293 cells stably carrying a κB promoter luciferase reporter were cultured in DMEM , 10%FCS ( tetracycline-negative; PPA laboratories ) , 1%Penicillin /Streptomycin supplemented with 5 µg/ml blasticidin and 125 µg/ml zeocin . Macrophages and Dendritic Cells were differentiated ( 7 days ) and maintained in RPMI , 10%FCS , 1%Penicillin /Streptomycin , 55 µM 2-mercaptoethanol , supplemented with 10% M-CSF or GM-CSF enriched medium obtained from the supernatant of LCCM Hybridomas or B16 cells in the same medium , respectively . T-cells were cultured and activated in RPMI , 10%FCS , 1%Penicillin/Streptomycin , 55 µM 2-mercaptoethanol , supplemented with 1% glutamine and Con A ( 2 µg/ml; Amersham Pharmacia Bioscience ) . B cells were cultured in OPTI-MEM I+GlutaMAX-I , 10%FCS , 1%Penicillin /Streptomycin , 55 µM 2-mercaptoethanol . ER-Hox B8 wt and IFNAR−/− cells were cultured in RPMI , 10% FCS , 1%Penicillin /Streptomycin , 55 µM 2-mercaptoethanol , supplemented with 1% GM-CSF enriched medium and 1 µM βestridiol ( Sigma ) . Differentiation was achieved by washing cells and culturing them in RPMI , 10%FCS , 1%Penicillin /Streptomycin , 55 µM 2-mercaptoethanol , supplemented with 10% GM-CSF enriched medium . shHela cells were generated by retroviral infection with lentivirus driving shRNA against luciferase ( 5′-GUGCGCUGCUGGUGCCAAC-3′ ) and Noxa ( 5-GAAGGTGCATTCATGGGTG-3 ) in a GFP expressing lentiviral vector pLVTHM . Lentiviralparticles were generated by transfecting 293FT cells together with the packaging vectors pMD2 . G and psPAX2 . Surviving cells were grown and collected and sorted to produce polyclonal GFP+ populations as reported earlier [45] . Chicken embryonic fibroblasts ( CEF ) were prepared freshly from 10 day old embryos , cultured in Earl's minimum essential medium and used in the second passage . MVA and MVAΔF1L were routinely propagated and tittered by vaccinia virus-specific immunostaining on CEF . Viruses were purified by ultracentrifugation through sucrose and virus stocks were maintained at −80°C in aliquots containing ≥ 2×108 IU/ml [46] . MVA expressing the enhanced green-fluorescence gene under control of the VV natural early/late promoter P7 . 5 ( MVA-GFP ) has been described previously [47] . MVA viruses were propagated and titrated by following standard methodology [46] . Antibodies used for Western blot were specific for h/mBim , Bad , Puma ( Cell Signalling ) hNoxa ( Alexis ) , Bid ( polyclonal rabbit anti-Bid was kindly provided by David Huang , Melbourne ) β-actin , β-tubulin ( Sigma ) , F1L ( polyclonal rabbit anti-F1L was kindly provided by Antonio Postigo and Michael Way , London ) , Bak ( NT; Upstate ) , cytochrome c , hMcl-1 ( BD Pharmingen ) , mNoxa , mIL-18 ( Abcam ) , H3L ( Polyclonal antiserum specific for VACV H3 was generated by repeated immunization of a rabbit using a synthetic peptide representing amino acids 247–259 within the H3 protein as a conjugate to keyhole limpet hemocyanin described by others [48] ) , GAPDH ( Millipore ) , IRF3 ( Santa Cruz Biotechnology ) Caspase-8 ( Cell Signalling ) , Bak ( 4B5 , kindly provided by Ruth Kluck , Melbourne ) . Spleen single cell suspensions from wt , Bim−/− , Noxa−/− , and Bim−/−Noxa−/− mice ( all C57BL/6 background ) were resuspended in Amonium chloride lysis buffer ( 150 µM NH4Cl , 10 mM NaHCO3 , 1 mM Na2EDTA pH = 7 , 4 ) for 5 min and then washed in PBS . Cells were then further resuspendedin sorting buffer ( PBS , 0 . 5%BSA ) , washed and were labelled with anti-B220-FITC ( BD Pharmingen ) , FITC-anti-NK 1 . 1 ( BD Pharmingen ) , and anti-MHCII-FITC ( BD Pharmingen ) antibodies for 30 min at 4°C . Labelled cells were washed and resuspended with anti-FITC micro beads ( MiltenyiBiotec , Auburn , CA ) for 30 min at 4°C . Labelled cells were removed on MACS columns ( MiltenyiBiotec , Auburn , CA ) in order to obtain T cell populations . To check purity cells were stained with CD3-APC antibody ( Purity ≥ 90% ) . Cells were activated with ConA ( 2 µg/ml ) for 3 days and split for infection . Splenocytes from mice of the genotypes above were resuspended in PBS , washed and were labelled with anti-B220 micro beads ( MiltenyiBiotec , Auburn , CA ) for 30 min at 4°C or with FITC-anti-B220 and treated as above . Labelled cells were purified on MACS columns ( MiltenyiBiotec , Auburn , CA ) . Dendritic cells from wt , Bim−/− , Noxa−/− , Bim−/−Noxa−/− , IRF-1−/− , IRF1−/−3−/− , IRF-1−/−7−/− , IRF-1−/−3−/−7−/− and IFNAR−/− mice ( all C57BL/6 background ) were derived from bone marrow single cell suspensions cultured with enriched GM-CSF medium for 7 days . On day 1 after isolation cells were passaged to separate them from attached fibroblasts . Medium was further enriched with GM-CSF on day 3 . Floating cells were collected and split for treatment on day 7 . Macrophages were isolated in the same way but with enriched M-CSF and attached cells were used for infection on day 7 . Bim−/− , Noxa−/− , Bim−/−Noxa−/− mice were kindly provided by Claire Scott , Philippe Bouillet or Andreas Strasser , Melbourne . p53−/− mice were kindly provided by David Huang , Melbourne . IRF-1−/− , IRF-1−/−3−/− , IRF-1−/−7−/− mice were kindly provided by Hermann Wagner , Munich . IFNAR−/− and IRF-1−/−3−/−7−/− mice were kindly provided by AdmarVerschoor , Munich . Cell death was induced by treatment with MVA or MVAΔF1L ( M . O . I = 10 ) in DMEM medium 2%FCS enriched to 10% FCS 8 h . p . i for 20 h , ultra violet radiation ( UV , 100 J/m2 ) 20 h , etoposide ( 10 µM , Sigma ) 20 h , or staurosporine ( 1 µM ) 8 h . In the indicated experiments cells were exposed to MG-132 ( 40 µM ) for 4 h , 16 h . p . i . Floating cells and trypsinised attached cells were combined and incubated in PBS containing 50 µg/ml propidium iodide ( PI ) for 5 min on ice before measuring PI uptake by flow cytometry . Alternatively floating cells and trypsinised attached cells were combined and incubated in 3 . 7% paraformaldehyde for 20 min at room temperature , washed three times in PBS and incubated in mild-permeabilisation buffer ( PBS , 3% FCS , 0 . 5% Saponin ) for 20 min . Cells were stained with primary monoclonal active caspase 3 antibody ( BD Pharmingen ) and with FITC- goat anti rabbit secondary antibody ( Dianova ) and measured for FITC+ cells by flow cytometry . Cells were harvested , washed in ice cold PBS , and resuspended at 1×107cells/ml in permeabilisation buffer ( 20 mM HEPES/KOH pH 7 . 5 , 100 mM sucrose , 2 . 5 mM MgCl2 , 100 mMKCl , 0 . 025% digitonin , 1× complete protease inhibitors ( Roche ) ) . Cell membrane permeabilisation was verified by uptake of trypan blue after incubation on ice for 10 min . Membrane ( pellet ) and cytosolic ( cytosol ) fractions were separated by centrifugation at 13 , 000× g for 5 min . In the case of cytochrome c release assays the pellet fraction was resuspended in digitonin-free permeabilisation buffer and treated with Bim BH3 peptide ( MRPEIWIAQELRRIGDEFNA ; Biosynthan GmbH; HPLC cleanness >90% ) for 30 min at 30°C . Supernatant and membrane ( pellet ) fractions were separated by centrifugation ( 13 , 000× g for 5 min ) . Cytosol , supernatant and pellet fractions were assessed by Western Blot for cytochrome c , Bak as a membrane control and F1L as an infection control . In the case of IRF3 translocation assayspellet fractions were resuspended in RIPA buffer ( Sigma ) wit 1× complete protease inhibitors and prepared along side the cytosol fractions for Western blotting . MEFs carrying the tetracycline-repressor were generated as described for HeLa cells [49] . Cells were infected with MVA or MVAΔF1L for 8 h prior to addition of tetracycline ( 0 . 1 µg/ml ) for further 12 h . Mice ( n = 3 ) were infected i . v . with 3×108 IU of MVA-GFP for 3 hours before spleens were harvested . Splenocytes were stained with the following antibodies for phenotypical analysis: CD11c APC ( HL3 ) , CD3 PB ( 500A2 ) , Gr-1 PE ( 1A8; all from BD Pharmingen ) , CD11b PerCP-C5 . 5 ( M1/70 ) , MHCII eF450 ( M5/114 . 15 . 2 ) and B220 PE ( RA3-6B2; all eBioscience ) . Fluorochrome-conjugated isotype-matched monoclonal antibodies were used as controls . Anti- CD16/CD32-Fc-Block ( BD Biosciences ) was included . Propidium iodide ( Molecular Probes ) was added immediately before analysis for live/dead discrimination . Data were acquired by FACS analysis on a FACSCanto ( BD Biosciences ) and analyzed with FlowJo software ( Tree Star , Inc . ) . Macrophages ( MHC II high /CD11c intermediate / CD11b high ) , dendritic cells ( MHC II high /CD11c high ) , B cells ( MHC II high-intermediate /CD45R ( B220 ) high ) , Granuloytes ( Gr-1 high / CD11b high ) an T cells ( CD3+ ) were identified within GFP+ splenocytes . Similarly , three mice were simultaneously infected i . v . with 3×109 IU of MVA-GFP for 3 hours before spleens were harvested . Splenocyte preparations were gated for GFP+ cells . Splenocytes were further assessed for the indicated phenotypical markers . Macrophages ( CD11b+/ Gr1−; eBioscience/ BD Pharmingen ) DC ( CD11c; BD Pharmingen ) T cells ( CD3; BD Pharmingen ) , B cells ( CD19; BD Pharmingen ) and neutrophil granulocytes ( CD11b+/Gr-1+ ) . MEFs ( wt , IRF1−/− , IRF3−/−7−/− , and MAVS−/− ) were seeded at 50 , 000 cells/well in a six well plate . 24 h later cells were transfected using Fugene liposomes ( Roche ) with different reporter constructs of the Noxa promoter ( pGL3-basic plasmid carrying wt or a double mutant Noxa 1 . 2 kb promoter region[24] ) or pEGFP-C1 as a transfection control ( constructs were kindly provided by Christophe Lallemand , Villejuif ) . Twenty-four hours later transfected cells were harvested and in cases when the pEGFP controls were effectively transfected to values above 20% ( EGFP positive cells were determined by flow cytometry ) , were split in 96 well plates at 5 , 000 cells per well . 16 h post seeding cells were infected with MVA or MVAΔF1L for further 8 h . Cells were washed , lysed in lysis buffer ( Promega ) , and incubated while gently shaking at 37°C for 30 min . Luciferase activity was detected using Orion MicroplateLuminator ( Berthold Detection Systems ) and an luciferase detection buffer ( Promega ) . Wt macrophages were seeded at 300 , 000 cells/well . Twenty-four hours post seeding cells were infected with MVA or MVAΔF1L for 8 h . Supernatant was collected and stored at −80°C . IFN-β was detected using INF-β capture antibody ( US Biological ) and anti-mouse INF-β primary detection antibody ( TEBU BIO ) . T-REx 293 cells stably carrying a κB promoter luciferase reporter were seeded at 20 , 000 cells/well in a 96 well plate . Twenty-four hours post seeding cells were infected with MVA or MVAΔF1L for 8 h and treated with recombinant human IL-1β ( Peprotech ) 10 ng/ml for a further 4 h . Luciferase activity was detected as above . HeLa cells were seeded at 80 , 000 cells per well and 24 h later transfected with LipofectamineRNAiMAX ( Invitrogen ) as instructed by the manufacturer . A total of 30 pmoles of siRNA directed against IFNAR ( AACAGCCAUUGAAGAAUCUUC ) [50] , IRF3 ( AACCGCAAAGAAGGGUUGCGU ) [50] , MAVS ( CCACCUUGAUGCCUGUGAA ) , RIG-I ( GUAUCGUGUUAUUGGAUUA ) , MDA-5 ( AUCACGGAUUAGCGACAAA ) , TLR3 ( CAGCATCTGTCTTTAATAA ) or STING ( GCATCAAGGATCGGGTTTA ) were used in each transfection ( these sequences were kindly provided by Robert Besch , Munich ) . Twenty-four after transfection cells were infected with MVA or MVAΔF1L for 20 h . HeLa cells cells were seeded at 80 , 000 cells per cover slip containing well and 24 h later treated with siRNA as above . ER-HoxB8 macrophages were seeded at 150 , 000 cells percover slip containing well . MVA infection was carried out as above for the desired times of infection . Cells were washed and fixed in 3 . 7% formaldehyde prior to permeabilizing with 0 . 2% Triton-X in PBS . Cells were washed , blocked in 5%BSA PBS and incubated for 45 min in blocking buffer with a 1∶100 dilution of anti-IRF3 ( Santa Cruz Biotechnology ) . Cells were further washed and incubated in blocking buffer with a 1∶500 dilution of dyelight conjugated anti-rabbit ( Dianova ) antibody and 1∶15000 dilution of Hoechst dye ( Sigma ) for 45 min . Samples were further washed and slides were mounted . Images were taken with a Zeiss Axioplan 2 Imaging microscope using the 40× magnification lens for all images . MVA or MVAΔF1L ( 3×106 IU ) were resuspended in 1 ml DMEM containing 2%FCS and treated with 1 µg/ml of psoralen ( 4 -aminomethyl-trioxsalen ) and then irradiated with UV for 15 min in a Stratalinker 1800 UV crosslinking unit ( Stratagene , La Jolla , CA , USA ) . 1×106 uninfected or MVA or MVAΔF1L infected cells ( 4 hp . i ) were trypsinized and collected for total RNA purification . Isolation of total RNA was done with the RNeasy Mini Kit ( Qiagen ) using QIAshredder ( Qiagen ) for lysate homogenization as stated in the manufacturer's instructions . RNA concentrations collected were in the range of 150–250 ng/µl and had A260∶A280 values of 1 . 8–2 . 2 . Total RNA was stored at −80°C until used for transfections . Total RNA was transfected into HeLa cells with Fugene at a ratio of 2 µg RNA and 8 µl Fugene . HeLa cells were simultaneously transfected with pppRNA ( 1 µg∶8 µl Fugene ) or poly ( I∶C ) ( 10 µg∶8 µl Fugene ) . HeLa cells were transfected with pEF-BOS plasmids containing FLAG-tagged RIG-I , RIG-I helicase fragment RIG-IC , LGP2 or pEGFP ( transfection control ) using Fugene according to the manufacturers protocol ( RIG-I plasmids we kindly provided by Katharina Eisenächer and Anne Krug , Munich ) . The ratio Fugene ( µl ) ∶DNA ( µg ) was optimized to 8∶2 . Transfected cells were infected with MVAΔF1L for 20 h . Cells were assessed for death or for FLAG expression . Cells were fixed in 3 . 7% paraformaldehyde for 20 min at room temperature , washed three times in PBS and incubated in mild-permeabilisation buffer ( PBS , 3% FCS , 0 . 5% saponin ) for 30 min . Cells were stained with anti-FLAG M2 mouse monoclonal antibody ( Sigma ) and goat anti-mouse Alexa 488 labeled secondary antibody ( Molecular Probes ) and analyzed by flow cytometry .
|
Viruses have come up with a diverse set of mechanisms to stop infected cells from committing suicide and hence secure their own propagation . In this study we use the DNA virus Modified Vaccinia virus Ankara , a highly attenuated version Vaccinia Virus , to study how cells detect viral infection and induce apoptosis . Modified Vaccinia virus Ankara is currently in clinical trials for its use in various vaccination protocols . By using a broad array of immortalized and primary cell types we observed that viral infection induced programmed cell death was controlled by proteins predominantly involved in detection of viral RNA , in particular proteins involved in the type 1 interferon response . The novelty of our findings lies on the observation that not only can RNA from DNA viruses be detected and activate the type 1 interferon response to infection , but that these responses can also directly modulate the levels of proteins regulating programmed cell death . Future treatments of infections by viral pathogens could exploit the synergistic ability of the type 1 interferon responses and programmed cell death in order to inhibit viral propagation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"death",
"signal",
"transduction",
"signaling",
"in",
"cellular",
"processes",
"molecular",
"cell",
"biology",
"cell",
"biology",
"viral",
"immune",
"evasion",
"virology",
"biology",
"microbiology",
"apoptotic",
"signaling",
"molecular",
"biology",
"antiapoptotic",
"signaling"
] |
2011
|
Induction of Noxa-Mediated Apoptosis by Modified Vaccinia Virus Ankara Depends on Viral Recognition by Cytosolic Helicases, Leading to IRF-3/IFN-β-Dependent Induction of Pro-Apoptotic Noxa
|
While a primary genital tract infection with C . trachomatis stimulates partial-protection against re-infection , it may also result in severe inflammation and tissue destruction . Here we have dissected whether functional compartments exist in the genital tract that restrict Th1-mediated protective immunity . Apart from the Th1-subset , little is known about the role of other CD4+ T cell subsets in response to a genital tract chlamydial infection . Therefore , we investigated CD4+ T cell subset differentiation in the genital tract using RT-PCR for expression of critical transcription factors and cytokines in the upper ( UGT ) and lower genital tract ( LGT ) of female C57BL/6 mice in response to C . trachomatis serovar D infection . We found that the Th1 subset dominated the UGT , as IFN-γ and T-bet mRNA expression were high , while GATA-3 was low following genital infection with C . trachomatis serovar D . By contrast , IL-10 and GATA-3 mRNA dominated the LGT , suggesting the presence of Th2 cells . These functional compartments also attracted regulatory T cells ( Tregs ) differently as increased FoxP3 mRNA expression was seen primarily in the UGT . Although IL-17A mRNA was somewhat up-regulated in the LGT , no significant change in RORγ-t mRNA expression was observed , suggesting no involvement of Th17 cells . The dichotomy between the LGT and UGT was maintained during infection by IL-10 because in IL-10-deficient mice the distinction between the two compartments was completely lost and a dramatic shift to the predominance of Th1 cells in the LGT occurred . Unexpectedly , the major source of IL-10 was CD11c+ CD11b+ DC , probably creating an anti-inflammatory privileged site in the LGT .
Chlamydia trachomatis is an intracellular bacterium that infects the genital and ocular mucosae . The genital tract infection is the number one cause of bacterial sexually transmitted disease ( STD ) world-wide . Because the infection is asymptomatic in up to 70% of females and can result in severe damage of the reproductive tract , it is one of the major causes of tubal factor infertility [1] . It is generally agreed that the best protection against infection and sequelae could be achieved by an effective vaccine . However , vaccine development has been hampered by our poor understanding of protective immune mechanisms in the genital tract . In particular , the dichotomy between effector and regulatory functions that , on the one hand eliminate infection and on the other , could prevent immunopathology from developing , is inadequately defined for genital tract chlamydial infections . It is known that chlamydial infection of the genital tract stimulates a complex array of host innate and adaptive immune responses . Cells of the innate immune system react rapidly to recognize and limit the infection , and ultimately influence the outcome of infection through the modulation of the adaptive immune response . Studies have shown that CD4+ T cells and Th1-cells , in particular , are necessary for the effective clearance of Chlamydia from the genital tract [2] , [3] , [4] , [5] , [6] . Protective immune responses to other infections such as herpes simplex virus-2 ( HSV-2 ) and Leishmania , are also Th1-mediated and critically dependent on IFN-γ [7] , [8] . However , we know that activated CD4+ T cells differentiate into a number of T helper subsets , including: T-helper 1 ( Th1 ) , Th2 , Th17 and several subsets of T regulatory ( Treg ) cells , each subset capable of secreting a distinct cytokine profile ( reviewed [9] ) . Differentiation of Ag-primed CD4+ T cell subsets is critically dependent on the cytokine milieu that regulates CD4+ T cell subset differentiation ( reviewed [10] ) . Early on after cognate interaction with an antigen-presenting cell ( APC ) , a developmental program mediated by a group of enzymes known as transcription factors is activated in the T cells . This enzymatic activity results in the removal of covalent modifications from histone tails , and together with DNA methylating enzymes , activates selected cytokine genes . This process allows for the expression of a signature profile of cytokine genes specific for that T cell subset . At the same time it silences the expression of cytokines and transcription factors of the opposing subsets , thereby resulting in lineage restriction [11] . Although not exclusively expressed in CD4+ T cells , expression of the transcription factors T-bet , GATA-3 , RORγ-t and FoxP-3 can be used to identify Th1 , Th2 , Th17 and some Treg subsets , respectively [12] , [13] , [14] , [15] . Th1 cells are associated with the production of IFN-γ and strong cell-mediated immunity , which is thought to be the primary mechanism for clearance of Chlamydia from the genital tract and protection against reinfection [3] , [5] . The Th2 subset is associated with secretion of the cytokines IL-4 , IL-5 , IL-13 and antibody production , and is not effective in the defense against Chlamydia [16] , [17] . Studies suggest that Th2-driven antibody production is of subordinate importance during a primary chlamydial infection , although it may contribute to protection against re-infection [18] . The Th17 subset has been ascribed critical roles in several infection models and autoimmune diseases , through the production of IL-17 and IL-23 . Recently , it was demonstrated that lung chlamydial infections were dependent on IL-17 for Th1 protection to develop [19] . However , presently we do not know whether IL-17 or Th17 cells play any protective role in genital tract chlamydial infections [20] , [21] . Advances in our understanding of the immunobiology of the genital tract and better knowledge about CD4+ T cell subset immunity in response to genital tract chlamydial infections are critical elements for the development of effective vaccines . Treg cell subsets have been ascribed a dampening function on inflammation . Tregs can be divided into the naturally occurring Tregs and inducible Tregs ( iTreg ) , according to cell surface markers and cytokine producing abilities , primarily IL-10 or TGF-β . Tregs have been documented in the context of many bacterial infections , including intracellular infections with Salmonella typhimurium [22] . However , the analysis of Treg subsets and their actions in the genital tract mucosa are few , and the role of Tregs in chlamydial infections is still quite unclear . We recently reported that ICOS-deficient mice , have impaired Treg-development in response to C . trachomatis genital tract infection , which resulted in significantly augmented local inflammation and a more effective clearance of bacteria compared to that found in wild-type ( WT ) mice [23] . The relative roles of Th1 , Th2 , and to a lesser extent , Th17 cells , have been described for other bacterial infections [21] . However , their contribution to resistance and immunopathology against a genital tract chlamydial infection is incompletely known . Therefore , we undertook the present study to gain insight into the development of different effector and regulatory CD4+ T cell-subsets in response to a genital tract infection in mice with C . trachomatis serovar D . More specifically , we sought to understand the balance between effector and regulatory CD4+ T cell subsets in genital tract protective immunity and whether any distinct functional anatomical compartments could be identified .
It has previously been shown that a primary infection with C . trachomatis results in partially protective immunity against reinfection with the same serovar [24] . Following intravaginal infection , we found that peak shedding occurred between 5 and 10 days after infection . Importantly , by day 10 , clearance of the infection in the genital tract had begun , and was completely eliminated by day 32 ( Fig . 1A ) . To secure that the EIA-method used for detection of EBs reflected an ongoing infection we assessed inclusion forming units ( IFU ) in samples taken at some critical time points ( Fig . 1A ) . This analysis demonstrated good correlation with the EIA detection method , albeit assessment of IFU was more sensitive at later time points , showing a higher level of infected mice compared to the EIA-method ( Fig . 1A ) . In agreement with previous work , highly immune mice exhibited strong resistance against reinfection , suffering only a transient infection , with less than 40% of animals infected after 4 days ( Fig . 1B ) [24] . We and others have shown that CD4+ T cells are crucial for clearance of a primary infection with C . trachomatis from the genital tract and the development of protective immunity [3] , [6] , [25] . In accordance , intense CD4+ T cell infiltration during infection can be seen throughout the genital tract ( Fig . 1C ) . In order to investigate CD4+ T cell differentiation during infection , we carefully dissected the UGT , which consisted of the uterus and uterine horns , from the cervix and proximal vaginal tissue of the LGT , for analysis by RT-PCR ( Fig . 1C ) . Throughout the study this anatomical distinction was kept , separating UGT from LGT . Little is known about the kinetics of differentiation of T cell subsets in the genital tract during infection . CD4+ T cells in the early phase of the infection were more frequent in the LGT than in the UGT . However , by day 10 of the infection CD4+ numbers had begun to increase dramatically in the UGT ( Fig . 2A ) . Parallel to this we found increases in mRNA expression of several important cytokines . Whereas increases in IFN-γ mRNA expression were seen from day 10 of infection in the UGT , little change was observed in the LGT ( Fig . 2B ) . This pattern was also confirmed at the protein level by labeling of IFN- γ in frozen sections or production of IFN-γ by isolated CD4+ T cells after stimulation with PMA/ionomycin in UGT , but not by CD4+ T cells of the LGT , of infected mice ( Fig . 2C ) . By contrast , IL-10 mRNA was increased in the LGT , in particular , with high expression levels recorded on day 24 ( Fig . 2B ) . Hence , the LGT and UGT were characterized by distinct and different cytokine responses , with IFN-γ expression in the UGT and IL-10 dominating the LGT ( Fig . 2B ) . Comparatively weak expression of mRNA for IL-4 was recorded in both UGT and LGT , but the latter showed higher levels on day 15 following infection ( Fig . 2B , 2D ) . Evidence of Th17 subset activity through IL-17A mRNA expression was found predominantly in the LGT at the later time points of infection ( Fig . 2B ) . The contrasting cytokine profiles of the UGT and LGT suggested that these compartments could be functionally different and subjected to unique regulatory control allowing different developmental or selection processes for the CD4+ T cell subsets . To rule out that the functional dichotomy observed between the UGT and LGT was due to differences in antigen/infectious load we followed the IFU in the respective tissue . Detection of IFUs in the UGT and LGT over time reflected an ascending infection and comparable antigen loads ( IFUs ) in the UGT and LGT ( Fig . 3A ) . To confirm this , we challenged the mice with different doses of EBs or heat-killed EBs and we observed a similar dominance of Th1 cells in the UGT and practically no IFN-γ in the LGT irrespective of the challenge dose ( Fig . 3B ) . To further investigate the nature and location of CD4+ T cell subsets during infection , we undertook RT-PCR analysis of T helper cell differentiation by monitoring transcription factor mRNA expression in the genital tract . The transcription factor T-bet has been shown to be the master regulator of CD4+ T differentiation into Th1-type cells , which controls the expression of IFN-γ , in addition to silencing T cell transcription factors of opposing T helper-subsets [14] . The activity of T-bet in the PALN compared to the ILN was higher at all measured time points during infection ( Fig . 4A ) . Expression of T-bet mRNA in the UGT increased to a peak on day 15 ( Fig . 4A ) , representing an 11-fold increase of expression from levels in naïve tissue ( Fig . 4D ) . In striking contrast , T-bet mRNA expression was not increased in the LGT following infection ( Fig . 4A ) . Furthermore , the transcription factor GATA-3 , essential for Th2 differentiation , peaked later than T-bet transcription on day 24 with a 9-fold increase in mRNA for GATA-3 exclusively up-regulated in LGT and not in the UGT ( p<0 . 05 ) ( Fig . 4B , D ) . Of note , in naïve and infected mice GATA-3 mRNA expression was higher in the ILN than in the PALN , which contrasted with the pattern seen for T-bet mRNA expression ( Fig . 4B , D ) . These expression patterns for Th1 and Th2 activity was also obtained when the transcription factor mRNA expression was normalized against CD3-□ mRNA levels in the respective tissues ( Fig . 4E ) . Thus , we found distinct and unique expression patterns of T-bet ( Th1 ) and GATA-3 ( Th2 ) mRNA in the genital tract in response to a C . trachomatis infection with a dominance of T-bet in the UGT and PALN , while GATA-3 was exclusively upregulated in the LGT , supporting that the UGT and LGT were functionally separate compartments . Th17 is a recently identified T helper subset that has been ascribed important roles in tolerance , autoimmune diseases and infections [11] , [20] , [26] . This population of T helper cells remains poorly studied in the genital tract , and to date , its role in Chlamydia genital tract infection has not been described . The Th17 subset differentiates under the control of the transcription factor RORγ-t [15] . Thus , we used RT-PCR to report the presence of Th17 cells through the expression of RORγ-t mRNA during Chlamydia genital tract infection . The expression of RORγ-t mRNA in the draining lymph nodes did not significantly change from naïve levels during infection , although there was an initial decrease in expression from naïve levels observed in the ILN ( Fig . 4C , D ) . During homeostatic conditions in the UGT of naïve mice , RORγ-t mRNA expression was low , whereas these levels were somewhat higher in the LGT ( Fig . 4C , E ) . However , following infection , RORγ-t mRNA expression decreased in the LGT and UGT ( Fig . 4E ) . These data suggested that the Th17 subset were not expanded in the genital tract in response to a C . trachomatis infection . The balance between CD4+ effector cell populations and Tregs is considered critical for limiting the immunopathological outcome of a chlamydial infection . We have recently shown that lack of FoxP3+ Tregs , as seen in the ICOS-deficient mice , increases the risk of developing severe immunopathology following a genital tract infection with C . trachomatis [23] . Here we found an increase in expression of FoxP3 mRNA in the LGT and UGT ( Fig . 5B ) . FoxP3 mRNA levels in the UGT increased gradually from day 10 , and by day 24 were at 30-fold of those observed in naïve mice , whereas levels in the LGT had increased 6-fold by day 24 ( Fig . 5B , C ) . By contrast , RT-PCR analysis of the draining lymph nodes revealed little activity and only decreased FoxP3 mRNA levels followed upon infection ( Fig . 5A ) . To conclude , we observed a relative increase in FoxP3 mRNA expression in the UGT , paralleling the increase in mRNA expression of T-bet ( Fig . 5B ) . By contrast , the LGT , dominated by GATA-3 mRNA expression , and exhibited less of a change in FoxP3 mRNA in response to a genital tract infection with C . trachomatis . Given that we observed an association between enhanced T-bet mRNA expression in the UGT and the PALN , while ILN only weakly expressed T-bet and more distinctly GATA-3 mRNA , we hypothesized that imprinting of UGT CD4+ T cells could have occurred in the PALN rather than in the ILN . To investigate to what extent the differences in transcription factor mRNA expression in UGT and LGT reflected differential homing properties acquired by primed T cells in the PALN and ILN , we adoptively transferred GFP-transgene expressing T cells from either the ILN or PALN of infected or naïve mice into naïve recipient C57Bl/6 mice or mice that had been infected 10 days earlier . Irrespective of whether the CD4+ T cells were isolated from the PALN or ILN , we found T cells were able to home to both the UGT and the LGT of Chlamydia-infected mice , albeit more cells ended up in the UGT ( Fig . 6A , B ) . Of note , GFP+ T cells could not be found in the genital tract of adoptively transferred naïve mice , indicating that homing of specific T cells to the genital tract does not occur in the absence of infection ( data not shown ) . Therefore , CD4+ T cells acquire homing properties in the PALN or ILN that do not discriminate between UGT and LGT , suggesting that these two functionally distinct compartments appeared to be influenced by local factors in the respective tissues , rather than unique imprinting by antigen-presenting DC restricted to the PALN or the ILN . Thus , the anatomically distinct T helper subset profiles observed could have been influenced by local factors in the tissues . Since we observed a significantly increased level of IL-10 mRNA expression in the LGT on day 24 of the infection and that this cytokine is a particularly strong inhibitor of Th1-development we analyzed the possible source for the production of IL-10 in the tissue . To locate the cellular source of IL-10 , we isolated distinct populations of cells , with purity of over 97% , from the genital tract of naïve mice and Chlamydia infected mice . Surprisingly , the major source of IL-10 mRNA in the LGT of infected mice was found to be the classical DC population ( cDC; CD11b+CD11c+ ) ( Fig . 7A , B ) . Labeling of frozen sections of LGT with anti-IL-10 also confirmed the presence of CD11c+ cells producing this cytokine on day 24 following infection ( Fig . 7C ) . Moreover , after saponin extraction of cytokines in biopsies of LGT we could detect IL-10 by ELISA , but only in samples from infected mice ( Fig . 7C ) . The high IL-10 mRNA expression in the cDC of the LGT was observed on day 24 after infection , while at earlier time points the IL-10 mRNA expression in LGT did not differ from that in naïve mice , although this was substantially higher levels than in UGT cDC . ( Fig . 7D ) . Interestingly , on day 24 post-infection we also observed an increase in the size of the cDC population in the LGT ( Fig . 7D ) . By contrast , only weak expression of IL-10 mRNA was found in macrophages ( CD11b+F4/80+ ) , plasmacytoid DC ( pDC; CD11c+CD11b-CD19-B220+ ) , CD8α+ DC ( CD11c+CD8α+ ) , CD4+ T cells ( CD4+ CD3+ ) , or epithelial sheets ( Fig . 7B and data not shown ) . To rule out that Tregs were directly or indirectly responsible for the IL-10 dominance in the LGT we undertook adoptive transfer experiments with CD4+ T cells from IL-10-deficient ( IL-10−/− ) or wild-type ( WT ) mice injected into nu/nu mice . Following a genital tract primary infection , nu/nu mice responded with IL-10 mRNA expression levels in the LGT of comparable magnitude irrespective of if the CD4+ T cells were IL-10-deficient or normal ( Fig . 7E ) . As before , the IL-10 mRNA was restricted to the LGT and only expressed at low levels in the UGT of infected mice ( Fig . 7E ) . This result clearly demonstrated that CD4+ T cells were not directly or indirectly responsible for the IL-10 production in the LGT . Taken together , the LGT mucosa appeared to be a privileged tissue through anti-inflammatory activity , provided by regulatory cDC producing IL-10 . Hence , our data suggested that Th1-effector T cell activity in the LGT in response to a genital tract chlamydial infection was restricted by DCs . Finally , to test the notion that IL-10 was a locally produced factor responsible for establishing a privileged compartment in the LGT we undertook experiments in IL-10−/− mice . Given that it has been reported that IL10−/− mice display enhanced immunity to chlamydial infection , we asked if the LGT compartment was dominated by Th1 effector cells rather than Th2 cells , as seen in WT mice [25] , [27] . To this end , IL-10−/− mice were infected with C . trachomatis and the mRNA expression of transcription factors T-bet , GATA-3 , RORγ-t and FoxP3 was analyzed by RT-PCR . We found that IL-10−/− mice displayed enhanced clearance of infection , as reported earlier [27] . In fact , elimination of bacteria was complete by day 15 in IL-10−/− mice , at a time point when more than 40% of WT mice remained infected ( Fig . 8A , B ) . With regard to transcription factor expression we found that T-bet mRNA expression in the LGT of IL-10−/− mice was strikingly up-regulated , representing a 50×106-fold increase over WT levels at the same time point ( Fig . 8C ) . This dramatic shift in T-bet and Th1-development was clearly a local consequence of lack of regulatory IL-10 in the LGT as the level of T-bet mRNA expression in the UGT was relatively unchanged in these mice ( Fig . 8C ) . Also , GATA-3 and RORγ-t mRNA expression levels both in the UGT and LGT were relatively unaffected in the absence of IL-10 , further supporting the notion that IL-10 in LGT renders this tissue a status of an anti-inflammatory privileged site ( Fig . 8D , E ) . Interestingly , whereas we had observed that T-bet mRNA expression levels in the UGT of WT mice were accompanied by a corresponding increase in FoxP3 mRNA expression , this did not apply to the LGT in IL-10−/− mice , where FoxP3 mRNA expression levels were not significantly changed ( Fig . 8F ) . The absence of IL-10 did not change TGF-β mRNA expression levels in either the UGT or LGT compared to that seen in WT mice ( not shown ) . These results favor that local production of IL-10 in response to a genital tract infection with C . trachomatis prevents the development of strong Th1-immunity in the LGT . Hence , we propose that cDC derived IL-10 provides conditions for a privileged compartment in the LGT .
In the present study we set out to learn more about the different effector and regulatory CD4+ T cell populations induced by a primary genital tract infection with C . trachomatis , an obligate intracellular bacterium . Unexpectedly we found that a key player in establishing the peripheral CD4+ T cell repertoire in the local genital tract tissue was a population of classical CD11c+ CD11b+ DCs , which produced significant levels of IL-10 . This production appeared to prevent Th1-dominance in the LGT , whereas the latter was clearly the main CD4+ T cell population in the UGT . Hence , we propose that anti-inflammation prevails in the LGT as a consequence of regulatory cDCs producing IL-10 . This could possibly impair anti-infectious effector functions , while preventing unwanted tissue destruction exerted by the Th1 cells during a Chlamydia-infection . The LGT could , thus , be viewed as a privileged site . Here we report a functional distinction between the LGT and UGT , driven by a local cytokine production , as immune protection against a genital tract infection develops . Surprisingly , we found that IL-10 played a critical role for this effect . The evidence for that was derived from experiments in IL-10−/− mice where this functional dichotomy was completely lost and both tissues were dominated by CD4+ Th1-cells . Furthermore , it was clear that imprinting of homing abilities in newly primed CD4+ T cells in the draining lymph nodes , ILN or PALN , did not support such a dichotomy , since irrespective of their origin , CD4+ T cells were found in both UGT and LGT . Rather , it appeared that the IL-10 production in the LGT influenced the accumulation and/or differentiation of Th2 cells and prevented the expansion of Th1 cells in that tissue . Furthermore , adoptive transfer experiments showed that the ability to produce IL-10 was not a critical property of the CD4+ T cells as such since IL-10-deficient cells in nu/nu mice did not alter the functional dichotomy between LGT and UGT seen in WT mice . Collectively our data suggest that genital tract cDC could exert a regulatory function by influencing the local CD4+ T cell repertoire in the LGT in response to a genital tract chlamydial infection . CD4+ T cells are of fundamental importance for protection against C . trachomatis genital tract infections , yet a detailed understanding of functional qualities of CD4+ T cell subsets during infection is limited [5] , [6] , [18] . Such studies have been technically demanding especially because of the difficulties in isolation of T cells from the tissues [28] . To circumvent this problem , we developed RT-PCR assays to detect mRNA expression of transcription factors and cytokines in the genital tract . Hence , for the first time , using very specific and sensitive methods , we were able to follow global CD4+ T cell differentiation and immune regulation in the genital tract in response to a C . trachomatis infection . Quite unexpectedly , we found evidence that unraveled a complex system of expansion and regulation of CD4+ T helper subsets in the UGT and LGT , establishing a clear functional dichotomy between the two compartments . Whereas , a strong Th1 profile was induced in the UGT with an exclusive presence of IFN-γ , the LGT appeared to be a privileged site with anti-inflammatory IL-10 and a dominance of Th2 cells . Our initial theory was that the dichotomy between LGT and UGT was established by imprinting different homing properties on CD4+ T cells by antigen-presenting cDC in the regional lymph nodes . This was because we observed an up-regulation of T-bet mRNA in the PALN and of GATA-3 mRNA in the ILN we thought that a differential homing pattern could have explained the selective enrichment of Th1 and Th2 cells in the UGT and LGT , respectively . However , this notion was not supported by our finding that T cells from both PALN and ILN were capable of homing to the UGT and LGT . Rather , an alternative explanation for the dichotomy was considered , namely that local production of regulatory cytokines in the LGT promoted Th2 cells and largely prevented Th1 cells in this tissue . This latter theory was also supported by two observations; i ) that local cDC in the LGT produced IL-10 and that ii ) in IL-10 deficient mice the LGT environment was changed and instead hosted a dominant Th1 cell population in response to the genital tract infection . Moreover , primed and tissue migrating CD4+ T cells , including Tregs , were not required to produce IL-10 to allow for the dichotomy between LGT and UGT . The importance of local production of immune regulating factors in the genital tract is supported by findings reported by Maxion et al . , who showed that chemokines associated with Th1 responses , namely CXCL10 , CXCL9 and CCL5 were found exclusively in the oviducts , while the Th2-associated chemokine CCL11 was elevated primarily in the cervical region following infection [29] . However , to accommodate our results with these findings we must assume that the selective expression pattern of chemokines in the LGT is influenced by IL-10 and that in the absence of this cytokine the expression of CXCL10 , CXCL9 and CCL5 prevails also in the LGT , allowing for the migration and accumulation of Th1 cells in both UGT and LGT , in agreement with what we observed in IL-10-deficient mice . Indeed , previous studies have documented that endogenous IL-10 plays a crucial down-modulating role on both CC and CXC chemokine expression and neutrophil influx , in e . g lung and gut intestinal tissues [30] , [31] . Moreover , several reports associate CXCL10 expression with tissue-recruitment of Th1 cells and IL-10 production strongly inhibits CXCL10 expression [32] , [33] . Hence , we propose that local IL-10 production by cDC is the key factor in maintaining LGT an anti-inflammatory privileged site , down-modulating CXCL10 and preventing Th1 cell influx . Whether withdrawal of IL-10 in the LGT , as in IL-10−/− mice , allows for CXCL10 , CXCL9 and CCL5 chemokine production to increase will be investigated in future studies . In the present study we have documented IL-10 producing cDC in the LGT , which appeared to control the distribution of CD4+ effector T cells and secured that the LGT was a privileged compartment during C . trachomatis infection . However , to unequivocally document that IL-10 producing cDC in the LGT were responsible for the lack of Th1 cells and dominance of Th2 and GATA3-expressing cells in the LGT , we would have to engineer a mouse model where these cells could be selectively depleted . Unfortunately , there exists no such model at present given that depletion of cDC in general in the CD11c-DTR ( diphtheria toxin receptor ) mouse model would take away all DCs , leaving no DCs for priming of a T cell response , plus the fact that repeated injection of DT would be required , which also depletes other cell subsets , including plasma blasts , activated CD8+ T cells , NK cells and some populations of macrophages [34] , [35] , [36] , [37] . We are , therefore , currently exploring the possibility to generate a mouse model with IL-10-deficiency in the cDC population through mating mice with a loxP IL-10 gene with mice CD11c-cre mice . We found evidence that not only CD4+ effector T cells , but also Tregs , may be differentially distributed to the UGT and LGT . We found that increases in FoxP3 mRNA expression occurred in the UGT already 10 days after inoculation , suggesting that early Th1 migration into the UGT promoted the establishment of FoxP3+ Tregs in this tissue . Whereas there was a clear association between increases in Th1 cells and Tregs in the UGT , such a pattern was not as clear for the LGT , as the RT-PCR detection revealed only 6-fold increases in FoxP3 mRNA in the LGT as compared to nearly 40-fold increases in the UGT over the course of the infection . Functionally , this observation agrees well with many previous reports because Tregs could effectively dampen the Th1 activity and protect against tissue damage [38] . As IL-10 from cDC appears to be the main anti-inflammatory factor in the LGT , Tregs could fulfill this function in the UGT , again pointing to the dichotomy and very compartmentalized functions of the LGT and UGT . At present we can only speculate that the mechanism by which Tregs limit immunopathology in the UGT is through TGF-α or other mechanisms rather than IL-10 , since IL-10 mRNA levels were low in the UGT [39] . Studies of Th17 cells have largely demonstrated their involvement in autoimmunity and recently also their role in host defense against bacteria [20] , [21] , [40] , [41] . Th17 cells are induced in a number of bacterial infections including Salmonella enteritidis , Helicobactor pylori , and Mycobacterium tuberculosis [21] , [41] , [42] , [43] . Although there are 6 members of the IL-17 family , Th17 cells produce large amounts of IL-17A and , therefore , most effector functions of Th17 cells have been attributed to the production of this cytokine . Th17 cells are induced by the cytokines TGF-β and IL-6 , which have both been shown to be produced in response to C . trachomatis [44] , [45] , [46] . Contrary to other studies of bacterial infections we failed to detect any major alterations in the Th17 subset in either the UGT or the LGT , as assessed by detection of mRNA for RORγ-t . We must , therefore , conclude that this subset appears not to play a role in host protection against a genital tract infection with C . trachomatis . This result is also at variance with a recent study by Bai et al . , who reported on lung infections induced by Chlamydia muridarium , the mouse-specific Chlamydia species . These authors treated mice with neutralizing anti-IL17 antibodies and found poor immunity to infection [19] . As IL-17 has not been found to inhibit chlamydial growth the effect of anti-IL17 antibody treatment was rather attributed to poor IL-12 and strong IL-10 production by DC , leading to reduced Th1-effector functions [47] . Similar to the chlamydial lung infection , the Th17 response in M . tuberculosis infections is induced rapidly and is necessary to attract Th1 cells to the lung to enhance the adaptive immune response [43] . Although , we observed some increases in IL-17 mRNA in the LGT by day 24 following infection a concomitant increase in RORγ-t mRNA was not observed . Hence , we do not think Th17 cells are prominent in chlamydial genital tract infections . It should be emphasized that the Th17 subset is not the only cell type capable of producing IL-17; γδ+ T cells [48] , NKT cells [49] , neutrophils [50] and even FoxP3+ T cells [51] have also been shown to produce this cytokine . However , because of the complexity of the CD4+ T cell response that we observed a similar type of anti-IL17 antibody treatment experiment should be performed to rule out any involvement of IL-17 in protective immunity against C . trachomatis genital tract infections . Recently , Moniz et al . described two subsets of DC in the mouse genital tract in response to C . muridarium infection [52] . These authors found that DC primed Th1 cells , while pDC produced IL-6 and IL-10 and primed non-Th1 cells . Our findings are somewhat T at variance with this observation , in that pDC from the LGT did not produce significant levels of IL-10 , while cDC did . However , the cDCs of the LGT may represent a subset of cells resembling those found in the gut intestine . Recent elegant studies by Varol et al . have documented several subsets of lamina propria DCs , especially a population of non-monocyte-derived CX3CR1- CD103+ were reported critical for homeostasis , whereas a monocyte derived DC population failed to control inflammation [53] . Also in the Peyer's patches , a DC subset has been reported which preferentially secretes IL-10 and generates Th2 responses [54] . By contrast , splenic DC produce IL-12 and favor Th1 generation . In agreement with this latter notion , Th1 instead of Th2-responses were induced in IL-10−/− mice . Th1 polarization of the immune response in the absence of IL-10 correlated also with better protection . IL-10 producing cDCs following chlamydial infection of the lung have been shown to reduce allergen-specific cytokine production and CD4+ T cell responses [55] . In the lung infection model of C . muridarium , adoptively transferred DCs prevented Th1 cell expansion , indicating that DC in chlamydial infections have a regulatory function [19] . These DC produced high levels of IL-10 , which resulted in poor Th1 expansion and poor clearance of bacteria from the lung [19] . The most important observation in the present study was that IL-10 production by cDC coincided with a lack of Th1 expansion in the LGT . Previous studies have reported that in IL-10−/− mice , resistance to Chlamydia was associated with early maturation and activation of DCs in the draining lymph nodes , enhanced antigen presentation and stimulation of increased IFN-γ production from the T cells [25] . Our present data complement these findings by showing that IL-10 strongly regulates the presence of Th1 cells in the LGT . Interestingly , this effect was restricted to Th1 functions , since no changes in Th2 or Th17 transcription factor mRNA expression were observed in the LGT of IL-10−/− mice . Whether these regulatory cDC migrated into the tissue or were resident in the LGT as the genital tract C . trachomatis infection ascended is not known . As aforementioned , the literature support for that they were derived from monocytes is weak as such DCs have nearly always been associated with pro-inflammatory responses [53] . Rather , it appears that the LGT cDC are bone-marrow derived and at least three subsets of these cells have been described in the mouse vaginal epithelium [56] . We believe that a better understanding of the functional dichotomy between UGT and LGT and the role of cDC-derived IL-10 in regulating the CD4+ T cell repertoire in the LGT is of vital importance to the development of future effective and safe local vaccines against C . trachomatis . Such knowledge could also have important implications for how to prevent the immunopathology associated with genital tract chlamydial infections .
A human genital tract clinical isolate of C . trachomatis serovar D was propagated in buffalo-green monkey kidney cells and purified by centrifugation . Chlamydia IFU were enumerated using the method described below and stored in sucrose–phosphate–glutamate ( SPG ) buffer at −80°C . The infectivity of C . trachomatis stocks of elementary bodies ( EBs ) was tested by intravaginal infection of C57BL/6 mice using a range of doses . We found that 106 inclusion forming units ( IFU ) was required for 100% of the mice to be infected on day 8 . 6–8 week old female C57BL/6 mice were purchased from Taconic ( Denmark ) . IL10−/− mice were bred at the Department of Experimental Biomedicine at the University of Gothenburg , Sweden [57] . UBI-GFP/BL6 mice [58] , which express a transgene coding for green fluorescent protein ( GFP ) under control of the human ubiquitin C promoter , were used for transfer experiments . Nu/nu mice were purchased from Taconic ( USA ) . All experiments include DepoProvera treated ( 2 . 5 mg subcutaneously 7 days prior to analysis ) naïve controls of each group . Mice were maintained under specific pathogen-free conditions , according to FELASA specified guidelines , at the Department of Experimental Biomedicine at the University of Gothenburg , Sweden . Approval was obtained from Swedish Animal Welfare Agency . Mice were given 2 . 5 mg subcutaneous injection of medroxyprogesterone acetate ( DepoProvera , Pharmacia Sverige AB ) 7 days prior to the inoculation of approximately 1×106 inclusion forming units ( IFU ) , 105 IFU , or 106 heat-killed IFU of C . trachomatis elementary bodies ( EBs ) intravaginally . Four weeks after the resolution of the primary infection , the inoculation procedure was repeated in a manner identical to that described for the primary infection . Bacterial shedding was monitored at 2 , 4 , 8 , 16 and 24 days post-reinfection . Detection of C . trachomatis infection was performed in two ways . Firstly , bacterial shedding was monitored at 8-day intervals using a commercial MicroTrak II Chlamydia EIA kit ( Trinity Biotech plc . ) according to manufacturer's instructions . Samples with an absorbance greater than the provided cut-off value were considered positive for chlamydial shedding . This detection method correlates closely with assessment of inclusion forming units ( IFU ) as described [2] . To confirm the level of infection at each time point , the number of IFU were enumerated by infection of HeLa cell monolayers , as previously described [59] . Swabs were collected in SPG buffer , vortexed and centrifuged at 13000 rpm for 10 minutes . Samples were then sonicated for 30 seconds . EBs were then added to HeLa cell monolayers and centrifuged at 280 ×g at room temperature for 1 hour followed by incubation at 37°C for 30 minutes . Plates were washed with 1× with HBSS and culture medium containing cycloheximide ( 2 µg/ml ) was added . After 40 hours incubation at 37°C in a 5% CO2 , infected monolayers were fixed by addition of 100 µl/well of methanol at room temperature for 20 minutes . The infected monolayers were stained with biotin-conjugated anti-MOMP antibody ( Abcam ) , followed by streptavidin-alkaline phophatase ( Dako ) , and developed by the addition of 1-Step NBT/BCIP reagent ( Pierce ) . The reaction was stopped by rinsing with water , and the plates were allowed to air dry before counting . The uterus/uterine horns ( upper genital tract; UGT ) , the vagina/cervix ( lower genital tract; LGT ) , para-aortic lymph nodes ( PALN ) and inguinal lymph nodes ( ILN ) were collected at the indicated time points . Tissues were stored in RNAlater ( Qiagen ) before total mRNA was isolated using Qiagen homogenizer and RNAeasy minicolumns ( Qiacube , Qiagen ) according to manufacturer's instructions . The resulting extraction was used for cDNA synthesis using oligo ( dT ) primer and SuperScript RT ( Invitrogen Life Technologies ) and analyzed by RT-PCR . Primers ( MWG-biotech ) used for the determination of transcription factor mRNA levels using SYBR green technology were as follows: GATA-3 forward ( 5′- CTT ATC AAG CCC AAG CGA AG -3′ ) , GATA-3 reverse ( 5′- CCC ATT AGC GTT CCT CCT C -3′ ) , T-bet forward ( 5′- TCAACCAGCACCAGACAGAG -3′ ) , T-bet reverse ( 5′- AACATCCTGTAATGGCTTGTG -3′ ) , Foxp3 forward ( 5′- AGC TGG AGC TGG AAA AGG A -3′ ) , Foxp3 reverse ( 5′- GCT ACG ATG CAG CAA GAG C -3′ ) , IFN-γ forward ( 5′- GCC ATC AGC AAC AAC ATA AGC -3′ ) , IFN-γ reverse ( 5′- TGA GCT CAT TGA ATG CTT GG -3′ ) , IL-10 forward ( 5′- GCT CCT AGA GCT GCG GAC T -3′ ) , IL-10 reverse ( 5′- TGT TGT CCA GCT GGT CCT TT -3′ ) , IL-4 forward ( 5′- CATCGGCATTTTGAACGAG -3′ ) , IL-4 reverse ( 5′- CGAGCTCACTCTCTGTGGTG -3′ ) , IL-17A forward ( 5′- TGT GAA GGT CAA CCT CAA AGT C -3′ ) , IL-17A reverse ( 5′- AGG GAT ATC TAT CAG GGT CTT CAT T -3′ ) RORγ-t forward ( 5′- GGT GAC CAG CTA CCA GAG GA -3′ ) , RORγ-t reverse ( 5′- CCA CAT ACT GAA TGG CCT CA -3′ ) , TGF-β forward ( 5′- CAC CGG AGA GCC CTG GAT A -3′ ) , TGF-β reverse ( 5′- TTC CAA CCC AGG TCC TTC CTA -3′ ) , CD3□ forward ( 5′- CCAAGGAAACCAACTGAGGA -3′ ) , CD3□reverse ( 5′- TTGATTCTGGGTGCTGGATAG -3′ ) , HPRT forward ( 5′- TCC TCC TCA GAC CGC TTT T -3′ ) , HPRT reverse ( 5′- CCT GGT TCA TCA TCG CTA ATC -3′ ) . RT-PCR was performed using the LightCycler System and Relative Quantification software ( Roche Diagnostics , GmbH ) . Results were expressed as a normalized ratio of the target mRNA to housekeeping mRNA . The reproductive tract , including the uterus/uterine horns ( upper genital tract; UGT ) and the vagina/cervix ( lower genital tract; LGT ) , was removed , snap frozen in TissueTek OCT Compound ( Histolab Products AB ) and stored at –80°C within 2 hours . Cryostat sections of 7 µm were fixed in acetone before blocking with 0 . 3% H2O2 , blocking using an avidin-biotin blocking kit ( Vector Laboratories ) , followed by 20% normal horse serum . Sections were incubated with anti-CD4-biotin ( BD PharMingen ) followed by anti-rat IgG and developed using peroxidase-conjugated avidin ( DAKO Cytomation ) and a commercial peroxidase AEC substrate ( Sigma-Aldrich ) . Sections were counterstained with HTX and mounted with Faramount ( Histolab Products AB ) . For intracellular cytokine staining , sections were fixed as above and permeabilised in 0 . 1% saponin/PBS . Sections were incubated with biotin-conjugated anti-IL-10 or FITC-conjugated anti-IFN-γ and biotin-conjugated anti-CD4 or FITC-conjugated anti-CD11c ( BD Pharmingen ) , followed by streptavidin-conjugated TxRd ( Vector ) and Topro-3 ( Invitrogen ) . Negative controls were stained with isotype-matched irrelevant antibodies or the secondary antibody in absence of a primary antibody . Sections were visualised using a Leica LSC microscope or Zeiss LSM 510 Meta confocal microscope . The draining lymph nodes ( ILN and PALN ) of the genital tract were harvested from UBI-GFP/BL6 mice , after 7 days of infection with C . trachomatis . CD4+ T cells were purified by negative selection using MACS . Briefly , single cell suspensions were prepared and incubated with CD4+ T cell biotin antibody cocktail ( Miltenyi Biotec ) for 10 minutes at 4°C , followed by biotin beads for 15 minutes . The cell suspension was transferred to a CS MACS column according to the manufacturer's instructions ( Miltenyi Biotec ) . Purity of eluted T cells was controlled using FACS ( typically around 90% ) . 1×106 cells were injected i . v . into recipient DepoProvera-treated C57BL/6 mice either before or after 10 days of infection . For nu/nu experiments , CD4+ T cells were isolated as describe from naïve IL10−/− or C57BL/6 ( IL-10+/+ ) mice and 1×106 cells were injected i . v . into recipient DepoProvera-treated nu/nu mice . One day later , two groups of nu/nu mice were infected with C . trachomatis , while two remained uninfected controls . Isolation of lymphocytes from the UGT and the LGT was performed as previously described [60] . Briefly , the genital tract tissue was washed in calcium- and magnesium-free HBSS ( CMF-HBSS; Life Technologies ) , supplemented with 25 mM HEPES ( Life Technologies ) , and then incubated at 37°C in CMF-HBSS containing 5 mM EDTA ( Merck ) and 10% heat-inactivated horse serum ( Life Technologies ) . Following each incubation , the supernatant containing the sloughed epithelial cells and the intraepithelial lymphocytes were collected , centrifuged and prepared for RNA isolation . For isolation of the mucosal lymphocytes the remaining tissue was incubated three times for 60 minutes with collagenase D ( 120IU/ml; Sigma-Aldrich ) dissolved in RPMI 1640 containing 25 mM HEPES and 20% inactivated horse serum ( Life Technologies ) . Cell suspensions were washed and stained with anti-CD11c , anti-CD11b , anti-MHCII , anti-CD4 , and anti-CD8 ( BD Pharmingen ) for 30 minutes on ice . The cells were then washed twice with PBS containing 0 . 1% BSA and sorted using a FACSAria ( BD biosciences ) . Sorted cells were sorted into PBS/0 . 1% BSA , immediately centrifuged and resuspended in 350 µl buffer RLT ( Qiagen ) for subsequent RNA extraction as described above . Sorted populations were defined as following: CD4; CD4+CD3+ , CD80+; CD11c+CD8α+ , MΦ; CD11b+F480+ , pDC; CD11c+CD11b-CD19-B220+ , cDC; CD11b+CD11c+ . Cytokine levels were measured from genital tract tissue or from in vitro stimulated sorted CD4+ T cells . Briefly , tissues were weighed and immediately placed in 10 volumes ( wt/vol ) of a protease inhibitor cocktail containing 10 mM EDTA , 2 mM PMSF , 0 . 1 mg/ml soybean trypsin inhibitor , 1 . 0 mg/ml BSA , PBS , pH 7 . 0 . Tissues were incubated at 2% saponin at 4°C over night . Samples were clarified by centrifugation at 13000× g for 10 minutes at 4°C . Protein concentration was determined using a cytometric bead array ( CBA; BD biosciences ) or ELISA ( BD biosciences ) . For sorted cells , CD4+ T cells were incubated at 37°C , 5% CO2 , for 72 hours in the presence of PMA ( Sigma ) at 10 ng/mL and ionomycin ( Sigma ) at 1 µg/mL . Cell culture supernatants were analyzed by cytometric bead array ( CBA ) according to the manufacturer's instructions ( BD biosciences ) or by intracellular cytokine staining for analysis by flow cytometry . CD4+ T cells isolated from the genital tract of infected or naïve mice were stimulated with PMA/ionomycin for 3 days as described above . Cell suspensions were incubated for the final 5 hrs at 37° in the presence of 5 µg/ml Brefeldin A ( Sigma-Aldrich ) . Cells were stained for surface molecules , fixed with 2% formaldehyde ( HistoLab Products AB ) and re-suspended in permeabilization buffer containing HBSS , 0 . 5% bovine serum albumin ( BSA ) , 0 . 5% saponin and 0 . 05% azide . FITC-conjugated anti-IL-4 or IL-5 were added . Cells were detected using an LSRII flow cytometer ( BD Biosciences ) with diva software ( BD Biosciences ) . Data were analysed using Flowjo software ( Tree Star Inc ) . Mann-Whitney or Dunnett's C non-parametric tests were used for analysis of significance . *p<0 . 05 , **p<0 . 01 denotes statistically significant differences . The following are the GeneIDs ( Database: Entrez Gene ) for each gene analyzed in this manuscript , given as gene name ( official symbol GeneID: # ) : T-bet ( Tbx21 GeneID: 57765 ) ; GATA-3 ( Gata3 GeneID: 14462 ) ; RORγ-t ( Rorc GeneID: 19885 ) ; FoxP3 ( Foxp3 GeneID: 20371 ) ; IFN-γ ( Ifng GeneID: 15978 ) ; IL-17A ( Il17a GeneID: 16171 ) ; IL-10 ( Il10 GeneID: 16153 ) ; IL-4 ( Il4 GeneID: 16189 ) ; CD3γ ( Cd3g GeneID: 12502 ) , HPRT ( Hprt1 GeneID: 15452 ) .
|
The immune response to the genital tract pathogen C . trachomatis can result in a number of pathological outcomes including tubal scarring and consequently , infertility . CD4+ T helper 1 ( Th1 ) cells are critical for host protection against infection , but may also contribute to immunopathology . Apart from the Th1 cells , little is known about the role of other CD4+ T cell subsets in response to a genital tract chlamydial infection . By tracking the development of T helper cells in the genital tract using RT-PCR for distinct transcription factors associated with these subsets , we found vastly different immune responses in the upper genital tract ( UGT ) compared to the lower genital tract ( LGT ) of female mice during infection . The LGT was dominated by anti-inflammatory IL-10 production from dendritic cells ( DC ) and the non-protective Th2 subset . In contrast , the upper genital tract was populated by protective-Th1 cells . In the absence of IL-10 , though , the LGT and UGT were both dominated by Th1 cells , arguing that DC-derived IL-10 secures an anti-inflammatory privileged site in the LGT . These findings provide a break-through in our understanding of functional compartments in the genital tract immune system with potentially strong impact on vaccine development .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases/sexually",
"transmitted",
"diseases",
"immunology/reproductive",
"immunology",
"microbiology/immunity",
"to",
"infections",
"immunology/immunomodulation",
"immunology/immune",
"response",
"immunology",
"infectious",
"diseases/bacterial",
"infections",
"immunology/immunity",
"to",
"infections"
] |
2010
|
The Female Lower Genital Tract Is a Privileged Compartment with IL-10 Producing Dendritic Cells and Poor Th1 Immunity following Chlamydia trachomatis Infection
|
Pentameric ligand-gated ion channels ( pLGICs ) mediate intercellular communication at synapses through the opening of an ion pore in response to the binding of a neurotransmitter . Despite the increasing availability of high-resolution structures of pLGICs , a detailed understanding of the functional isomerization from closed to open ( gating ) and back is currently missing . Here , we provide the first atomistic description of the transition from open to closed ( un-gating ) in the glutamate-gated chloride channel ( GluCl ) from Caenorhabditis Elegans . Starting with the active-state structure solved in complex with the neurotransmitter L-glutamate and the positive allosteric modulator ( PAM ) ivermectin , we analyze the spontaneous relaxation of the channel upon removal of ivermectin by explicit solvent/membrane Molecular Dynamics ( MD ) simulations . The μs-long trajectories support the conclusion that ion-channel deactivation is mediated by two distinct quaternary transitions , i . e . a global receptor twisting followed by the radial expansion ( or blooming ) of the extracellular domain . At variance with previous models , we show that pore closing is exclusively regulated by the global twisting , which controls the position of the β1-β2 loop relative to the M2-M3 loop at the EC/TM domain interface . Additional simulations with L-glutamate restrained to the crystallographic binding mode and ivermectin removed indicate that the same twisting isomerization is regulated by agonist binding at the orthosteric site . These results provide a structural model for gating in pLGICs and suggest a plausible mechanism for the pharmacological action of PAMs in this neurotransmitter receptor family . The simulated un-gating converges to the X-ray structure of GluCl resting state both globally and locally , demonstrating the predictive character of state-of-art MD simulations .
Pentameric ligand-gated ion channels ( pLGICs ) play a central role in the intercellular communication in the brain and are involved in fundamental processes such as learning , attention , and memory [1] . They are membrane-bound oligomeric proteins that convert a chemical signal , typically the local increase in the concentration of neurotransmitter , into an ion flux through the post-synaptic membrane [2] . At rest , the ion channel is closed and binding of the neurotransmitter to the extracellular ( EC ) domain elicits a fast isomerization , which results into the opening of a transmembrane ( TM ) pore and a corresponding flux of cations ( or anions ) that diffuse at rates approaching tens of millions of ions per second . This process is commonly referred to as “gating” [3] . Prominent members of the pLGIC family in humans include excitatory receptors like the nicotinic acetylcholine receptor ( nAChR ) , which are associated with a cationic channel , and inhibitory receptors like the GABAA receptor , which are linked with an anionic channel . Signal transduction by pLGICs is allosterically regulated by ligand binding at sites that are topographically distinct from the neurotransmitter-binding or orthosteric site . The design of small molecules able to activate ( agonists ) , inhibit ( antagonists ) , or modulate ( positive or negative allosteric modulators ) the function of pLGICs is critical for the development of pharmacological strategies against a range of neurological disorders including Alzheimer’s , Parkinson’s , schizophrenia , and depression . [4] Despite the fact that pLGICs and related dysfunction have attracted significant pharmacological interest , the molecular mechanism of signal transduction remains to be elucidated . X-ray crystallography of prokaryotic homologues identified in Gloeobacter violaceus ( GLIC ) [5] and Erwinia chrysanthemi ( ELIC ) [6] provided the first high-resolution descriptions of the open and closed states of the channel . X-ray structures of the eukaryotic glutamate-gated chloride channel ( GluCl ) from Caenorhabditis elegans , which was solved in complex with the endogenous neurotransmitter L-glutamate ( L-Glu ) and the positive allosteric modulator ivermectin ( IVM ) [7] , and later with phospholipids [8] , have provided detailed information on the interaction with a variety of modulatory ligands; see Fig 1 . And , high-resolution structures of GLIC at pH7 [9] and GluCl apo [8] , which both captured a pLGIC in the absence of agonist , demonstrated that ion gating is mediated by a large conformational change of the receptor , which involves both global twisting as originally proposed based on modeling [10] and the radial expansion or “blooming” of the EC domain . Finally , the most recent structural determinations of the GABAA receptor [11] , the 5-HT3 receptor [12] , the Gly receptor [13 , 14] , and the α4β2 nicotinic receptor [15] started to illuminate the details of the signal transduction mechanism: ( i ) visualizing a state that is most consistent with desensitization [11]; ( ii ) providing an atomistic description of the regulatory intracellular ( IC ) domain [12]; and shedding light onto pLGICs activation/deactivation by agonist versus antagonist binding [13 , 14] . The significant amount of structural information available in pLGICs provides opportunities to explore gating by all-atom Molecular Dynamics ( MD ) [16] . By analyzing the spontaneous relaxation of the open-channel structure triggered by the removal of agonist , MD simulations of GLIC [17] and GluCl [18] consistently pointed to the existence of an indirect coupling between twisting and blooming , suggesting a sequence of events linking ( allosterically ) neurotransmitter unbinding to pore closing [18] . Similarly , starting with the X-ray structure of GluCl with IVM and no L-Glu bound [19] or the mouse 5-HT3R in the absence of nanobodies [20] , the transition to a “water-conducting” channel was captured by introducing the neurotransmitter ( L-Glu and serotonin , respectively ) at the orthosteric site and relaxing the complex by μs-long MD . Although in both cases the physiological significance of the starting structures remains unclear , i . e . it does not correspond to the resting state visualized by GLIC at pH7 [9] or GluCl apo [8] , these studies evidenced a striking correlation between orthosteric agonist binding and pore opening . In addition , the analysis of 5-HT3R suggested that rotameric transitions of the pore-lining residues at positions 9′ and 13′ , which significantly enlarge the ion-pore diameter with minimal variation of the protein backbone , is key to stabilize an ion-conducting state [20] . Here , we report on the first atomistic description of the functional isomerization from active to rest ( un-gating ) in a pLGICs . Starting with the open-channel structure of GluCl in complex with L-Glu and the positive allosteric modulator IVM , we analyze its structural relaxation upon removal of the IVM by μs-long MD simulations in the native lipid-membrane environment . The calculated trajectories illuminate the spontaneous isomerization to a closed-channel form that is strikingly similar to the X-ray structure of GluCl apo , thus bridging two physiological states of the channel at atomic resolution . Analysis of two independent realizations of un-gating unveils a novel mechanism of pore closing and illustrates how agonist unbinding from the orthosteric site and/or the allosteric transmembrane site regulates the transition to a non-conductive state . These results provide fundamental insights onto the allosteric mechanism and regulation of pLGICs , offering a plausible interpretation of the pharmacological action of positive allosteric modulators ( PAMs ) .
The two μs relaxations of GluCl active with IVM removed were analyzed in greater detail to collect insights onto the allosteric mechanism that couples orthosteric agonist binding to pore closing almost 60Å away . The MD relaxation of GluCl promoted by the removal of IVM showed that ion-channel deactivation is mediated by two sequential isomerizations . By projecting the 2 . 5μs relaxation ( run A ) on the twisting and blooming ( polar tilting ) reaction coordinates an interesting scenario emerges ( Fig 8 ) . The results show that when IVM and L-Glu are both bound ( red ) , the receptor is stable in a globally untwisted or straight configuration with a contracted EC domain; the twisting and blooming angles fluctuate around average values of 15 . 3° and 7 . 6° , respectively . On the other hand , when no agonist is bound ( green ) , the receptor adopts a globally twisted configuration with a radially expanded EC domain; corresponding twisting and blooming angles are 21 . 3° and 12 . 1° , respectively . Interestingly , when IVM is removed ( blue ) , the receptor does not reach the resting state immediately . Rather , it evolves to an intermediate configuration that is significantly more twisted ( τ ≈ 18° ) but still preserves a contracted EC domain ( θp ≈ 8° ) . The kinetic stability of this marginally stable state is not negligible ( i . e . 400 ns ) and is time-correlated with the presence of L-glutamate ( Fig 2 , bottom ) , whose binding to the orthosteric site hinders the full twisting isomerization ( see above ) . To explore the significance of this kinetic intermediate sampled by MD on closing , an additional simulation of GluCl active with IVM removed was carried out by introducing harmonic restraints on the L-Glu/receptor interactions to stabilize neurotransmitter binding; see Methods . The L-Glu restrained simulation ( L-Glu* ) shows that orthosteric agonist binding alone stabilizes a unique quaternary organization of the pentamer , which is more twisted ( τ of 16 . 7° ) than the one stabilized by IVM binding ( τ of 15 . 3° ) and closely resembles the kinetic intermediate captured by un-restrained MD on closing; see Fig 8 . To collect additional information , the configuration of the orthosteric site , the allosteric site and the ion pore in the L-Glu* simulation were compared with those observed in the resting-state and active-state simulations . To this aim , the separation between the Cα atoms of Ser150 ( + ) from Arg56 ( - ) in the EC domain and G281 ( + ) from L218 ( - ) in the TM domain were monitored over time; these residues sit at the subunits interface and form strong interactions with either L-Glu or IVM in the X-ray structure of the active state [1] . The results ( S2 Table ) show that in the L-Glu* simulation the orthosteric pocket is active-like and is significantly more contracted than that in the resting state or at the end of the relaxation of GluCl with IVM removed . By contrast , the allosteric site in the TM domain is more resting-like and its cavity shrinks by 1 . 2 Å relative to active . Finally , analysis of the transmembrane pore by HOLE shows that the radius at the constriction point is 1 . 81 Å in the L-Glu* simulation , which closely matches the value measured in the active state with IVM bound ( 1 . 86 Å ) . Based on these observations , we conclude that the conformation of GluCl sampled by L-Glu* or in the early stages of the MD relaxation with IVM removed is a quaternary distinct open-channel form featuring a globally more twisted architecture and a closed allosteric site , but that still preserves the exact same open-pore configuration , i . e . it is functionally equivalent to the IVM bound state .
Pentameric ligand-gated ion channels ( pLGICs ) are transmembrane protein assemblies that mediate interneuron communication by opening an ion pore in response to increased levels of neurotransmitter at synaptic terminals . Despite the recent availability of high-resolution structures from prokaryotes [5 , 6 , 9 , 25] , invertebrate [7 , 8] and vertebrate [12 , 13] eukaryotes including humans [11 , 14] , a detailed understanding of gating , i . e . the conformational transition leading to an open pore in response to agonist binding , is still missing . Only recently , a two-step asymmetric mechanism has started to emerge [1] . Here , we report on the functional isomerization of the glutamate-gated chloride channel ( GluCl ) from the active state stabilized by agonist binding at both the orthosteric ( L-glutamate ) and the allosteric transmembrane ( ivermectin ) sites to the resting state by microsecond , explicit solvent/membrane Molecular Dynamics ( MD ) . In the spirit of a previous study by us [18] , we have analyzed the spontaneous relaxation of the channel upon removal of ivermectin ( IVM ) and infer on the pore-closing mechanism . Importantly and unlike in previous reports [17 , 18] , the present simulations capture the full conformational transition to the physiological resting state at an atomic level of detail . Consistent with previous models [18 , 26] as well as X-ray crystallography of the end states [7 , 8] , the removal of IVM results in a closed-channel structure that is globally twisted and presents an expanded or bloomed configuration of the extracellular ( EC ) domain . Strikingly , the totally conserved prolines on the M2-M3 loop ( P268 ) , which are “out” in the active state , have all moved “in” at the end of the MD relaxation , and the transmembrane pore shuts at the position 9′ ( L254 ) , consistent with the X-ray structure of GluCl apo [8] . These results show that MD simulations started from the high-resolution structure of an open GluCl stabilized by ligands [7] were able to produce a closed-channel state that is both globally and locally consistent with the crystallographic result in the absence of ligands [8] . Hence , they suggest that state-of-art Molecular Dynamics may be used in some cases as a predictive tool . In addition , they provide the details of the closing isomerization , which are not directly accessible by experiments , shedding new light onto the gating/un-gating mechanism in pLGICs . The most important findings are summarized below . First , the simulation results indicate that in the absence of IVM the active state of GluCl is conformationally strained ( or tensed ) and relaxes to a globally twisted , closed-pore form that closely resembles the structure of GluCl apo [8] . Consistent with previous hypotheses [18 , 25] , they provide evidence that the closed-pore configuration is reached through an inward displacement of the M2-M3 loop at the EC/TM interface , which involves the passage of the bulky proline 268 past the tip of the β1-β2 loop . Thus , they support the conclusion that opening/closing of the channel is regulated by receptor twisting indirectly [18] and that gating/un-gating essentially corresponds to rearranging loops in the crowded environment of the EC/TM interface . However and in disagreement with previous conclusions [18] , we demonstrate that the position of the β1-β2 loop relative to the M2-M3 loop can be entirely controlled by the twisting isomerization , which appears to be the only molecular requirement for closing . In light of this , the present simulations are consistent with a mechanism for closing in which agonist unbinding is the initiating event , a global twisting isomerization follows , which shifts the β1-β2 loop away from the EC/TM interface , and the horizontal translocation of the M2-M3 loop coupled to an inward untilting of the pore-lining helices M2 completes the transition by shutting the pore at the position 9′ . Since the same global twisting transition appears to be regulated by L-Glu or IVM binding at topographically distinct sites , this analysis demonstrates that quaternary twisting alone provides an allosteric coupling for pore opening/closing , as originally proposed based on a model of α7 [10] . Second , our MD relaxations of GluCl with IVM removed reveal the existence of a previously unreported metastable state on closing . This “new” intermediate structurally corresponds to an open-channel form that is globally more twisted than GluCl with IVM bound [7] and features a closed allosteric pocket in the TM domain , but that still preserves the exact same open-pore configuration . As its kinetic stability is controlled by L-Glu unbinding , we conclude that neurotransmitter binding in the absence of allosteric modulators would stabilize a quaternary distinct open-channel form , consistent with recent reports on the Gly receptor [13] . Importantly , our analysis makes it clear that agonist binding at the orthosteric ( L-Glu ) or the allosteric transmembrane ( IVM ) sites modulate the same twisting isomerization with IVM binding promoting the transition to a super-untwisted state , which is expected to enhance the energetic barrier for closing even further . In this view , the distinct quaternary structure captured by MD in the early stages of the relaxation with IVM removed would be consistent with an active state elicited by neurotransmitter binding alone , i . e . in the absence of modulators . Since the coupling between quaternary twisting and pore-closing is indirect ( see above ) , this interpretation offers a molecular understanding of the pharmacological action of positive allosteric modulators ( PAMs ) , which would increase the open-pore probability with no effect on the ion flux by modulating the twisting isomerization . Last , our μs-long simulations of GluCl indicate that ion-channel deactivation or un-gating is composed of two distinct quaternary transitions , i . e . a global receptor twisting and the radial expansion or blooming of the EC domain , which are activated in this order to reach the physiological resting state ( S2 Video ) . However , they also suggest that receptor blooming is not strictly required for closing , which naturally questions the functional significance of the second isomerization . Our simulation analysis provides evidence that upon receptor twisting from active , the radial expansion of the EC domain that results in uncoupled EC subunits , is a stochastic event occurring on the μs timescale . Because this transition involves solvent exposure of large contact areas that are buried in the contracted form , the μs barrier for blooming probed by MD , e . g . in run B must be associated with breaking of extensive interactions at the subunit interfaces . On the other hand , the enhanced conformational variability observed in the bloomed EC domain ( S1 Video ) is likely to introduce a sizeable entropy stabilization , which must prevent the fast reverse transition to the contracted form as evidenced by the second part of the relaxation with IVM removed in run A and the room-temperature simulation of GluCl apo . Interestingly , the significant barriers for blooming/unblooming in a closed-pore receptor would be consistent with the existence of a pre-active intermediate state on gating [27–29] . Since the thermodynamic stability of this intermediate is supposed to be modulated by ligand binding at the orthosteric site [27] , the existence of a blooming isomerization with no ( direct ) consequence on pore opening/closing might be interpreted as a molecular mechanism for agonist selection . If so , agonism in pLGICs would be related to the ability of a given ligand to stabilize the globally twisted , EC-contracted , and pore-closed receptor , which was captured by MD in one of the two runs with IVM removed ( run B ) , versus the resting receptor . Although these conclusions remain speculative at this stage , the simulation results emphasize that pore-closing and un-gating , i . e . the transition to the physiological resting state , are not quite the same thing . The mechanistic scenario emerging from the simulations of GluCl prompts the comparison with previous models of gating [18–20 , 26 , 30 , 31] , which highlights similarities but also differences . Although existing models generally agree on that gating is mediated by a global isomerization involving both twisting and blooming , they do not provide a precise description of the mechanistic role of these movements , nor of their modulation by ligand binding events . Our present analysis illustrates how both twisting and blooming , and not only twisting [19] , fundamentally contribute to the functional isomerization to the physiological resting state , showing that un-gating is mediated by the sequential activation of twisting and blooming in this precise order , rather than the opposite [20] . Moreover , our interpretation offers a unifying mechanistic picture , in which the functional consequences of the more local changes at key sites ( i . e . the orthosteric site , the pore lumen , the EC/TM interface , etc . ) are subjected to the quaternary re-organization of the channel , which provide allosteric control through ligand-binding events at the subunit interfaces . In this framework , our analysis of GluCl suggests that pore closing is rate limited by the global twisting of the receptor rather than rotameric switching at the hydrophobic girdle [20] , which we have shown is irrelevant to water permeation in the open-channel form of GluCl stabilized by IVM binding ( S14 Fig ) . Similarly , it highlights that the structural rearrangements promoted by the removal of IVM , which have been described as specific to IVM binding and not related to gating [19 , 26] , are part of a more complex quaternary mechanism for gating/un-gating , which is more than the mere opening/closing of an ion pore . Remarkably and in agreement with our interpretation , the mechanistic role of the loops at the EC/TM interface for gating/un-gating in pLGICs has been recently recognized by the string method optimization of the gating pathway ( s ) in GLIC [31] . However , unlike in Ref . [31] , no evidence of causality between β-sheet expansion and pore closing was found in our simulated un-gating of GluCl ( see S15 Fig ) . In conclusion , we have reported on the spontaneous and complete pore-closing isomerization of the eukaryotic , pentameric ligand-gated ion channel GluCl as visualized by two independent 2 . 5 μs-long explicit solvent/membrane MD simulations . The availability of a time-resolved , atomistic description of the conformational transition between two physiological states corresponding to ion-channel deactivation provides new insights on the molecular mechanism and allosteric regulation of gating . These results considerably enrich our understanding of pLGICs function [3] and offer new opportunities to explore ligand modulation in this important family of neurotransmitter receptors .
A detailed description of the preparation of the active state model with and without IVM starting with the X-ray structure of GluCl with L-glutamate and ivermectin bound ( PDB 3RIF ) [7] is given in Ref . [18] . The L-Glu* simulation was started from the model built with IVM bound . Parameters for IVM were assigned using the CGenFF software [32 , 33] . An atomistic model of the resting state was prepared starting from the X-ray structure of GluCl apo ( PDB 4TNV ) [8] and following the same procedure . Three missing residues per subunits ( i . e . 103–105 ) , which form a loop inside the lumen of the channel in the EC domain were reconstructed by MODELLER [34] using the X-ray structure of GluCl with IVM bound as a template . Two intra-subunit disulphide bridges between the cysteine residues 130 and 144 , and residues 191 and 202 were built in all subunits as done for the active state . The structure of GluCl apo was analyzed by MOLPROBITY [35] and the suggested flips ( 30 total , S5 Table for a complete list ) were introduced before submitting the structure to energy minimization . The protonation state of the ionizable residues at pH 5 . 5 for consistency with the crystallization conditions [8] was assigned based on Poisson-Boltzmann calculations [36] and the multi-site titration approach [37] . Amino acids predicted in a non standard protonation state are listed in S4 Table . All-atom MD simulations of the GluCl pentamer with five L-Glu ligands bound to the orthosteric pocket were described in Ref . [18] . Here , the initial sub-μs simulations with and without the allosteric modulator ivermectin ( IVM ) were extended to 470 ns and 2 . 5μs , respectively . One independent 2 . 5μs-long simulation with IVM removed was initiated using a longer equilibration scheme ( 50 ns versus 2 ns ) to minimize the influence of the relaxation regime of the membrane environment on the conformational transition of the pLGIC . The MD simulation of GluCl apo in its explicit solvent and membrane environment was set up using the same protocol . The constructed system included one pentameric protein assembly , 42 . 034 water molecules , 324 POPC lipids , 119 Na+ , and 119 Cl− ions for a total of 197 . 056 atoms . The energetics were modeled using the all-atom CHARMM27 force field ( i . e . CHARMM22 [38] with CMAP corrections for backbone dihedrals [39] ) for the protein and CHARMM36 for the lipids [40] . The modified water model TIP3P [41] and the NaCl parameters from Roux and coworkers [42] were used for the solvent . The simulations were carried out in the isothermal-isobaric ( NPT ) ensemble using the highly scalable NAMD package [43] . The pressure was maintained constant to 1 . 01325 bar by the Berendsen barostat [44] , the temperature was controlled by the Langevin thermostat [45] at 300 K . A cutoff of 12 Å was used for the electrostatic interactions with a switch at 10 Å . The electrostatics were computed every two steps and the time step used was 2 fs . All covalent bonds involving hydrogen atoms were constrained with the SHAKE algorithm [46] . The simulation cell was allowed to fluctuate anisotropically while keeping a constant ratio between the x and y dimensions , which are parallel to the membrane plane . The molecular system was equilibrated for 2 ns ( NPT ensemble , 1 bar , 300 K ) , while the positional restraints on the heavy atoms of the protein were gradually turned off . Five observables were used to characterize the functional state of the pLGIC: the global twisting of the receptor , the tilting of the β-sandwiches in the EC domain , the configuration of the ion pore , its ion and water permeability , and the configuration of the orthosteric neurotransmitter site and the allosteric transmembrane site . All observables have been implemented in the program Wordom ( version 0 . 23-rc1 available at https://sourceforge . net/p/wordom/codehg ) to allow for efficient analysis of long MD trajectories . The global twisting ( τ ) was evaluated per subunit and defined as the angle spanned by the projections of the geometrical centers of the EC and the TM portions of each subunit on the pseudo-symmetry axis of the receptor . Geometrically , this angle measures the torsion of the EC domain relative the TM domain of the receptor around the pore axis . For the analysis , receptor twisting per snapshot was evaluated by averaging over the twist angle of its five subunits . The extracellular tilting was also evaluated per subunit and decomposed into polar ( θp ) and azimuthal ( θa ) components . These two angles were measured in the reference frame of each EC subunit with the Z-axis perpendicular to the plane of the membrane and the X-axis pointing outwards along the radial direction . By naming v → the vector defining the principal axis of the EC subunit , the polar ( radial ) tilt was measured as the angle between the Z-axis and the projection of v → on the XZ plane , whereas the azimuthal ( tangential ) tilt as the angle between the Z-axis and the projection of v → on the XZ plane . Similar to the twisting angle , the polar and azimuthal tilt used for the analysis correspond to averages over the five subunits per snapshot . The opening of the transmembrane pore was probed by measuring its radius at the constriction point ( residue 9′ , Leu 254 ) by HOLE [21] or the Cα cross section at position 9′ ( σ ) using a simple geometric definition [10]; the latter is referred to as the “ion-pore size” throughout the text . Pore dehydration and ion permeation were probed by counting the number of water molecules or ions within a cutoff distance of 2 and 5 Å , respectively , from the constriction point . The latter provides an orthogonal and not structure-based measure of pore opening . This analysis was done using the toolbox of VMD 1 . 9 . Simultaneously , water and ion permeability were analyzed by measuring the corresponding fluxes across the lipid membrane . To this aim , a transition event was defined as the translocation of a given particle ( a water molecule or an ion ) from one compartment of the simulation box to the other across the membrane . Water transitions were counted by monitoring the position of the oxygen atom over time . No distinction was made between upward or downward transitions , the total flux being the sum of the two . To avoid miscounting that result from particle diffusion through the periodic boundary along the Z direction , a layer of 5 Å was added at the top and the bottom of the simulation box . The same setup was used to monitor the ion flux . The configuration of the ligand-binding sites ( i . e . the orthosteric and the allosteric transmembrane sites ) were analyzed by monitoring one or more characteristic distances between residues selected based on their interaction with L-glutamate and ivermectin in the X-ray structure of the active state [7] . For the orthosteric site , the distance between the Cα atoms of Ser 150 ( + ) and Arg 56 ( - ) was used . For the allosteric site , following the work in Ref . [26] the distance between the Cα atoms of Gly 281 ( + ) and Leu 218 ( - ) was monitored . ( + ) and ( - ) refer to the principal and the complementary subunit , respectively . In the L-Glu* simulations , the position of L-Glu was controlled by restraining the distance between its side-chain carboxylic oxygens and the basic nitrogens of Arg56 ( - ) and its aminic nitrogen with the geometrical center of the aromatic ring of residues Tyr 200 and Tyr 151 to the crystallographic binding mode ( PDB 3RIF , see S4 Fig ) . To this aim , harmonic restraints with a force constant of 10 kcal/mol/Å2 were introduced [47] . Most of the trajectory handling and data extraction were done using Wordom [48 , 49] , VMD 1 . 9 and the python scientific packages Scipy , Numpy [50] and MDTraj [51] . All the plots were done using the python library Matplotlib 2 . 0 [52] .
|
Pentameric ligand-gated ion channels ( pLGICs ) control membrane conductance in living systems from bacteria to humans . Here , we report on μs-long , atomistic Molecular Dynamics simulations of the glutamate-gated chloride channel ( GluCl ) with an explicit treatment of the solvent and the membrane environment . The calculated trajectories provide a complete and time-resolved visualization of ion-channel deactivation in pLGICs . Strikingly , two independent simulations started with an open channel stabilized by agonist binding ( i . e . both L-glutamate and ivermectin ) converge to the X-ray structure of GluCl in the absence of ligands upon removal of the positive allosteric modulator ivermectin . These simulations visualize the gating isomerization in pLGICs with unprecedented space and time resolution and provide a plausible mechanism for the pharmacological action of positive allosteric modulators ( PAMs ) . Our analysis demonstrates in the clearest fashion the predictive power of state-of-art computer simulations and opens to the rational design of drugs targeting pLGICs’ dysfunction including Alzheimer’s , Parkinson’s , depression , or nicotine addiction .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"allosteric",
"regulation",
"medicine",
"and",
"health",
"sciences",
"chemical",
"characterization",
"molecular",
"dynamics",
"chemical",
"compounds",
"computational",
"biology",
"enzymology",
"electrophysiology",
"neuroscience",
"organic",
"compounds",
"isomerism",
"ion",
"channels",
"amino",
"acids",
"isomerization",
"enzyme",
"chemistry",
"research",
"and",
"analysis",
"methods",
"proteins",
"enzyme",
"regulation",
"chemistry",
"cyclic",
"amino",
"acids",
"proline",
"binding",
"analysis",
"biophysics",
"chlorides",
"physics",
"biochemistry",
"biochemical",
"simulations",
"stereochemistry",
"organic",
"chemistry",
"ligand-gated",
"ion",
"channels",
"physiology",
"biology",
"and",
"life",
"sciences",
"chemical",
"reactions",
"physical",
"sciences",
"computational",
"chemistry",
"neurophysiology"
] |
2017
|
Un-gating and allosteric modulation of a pentameric ligand-gated ion channel captured by molecular dynamics
|
Following antifungal treatment , Candida albicans , and other human pathogenic fungi can undergo microevolution , which leads to the emergence of drug resistance . However , the capacity for microevolutionary adaptation of fungi goes beyond the development of resistance against antifungals . Here we used an experimental microevolution approach to show that one of the central pathogenicity mechanisms of C . albicans , the yeast-to-hyphae transition , can be subject to experimental evolution . The C . albicans cph1Δ/efg1Δ mutant is nonfilamentous , as central signaling pathways linking environmental cues to hyphal formation are disrupted . We subjected this mutant to constant selection pressure in the hostile environment of the macrophage phagosome . In a comparatively short time-frame , the mutant evolved the ability to escape macrophages by filamentation . In addition , the evolved mutant exhibited hyper-virulence in a murine infection model and an altered cell wall composition compared to the cph1Δ/efg1Δ strain . Moreover , the transcriptional regulation of hyphae-associated , and other pathogenicity-related genes became re-responsive to environmental cues in the evolved strain . We went on to identify the causative missense mutation via whole genome- and transcriptome-sequencing: a single nucleotide exchange took place within SSN3 that encodes a component of the Cdk8 module of the Mediator complex , which links transcription factors with the general transcription machinery . This mutation was responsible for the reconnection of the hyphal growth program with environmental signals in the evolved strain and was sufficient to bypass Efg1/Cph1-dependent filamentation . These data demonstrate that even central transcriptional networks can be remodeled very quickly under appropriate selection pressure .
The incidence of invasive fungal infections has steadily increased within the past decades , largely because of a growing population of susceptible individuals , reflecting the progress of modern medicine in prolonging life even with severe underlying diseases and the increasing rate of immuno-deficient patients . One of the most frequently isolated fungi is Candida albicans , an ubiquitous and normally harmless commensal of the alimentary tract and mucocutaneous membranes . As an opportunistic pathogen , it can cause superficial infections like oropharyngeal candidiasis , especially in HIV patients , as well as life-threatening systemic infections with mortality rates up to 40% , even with current antifungal treatment options [1] . The transition from the commensal to a pathogenic state depends on the microbiota , the host response , and C . albicans activities , such as adhesion , secretion of hydrolases , metabolic adaptation , biofilm formation and , importantly , morphological plasticity , which includes the yeast-to-filament transition [2]–[7] . To survive and thrive in the many different niches inside the host , C . albicans must be able to adapt to changing environments and different stresses . In the short term , this occurs primarily by changes in gene expression and translation , and via post-translational modifications , but ultimately microevolutionary processes will play an important role . As a prominent example , White et al . [8] have shown that microevolution is the driving force behind the emergence of antifungal drug resistance . They demonstrated the de novo appearance of fluconazole resistance in evolving C . albicans strains in vivo [8] . Furthermore , clinical isolates generally exhibit large genetic variations , and microevolution can be observed both in vitro and in vivo [9] , [10] , indicating that this process plays an important role in host-pathogen interactions . Therefore , microevolution provides a source of variation for the adaptive response of C . albicans to challenging ( host ) environments . Different mechanisms account for the generation of new genotypic variants , including point mutations , amplification or deletion of chromosomal segments , chromosomal translocation or inversion , and whole chromosome aneuploidy . These genetic variations can affect expression of single genes or the structure of their encoded proteins as well as whole transcriptional networks via a mechanism known as transcriptional rewiring . In this process , the interaction between promoter regions and their corresponding regulators can be switched to different pairings , which in turn cause new connections to be formed between a signal and a transcriptional response [11] , [12] . Whereas many studies have explored the underlying mechanisms of drug resistance , the role that microevolution plays in host-pathogen interactions has rarely been investigated: Forche et al . [13] found that a C . albicans strain , passaged through a mouse host , responded by undergoing chromosome-level genetic variations , which were sufficient to generate new variants of C . albicans . The yeast-to-hyphae transition of C . albicans is central for pathogenicity [14] , [15] . Filamentation plays a pivotal role for adhesion to , invasion into and damage of epithelial and endothelial cells [2] , [16] , [17] . Upon internalization by macrophages , C . albicans induces host cell death by triggering pyroptosis , a form of programmed cell death [18] , [19] . However , later in the infection process the yeast-to-hyphae transition contributes to escape from the phagosome [19] , [20] . Morphology also plays a key role in host recognition [21] . Given the importance of morphology of C . albicans for pathogenicity , it is not surprising that the yeast-to-filament transition is induced by a wide range of environmental factors and conditions like high pH , host body temperature , CO2 , starvation and presence of serum , all of which act via several signaling pathways . Among them , the cAMP-dependent protein kinase A ( cAMP-PKA ) and the mitogen-activated protein kinase ( MAPK ) pathways , which target the transcription factors Efg1 and Cph1 , respectively , play a central role in hyphal formation [22] , [23] . This is demonstrated by a cph1Δ/efg1Δ double mutant , which is unable to form hyphae under almost all hyphae-inducing conditions in vitro ( except agar embedded conditions ) and which is probably the most commonly used mutant of C . albicans in a wide range of experiments [14] , [15] , [22] , [24] . Due to the central role of the yeast-to-filament transition for C . albicans virulence , we used the cph1Δ/efg1Δ double mutant as a model for evolutionary adaptation . To this end , we performed a series of co-culture passages of this mutant with macrophages . We expected that the hostile environment of the phagosome imposes a high selective pressure on the fungus favoring either intracellular adaptation or return to filamentation in order to escape . We performed phenotypic , transcriptomic and genomic analyses of the pre- and post-passaged strains to elucidate the degree of genetic plasticity of C . albicans when facing host stresses . We show that adaptation to macrophages leads to distinct phenotypic differences between the pre- and post-passaged strains with regained filamentation in the latter . As the causative mutation , we identified a heterozygous , non-synonymous single nucleotide exchange in the gene SSN3 , which encodes the cyclin-dependent kinase of a regulatory module of the Mediator complex . Our results demonstrate that the regulation of the morphological switch in C . albicans can be subject to microevolution .
To determine the ability of C . albicans to adapt to stresses inside phagocytes and to test the adaptability of the hyphal regulatory network , we first screened for mutants which are unable to escape from macrophages via filamentation response . We tested multiple C . albicans deletion strains with known defects in hyphal formation: strains lacking RAS1 , RIM101 , DFG16 , TEC1 , HGC1 , EED1 , or UME6 and the avirulent double deletion mutant lacking CPH1 and EFG1 [22] . Of these , only the cph1Δ/efg1Δ double mutant was completely unable to escape from macrophages even after 24 hours , while all other mutants still formed filaments inside the host cell and pierced the phagocyte membrane to some extent ( S1A Figure ) . Microscopy with FITC-labeled cph1Δ/efg1Δ cells revealed that these cells were viable and still able to replicate in the yeast form after ingestion by macrophages ( S1B Figure ) . Therefore , we chose the cph1Δ/efg1Δ strain for the following microevolution experiment . Cells of the murine macrophage cell line J774A . 1 were infected with the cph1Δ/efg1Δ double mutant at a macrophage-to-fungal ratio of 2∶1 and co-incubated . Every 24 hours , non-phagocytosed cells were removed and macrophages were lysed to harvest the phagocytosed cells . A defined fraction of this population was then transferred to a fresh macrophage population . After 19 passages , a significant morphological alteration became visible , as several phagocytosed cells started to form filaments . These filamenting cells became fixed in the population after additional 23 rounds of co-incubation . This morphologically distinct variant , evolutionary derived from the cph1Δ/efg1Δ mutant , was termed Evo . The absence of CPH1 and EFG1 in the Evo strain was verified by Southern blot analysis ( S1C Figure ) . To exclude temporary or epigenetic effects , the Evo strain was repassaged daily in liquid rich ( YPD ) medium without any selection pressure by host cells for 14 passages . The phenotype remained stable and no reversal was detected . To test whether the regained ability to form filaments was restricted to macrophage interactions or observed under additional hypha-inducing conditions , we analyzed the morphology of the Evo strain in the absence of host cells . In the cell culture medium DMEM with 10% serum at 37°C and 5% CO2 , clear filament formation of the Evo strain , but not the cph1Δ/efg1Δ strain , was observed ( Fig . 1 ) . Filamentous growth is associated with highly polarized ergosterol inclusion in membranes , which can be visualized by filipin staining [25] . As shown in Fig . 1 , Evo cells grown in the presence of serum exhibited intense filipin staining at the filament tips , equal to the wild type cells . Consistent with the defect in polarized growth , the cph1Δ/efg1Δ strain showed a more uniform filipin staining . Staining with calcofluor white for morphology analyses showed the expected true hyphae for the wild type and elongated yeasts for the cph1Δ/efg1Δ strain ( Fig . 1 ) . Interestingly , the Evo strain showed heterogeneous cell morphologies , i . e . a mixture of pseudohyphae with constrictions at the septa and true hyphae with parallel-sided walls ( Fig . 1 ) . The percentage of the different morphological forms was quantified using the morphological index ( MI ) [26] of individual cells after 4 and 12 hours of growth in serum ( S2A Figure ) . The MI for cph1Δ/efg1Δ was <2 . 5 at both time points , indicating yeast morphology . In contrast , most cells of the Evo strain grew as pseudohyphae ( MI 2 . 5–3 . 4 ) after 4 hours , while after 12 hours true hyphae were evident ( MI>3 . 4 ) in approx . 50% of the population . Both morphologies will be referred to as filaments . We then tested different classical hyphae-induction media for C . albicans to assess the extent of phenotype reversal to wild type morphology . In response to serum-containing YPD medium with 5% CO2 , the Evo strain initially formed filaments but switched back to yeast growth much earlier than the wild type ( S2B Figure ) . Filamentation ( mainly pseudohyphae ) also occurred in response to the amino sugar N-acetyl-D-glucosamine as sole carbon source and 5% CO2 ( S2B Figure ) . Finally , cells of the Evo strain were incubated in serum-containing water at 37°C in atmospheric air . Again , stable filamentation was induced , demonstrating that high CO2 is not absolutely necessary for filamentation of the Evo strain ( S2B Figure ) . In embedded media at 23°C ( S2C Figure . ) , deletion of EFG1 causes a hyperfilamentous phenotype [24] . Accordingly , the cph1Δ/efg1Δ strain was hyperfilamentous under these conditions . Interestingly , while cells of the Evo strain displayed an even more pronounced hyperfilamentous phenotype , it did not undergo filamentation on solid medium at 37°C , as seen in the cph1Δ/efg1Δ strain ( S2D Figure ) . In conclusion , our microevolution experiment led to the regained ability of filamentous growth in the cph1Δ/efg1Δ mutant in response to a diverse range of hyphae-inducing conditions , indicating that microevolutionary events had enabled this strain to bypass the dependency on Cph1 and Efg1 for filamentation in these media . Filamentous growth is an important contributing factor for the escape from macrophages . We therefore determined the amount of Evo cells that escaped from macrophages by piercing through their membranes after 4 h , 6 h and 8 h of co-incubation ( Fig . 2A ) . Both Evo and wild type , but not the cph1Δ/efg1Δ double mutant , were able to escape from macrophages . However , the piercing rate of the Evo strain was significantly lower than for the wild type at all time points . After 8 h of co-incubation nearly all wild type cells had escaped from the macrophages , but only about 25% of Evo cells . The delay in filamentation and the presence of pseudohyphae in the Evo strain may explain these differences . Next , we assessed the fungus' ability to invade oral epithelial cells . Invasion requires previous adhesion , and the cph1Δ/efg1Δ strain was almost entirely unable to adhere to epithelial cells ( Fig . 2B ) . Adhesion of the Evo strain was still reduced compared to the wild type , but significantly higher than for the double mutant ( Fig . 2B ) . This is reflected by the invasion capacity of the Evo strain , which was significantly lower than the wild type strain , but substantially higher than the cph1Δ/efg1Δ strain . Finally , we also investigated the potential of the Evo strain to damage macrophages and epithelial cells by measuring the release of lactate dehydrogenase ( LDH ) . After 32 hours of co-incubation , the Evo strain had damaged macrophages to the same extent as the wild type strain , and epithelial cells to a significantly higher degree than the cph1Δ/efg1Δ strain ( Fig . 2C ) . The Evo strain had thus regained abilities putatively relevant for systemic infections . Hence , the virulence of the Evo strain was tested in a murine model of hematogenously disseminated candidiasis . Survival was monitored over a period of 21 days . As predicted , mice infected with the Evo strain showed an intermediate and significantly different survival rate compared to mice infected with the wild type and cph1Δ/efg1Δ strains ( Fig . 2D ) . Histological examination of kidneys from infected animals revealed that the Evo strain retained its filamentous morphology in vivo , even though filaments formed by the Evo strain were shorter than by the wild type , and invasion into deeper layers of the kidney tissue was less pronounced ( Fig . 2D ) . In summary , the evolved changes in response to macrophages enabled the Evo strain not only to form filaments in vitro , but also in contact with host cells , which correlated with a higher virulence potential both in vitro and in vivo . Hyphal-associated virulence of C . albicans is not only due to filamentation per se , but also to the expression of hyphae-associated genes . In order to monitor the expression of typical hyphae-associated genes in the Evo strain , we measured the mRNA levels of HWP1 , ECE1 and ALS3 , all encoding hyphal cell surface proteins , and of EED1 , a gene that is associated with hyphal cell elongation [27] . An upregulation of all four genes in the Evo strain was confirmed by qRT-PCR after 1 hour of incubation in DMEM+10% FBS at 37°C and 5% CO2 ( Fig . 3A ) . HWP1 expression was similar in the Evo and WT strain , whereas ECE1 and ALS3 were higher expressed in the WT strain , and EED1 was more strongly upregulated in the Evo strain . Furthermore , we observed Als3 exposure on the surface of wild type and Evo cells by immunofluorescence , but not on the cph1Δ/efg1Δ strain ( Fig . 3B ) . This regained cell-surface exposure of the Als3 adhesin [28] is in accordance with the increased adhesion potential of the Evo strain . We were next interested if the filamentation program can be blocked by the quorum-sensing molecule farnesol . Very low concentrations ( 1 µM ) of farnesol in the medium resulted in a complete repression of filament formation in the Evo strain , whereas wild type cells still formed hyphae ( Fig . 3C ) . Consistently , farnesol treatment led to a dramatic repression of filament-associated gene expression ( Fig . 3D ) . By addition of exogenous dibutyryl-cyclic AMP ( db-cAMP ) to the farnesol-containing medium , filamentation was rescued in the Evo strain ( Fig . 3C ) . These data suggest a critical role for cAMP signaling in the filamentation process of the Evo strain . The yeast-to-filament regulatory network comprises many different transcription factors ( TFs ) . The filament-associated biofilm formation is controlled by a network formed by Bcr1 , Tec1 , Brg1 , Rob1 , Ndt80 and Efg1 [29] . Efg1 positively regulates all other TF genes in this network except ROB1 . We measured the transcription of these central TF genes at 30 min and 60 min after filament induction . As shown in Fig . 4A , we found an at least 1 . 5-fold upregulation of ROB1 and TEC1 after 30 min , and of BCR1 and BRG1 at both timepoints in the Evo strain . The wild type , however , showed only an increased expression of TEC1 at both timepoints and of BRG1 after 30 min . In contrast , most of these TF genes were down- or scarcely upregulated in the cph1Δ/efg1Δ strain ( Fig . 4A ) . Formation of wild type filaments is also regulated in part by CZF1 under certain conditions [30] . An increased expression of CZF1 , however , is not the cause for filamentation in the Evo strain . The CZF1 mRNA levels under serum induction did not greatly differ from the mRNA level in the cph1Δ/efg1Δ strain ( Fig . 4B ) . In addition , the mRNA level of UME6 , a key TF gene necessary for the maintenance of filamentation [31] , was upregulated in wild type cells at both time points but not in the cph1Δ/efg1Δ strain ( Fig . 4B ) . Interestingly , UME6 expression was more than 4-fold upregulated in the Evo strain after 60 min growth in serum-containing medium . C . albicans possesses an EFG1 homolog , EFH1 , and overexpression of this gene is known to induce pseudohyphal growth . In addition , like EFG1 , EFH1 is involved in the regulation of expression of filament-associated genes [24] . We found that EFH1 showed the strongest upregulation ( 7 . 4-fold ) among the tested TF genes in the Evo strain . However , deletion of EFH1 did not abolish filamentation of an Evo strain derivative ( Fig . 4C ) . Hence , the filamentation phenotype of the Evo strain was not linked to this TF . In summary , the Evo strain has regained most of the transcriptional hallmarks of filament production , including the upregulation of the central transcription factor genes TEC1 , BRG1 and UME6 . The few discrepancies to the wild type may partially explain the remaining differences in morphology . However , the late-phase upregulation of UME6 indicates that the filament maintenance of the Evo strain is similar to the wild type at the transcriptional level . Furthermore , the function of Efg1 was not replaced by Efh1 in the Evo strain . Our data indicated that the Evo strain regained the potential to produce hyphae , showed upregulation of transcription factor genes involved in filamentous growth and other hyphal associated genes , and regained a high virulence potential . The reduced virulence of the cph1Δ/efg1Δ strain is likely predominantly caused by the filamentation defects , however , Efg1 has also an important role in cell wall architecture [32] and the cell wall is essential for adhesion and invasive growth and thus for pathogenicity [33] . We therefore tested the Evo strain for cell wall defects by treatment with cell wall perturbing agents , i . e . congo red ( CR ) , calcofluor white ( CFW ) and sodium dodecyl sulfate ( SDS ) . As shown in Fig . 5A , the cph1Δ/efg1Δ strain was hypersensitive to all tested agents . In contrast , the Evo strain was as resistant as the wild type to CR and CFW , agents that disturb glucan and chitin architecture , respectively . The same phenotypic reversal was observed for the cell membrane disturbing agent SDS , suggesting a loose structure of the cell wall only in the cph1Δ/efg1Δ strain . These results indicate that the altered cell wall composition of the cph1Δ/efg1Δ strain was at least partially restored in the Evo strain . We therefore stained exposed mannan and β-1 , 3-glucan with fluorescently labeled concanavalin A ( ConA ) and anti-β-1 , 3-glucan antibody , respectively ( Fig . 5B ) . Quantification by FACS analysis displayed significantly reduced mannan and increased β-1 , 3-glucan signals on the surface of the cph1Δ/efg1Δ strain compared to the wild type strain . The Evo strain showed an intermediate mannan and wild type-like glucan exposure . The two MAP kinases , Cek1 and Mkc1 , become activated in wild type C . albicans upon treatment with cell wall disturbing agents [34]–[36] . After treatment with CR , both Cek1 and Mkc1 were phosphorylated in the Evo strain but not in the cph1Δ/efg1Δ strain ( Fig . 5C ) . Unusually , only phosphorylated Mkc1 could be detected in the wild type strain , which may be due to changes in the CR treatment protocol compared to previous experiments performed by another group [34] . However , these results show that the Evo strain regained the ability to phosphorylate Mkc1 and Cek1 in response to cell wall stress . To gain insight into the regulatory program of filamentation in the absence of CPH1 and EFG1 , we performed gene expression analysis by RNA sequencing under hyphae- and non-hyphae-inducing conditions . Both , the cph1Δ/efg1Δ and the Evo strain , were analyzed with a sequencing depth sufficient to cover the genome 75–350× . Expression ( RPKM≥1 ) was detected for 5 , 854 of the C . albicans open reading frames ( 94% ) , as well as for 561 nTARs ( novel transcriptionally active regions , [37] ) , 67 small nuclear RNAs and for 24 tRNAs ( see Materials and Methods for details and S2 Table for a complete list of all detected transcripts ) . Differential expression of selected genes was subsequently validated by qRT-PCR using biological replicates ( S3A Figure ) . After the transfer to filament-inducing conditions , 379 transcripts were significantly upregulated ( ≥2-fold , p<0 . 01 ) and 279 downregulated in the Evo strain . In the cph1Δ/efg1Δ strain , 255 transcripts were up- and 252 downregulated under the same condition . Within the group of upregulated transcripts , 209 genes were induced in both strains , while 46 transcripts were specifically induced in the cph1Δ/efg1Δ strain and 170 transcripts specifically in the Evo strain . 186 of the downregulated transcripts were repressed in both strains , whereas 66 and 93 transcripts were specifically repressed in the cph1Δ/efg1Δ and Evo strains , respectively ( S3B Figure ) . We investigated the expression of individual marker genes for filamentation [38] more closely ( S3C Figure ) . As expected , all eight genes of the core filamentation response ( ALS3 , ECE1 , DCK1 , HGT2 , HWP1 , IHD1 , RBT1 and orf19 . 2457 ) were significantly upregulated in the Evo strain under filament-inducing conditions . Four of these genes ( ECE1 , HWP1 , IHD1 and RBT1 ) and further filament-associated genes , like ALS1 , BRG1 and HGC1 were also upregulated in the non-filamenting cph1Δ/efg1Δ strain . Expression of filament-associated genes independent of any morphological transition has previously been described in the cph1Δ/efg1Δ mutant [38]–[40] . However , these genes were expressed at a significantly higher level in the Evo strain compared to the cph1Δ/efg1Δ strain under filament-inducing condition ( S2 DS5 Table ) . Overall , genes most highly expressed ( ≥5-fold ) in the Evo strain under filament-inducing condition are mainly hyphal-associated genes ( HWP1 , ECE1 , ALS3 , RBT1 , FRG2 , ALS1 and IHD1 ) . Furthermore , the expression of YWP1 , encoding a yeast-form cell wall protein , is downregulated in the Evo strain , while its expression did not change in the cph1Δ/efg1Δ strain . These results suggest that genes associated with C . albicans hyphae formation are also associated with filamentation of the Evo strain . Upregulation ( >1 . 5-fold , p<0 . 01; S2 DS1 and DS4 Table ) of DCK1 , LMO1 and CEK1 , which are required for filamentation under embedded conditions and for cell wall integrity [41] , was found solely in the Evo strain . This provides a possible explanation for the hyper-filamentous phenotype under embedded conditions as well as the increased resistance to cell wall perturbants compared to the double mutant ( Figs . 1+5 ) . To determine whether changes in the regulation of effector genes are reflected by an upregulation of specific TF genes , we also analyzed the expression levels of TF genes in the cph1Δ/efg1Δ and Evo strains under filament-inducing conditions in more depth ( S2 DS7 Table ) . A significantly higher expression of 21 TF genes was shared by both strains , and only five TF genes were specifically upregulated in the cph1Δ/efg1Δ strain as compared to the levels in the Evo strain . Interestingly , 17 TF genes had significantly higher expression specifically in the Evo strain and not in cph1Δ/efg1Δ , including three genes known to be important hyphal morphogenesis regulators: UME6 ( in agreement with previous qRT-PCR results ) , RIM101 and HAC1 . Eight of the higher expressed TF genes in the Evo strain have unknown biological functions . In the cph1Δ/efg1Δ strain , but not in the Evo strain , CPH2 , TEC1 and ACE2 , which encode TFs involved in hyphal growth , were significantly downregulated under filament-inducing conditions . Finally , a significantly lower expression was observed for NRG1 in the Evo strain , which codes for a repressor of hyphal development [42] . Hence , we scanned for Nrg1 binding sites ( A/C ) ( A/C/G ) C3T [43] in the putative promoter regions of genes specifically upregulated twofold in the Evo strain and detected the sequence motifs in 70% of these promoter regions ( S2 DS8 Table ) . With this , the Nrg1 binding motif is statistically overrepresented in promoters of upregulated genes ( p<0 . 01 ) when compared to promoters of all other genes . The downregulation of NRG1 in the Evo strain may therefore facilitate expression of filament-associated genes and hence filament formation . Further analyses indicated a significant upregulation of genes encoding for secreted aspartyl proteases ( SAP5 , SAP6 , SAP10 ) . In addition , significant differences in expression of genes associated with cell wall biosynthesis ( CHK1 , KRE6 , GLC3 , MP65 , ALG11 and MNT2 ) , alkalinisation ( ACH1 ) as well as of genes involved in glucose and galactose interconversion and uptake ( GAL10 , GAL1 , HGT2 , HGT4 , HGT12 and GSY1 ) were observed . In summary , our transcriptional analysis indicated that serial passage through macrophages led to substantial alterations of the global transcriptional profile . The programs and pattern we found differed clearly from the cph1Δ/efg1Δ mutant , and resembled more the well-known programs of the wild type strain . This is concomitant with and likely correlated with the regained ability of the Evo strain to induce filaments and to induce damage to host cells in vitro and in vivo . We went on to determine the genetic basis for the observed phenotypical differences . No obvious large-scale structural variations were detectable between the karyotypes of wild type , the cph1Δ/efg1Δ and Evo strains using pulsed field gel electrophoresis ( PFGE; S4A Figure ) . To detect possible loss of heterozygosity ( LOH ) events [44] , we analyzed four SNP-restriction fragment length polymorphism ( RFLP ) markers per chromosome [45] . No differences were detected between double mutant and Evo strain ( S3 DS1 Table ) . Taken together , these data show that no gross chromosomal rearrangements have occurred in the Evo strain . We re-sequenced the genomes of the Evo and the cph1Δ/efg1Δ strains to identify single nucleotide polymorphisms ( SNP ) that may have arisen during the microevolution experiment . Sequencing depth for cph1Δ/efg1Δ and Evo were 99× and 108× in average , respectively , with 98 . 8% of the C . albicans SC5314 reference genome covered in both cases . Comparison of both sequences revealed a chromosome 7 trisomy in the cph1Δ/efg1Δ strain , an aneuploidy that appears to have been lost during the evolution experiment ( S4B Figure ) . This is also reflected by a 1 . 5× higher mean transcription level of genes on chromosome 7 in the cph1Δ/efg1Δ strain ( S4C Figure ) . In addition , an amplification of URA3 on chromosome 3 was observed . URA3 was originally used as a marker to delete CPH1 and EFG1 in the cph1Δ/efg1Δ strain , and is now present in three copies in this mutant . The Evo strain contained 7–8 copies ( S4B Figure ) . A qPCR analysis on isolated gDNA supported these findings ( S4D Figure ) . PFGE and subsequent hybridization with a URA3 specific probe further revealed that all copies were located on the same chromosome ( S4E Figure ) . To exclude any possible contribution of multiple URA3 gene copies to the filamentous phenotype , the Evo strain was cured from URA3 with 5-fluoroorotic acid treatment [46] . This Evo Ura− strain was still able to filament , showing that URA3 copy number is not responsible for the filamentous phenotype ( S4F Figure ) . Additionally , after re-introduction of a single URA3 using the standard CIp10 plasmid at the RPS10 locus [47] , these strains exhibited the same adhesion , invasion and macrophage damage properties as their multi-URA3 counterparts ( S4G Figure ) . This indicates that the excessive URA3 copies do not have an influence on classical virulence properties of C . albicans . We observed a high number of SNPs in the cph1Δ/efg1Δ strain: altogether , 70 , 197 heterozygous and 3 , 156 homozygous SNPs were identified in cph1Δ/efg1Δ relative to the C . albicans SC5314 consensus reference genome ( Assembly 21 , [48] ) . Similarly , 72 , 315 heterozygous and 3 , 294 homozygous SNPs were identified in the Evo strain . These figures are consistent with those achieved when reads obtained by sequencing the genome of C . albicans SC5314 are aligned on the reference genome and reflect the high level of heterozygosity in C . albicans as well as putative sequencing errors and ambiguous positions in the reference genome ( homozygous SNPs ) . After combining these sets and filtering , only 329 putative SNPs were found to distinguish the cph1Δ/efg1Δ and Evo strains . Notably , polymorphisms at 209 of these positions are observed in the genomes of 19 clinical isolates , distributed over several C . albicans phylogenetic groups ( CdE , unpublished data ) . This suggests that they were most likely not responsible for the restoration of filamentation . Of the 120 remaining positions , 83 were in non-coding regions , 22 resulted in synonymous changes and 15 resulted in non-synonymous changes ( S3 DS2-4 Table ) . Finally , the RNA-Seq dataset was used as an additional source to detect SNPs specifically in expressed genes ( see Materials and Methods & S3 DS6&7 Table , ) : A total of 65 putative transcribed SNPs , both heterozygous and homozygous , were found in the Evo strain , of which 21 were located in non-coding regions . Inside ORFs , 26 caused a synonymous and 13 a non-synonymous nucleotide exchange . Of all 39 SNPs detected in coding regions , 24 were located in genes of the ALS gene family ( ALS2 and ALS4 ) , although these are likely false positives , as genes of the ALS family possess a very high sequence similarity and tandem repeat regions complicating read-mapping and SNP resolution [49] . Comparison of SNPs detected by RNA-Seq and Whole-Genome Sequencing revealed three SNPs shared by both detection methods . One SNP was located in a non-coding region between two uncharacterized genes ( orf19 . 351 and orf19 . 352 ) , while the other two were located inside ORFs . A SNP in ATP18 ( orf19 . 2066 . 1 ) resulted in a synonymous amino acid exchange , while the second SNP in SSN3 ( orf19 . 794 ) resulted in a heterozygous , non-synonymous Arg/Arg to Arg/Gln amino acid change . As the SNP at nucleotide position 1 , 055 in the SSN3 ORF ( Fig . 6A ) was detected in both analyses , we focused our investigation on this specific mutation . Ssn3 has been well characterized in Saccharomyces cerevisiae as an RNA polymerase II holoenzyme-associated cyclin-dependent kinase of the Mediator complex contributing to transcriptional control [50] . It was shown that Ssn3 promotes the degradation of the transcription factor Ste12 by phosphorylation and thereby regulates S . cerevisiae filamentous growth [51] . As depicted in Fig . 6B , the heterozygous Arg352Gln mutation of Ssn3 in the Evo strain is located within the activation segment of the protein kinase catalytic domain . An amino acid sequence comparison of C . albicans Ssn3 to sequences from S . cerevisiae , Cryptococcus neoformans , Mus musculus and Homo sapiens demonstrated this arginine residue to be conserved from fungi to mammals . The activation segment comprises several conserved structural features: the magnesium binding loop , the activation loop and the P+1 loop , in which the mutation occurred . While the activation loop is the site of regulatory phosphorylation in many kinases , the P+1 loop forms a pocket that recognizes the substrate protein [52] . To ascertain the impact of the SNP on filamentation induction , we selectively deleted either the mutated or the wild type SSN3 allele in the Evo strain , using the dominant selection marker SAT1 [53] . Sanger sequencing confirmed the exclusive presence of either one allele in the genome ( Fig . 7A ) . Strikingly , when incubated in DMEM with 10% serum at 37°C and 5% CO2 only the strain with the mutated allele still present ( Evo ssn3Δ/SSN3m ) was able to induce and maintain filamentation . The mutant containing only the wild type allele ( Evo SSN3/ssn3mΔ ) remained in the elongated yeast form , and thus presented the typical ancestral ( cph1Δ/efg1Δ ) phenotype ( Fig . 7A ) . In addition , only the Evo ssn3Δ/SSN3m strain could escape from macrophages by forming filaments like the wild type ( Fig . 7A ) . The damage capacity correlated with this ability to produce filaments: While Evo and Evo ssn3Δ/SSN3m strains showed the same levels of phagocyte lysis , the Evo SSN3/ssn3mΔ strain caused significantly less damage during co-incubation with macrophages . In fact , damage was indistinguishable from the original cph1Δ/efg1Δ strain ( Fig . 7B ) . In contrast , the deletion of the mutated allele had no influence on the hyphal development defect on solid medium ( S5A Figure ) and sensitivity to cell wall disturbing agents ( S5B Figure ) . To further ascertain that the SSN3 mutation alone is sufficient to allow filamentation in a cph1Δ/efg1Δ background , we created an independent cph1Δ/efg1Δ double mutant using the dominant selection marker SAT1 ( see S3 DS1 Table ) . Importantly , this cph1Δ/efg1ΔSAT1 strain contained neither the URA3 amplification nor the trisomy of chromosome 7 or other genetic alterations of the original cph1Δ/efg1Δ strain . In all our filamentation assays , this mutant behaved identical to the original cph1Δ/efg1Δ strain by not forming any hyphae ( Fig . 7A ) and hence not escaping from or damaging macrophages ( Figs . 7A & 7B ) . To isolate the effect of the mutated SSN3 , we followed several strategies with this new mutant: SSN3 overexpression strains were created of both the wild type and mutated ( SSN3m ) allele under the control of the strong ADH1 promoter ( see S1 Protocol ) . Strikingly , only the mutated allele allowed hyphae formation under inducing conditions in the cph1Δ/efg1ΔSAT1 strain ( Fig . 7A , lower left corner ) , even in the presence of the two native SSN3 alleles . Similarly , macrophage lysis was increased in the SSN3m overexpressing strain , but not under SSN3 overexpression ( Fig . 7B , right panel ) . Finally , we integrated the mutated SSN3 ( together with a SAT1 cassette ) into the SSN3 locus of cph1Δ/efg1ΔSAT1 , replacing one SSN3 allele and essentially reproducing the heterozygous situation of the Evo strain . Again , this strain behaved virtually identical to the Evo strain , both in forming hyphae ( Fig . 7A ) and in damaging macrophages ( Fig . 7B , right panel ) . In summary , these data show that a non-synonymous mutation in SSN3 that arose during our microevolution experiment is alone sufficient for regaining the ability to filament even in the absence of Efg1 and Cph1 .
Previous experimental studies on the acquisition of antifungal drug resistance and on stress-induced chromosome rearrangements have elegantly demonstrated the adaptive potential of C . albicans [44] , [54] . Here , we demonstrate – to our knowledge for the first time – that a complex trait such as the hyphal formation program of C . albicans can be subject to microevolution in the laboratory . The yeast-to-hyphae transition is of crucial importance for full C . albicans pathogenicity , which is reflected by its complex regulation [14] . Multiple overlapping as well as separate signaling pathways are activated by various environmental signals to regulate hyphae formation . Wild type hyphae are an important contributor to the fungus' ability to escape from engulfing macrophages . In contrast , the cph1Δ/efg1Δ mutant strain cannot escape by filament formation , yet is able to replicate inside macrophages and to block phagosome maturation . Therefore , we expected that the mutant strain would survive in the phagosome , albeit with reduced fitness compared to the wild type . We monitored the phenotypic changes of the cph1Δ/efg1Δ strain co-passaged with macrophages for 42 passages . On a comparatively short evolutionary timescale our experiment resulted in a strain which not only regained the ability to filament , but also re-acquired other important characteristics , like a more wild type-like cell wall structure and increased virulence . We were able to show that a minimal sequence alteration accounts for the striking phenotypic reversal to wild type-like filamentation: a single missense mutation in SSN3 . SSN3 encodes a fungal protein kinase , which phosphorylates various regulators in S . cerevisiae . Our data shows that it can become important for bypassing the requirements of Efg1 and Cph1 for filamentation in C . albicans . The in-depth characterization of the evolved strain revealed that the hyphal morphogenesis program can be induced by certain , but not all conditions which induce filamentation in the wild type strain . The fact that the Evo strain filaments in liquid , but not on solid media indicates an involvement of cAMP signaling and hence argues for a bypass of Efg1 functions rather than Cph1 [55]–[57] . This was further supported by three additional findings . First , the yeast-to-filament switch occurred in response to either serum , GlcNAc or CO2 , stimuli all known to trigger the activation of PKA signaling [23] , [58]–[60] . Second , filamentation was entirely blocked by the addition of the quorum-sensing molecule farnesol which represses both cAMP-PKA and MAPK signaling pathways [61] , [62] . The full restoration of filamentation when cAMP was added supports the involvement of the cAMP-PKA pathway . Third , the repressor of hyphae formation , Nrg1 is normally downregulated by the cAMP-PKA pathway , except in the presence of farnesol [63] . Transcriptome analysis showed NRG1 expression to be downregulated in the Evo strain , but not in the cph1Δ/efg1Δ mutant . As 70% of the upregulated genes in the Evo strain contain an Nrg1 binding site , these data emphasize the likely importance of Nrg1 levels on filamentation of the Evo strain . Given that the cph1Δ/efg1Δ mutant is strongly reduced in virulence [22] , the almost wild type-level virulence in the Evo strain in our murine model was striking . Examination of kidney sections revealed filament formation of the Evo strain in vivo . Compared to wild type filaments , these were shorter and resulted in less pronounced tissue invasion , which is likely associated with the lower overall virulence compared to the wild type . Three factors are likely to have contributed to the increased virulence of the Evo strain in the absence of Efg1 and Cph1: First , its ability to escape from macrophages like the wild type; second , its adhesion to host cells which was significantly higher than the cph1Δ/efg1Δ strain; and third , the ability to form filaments upon contact with epithelial cells , which is a prerequisite for both active penetration into and induced endocytosis by host cells [64] . Wächtler et al . [17] showed that filamentation alone is insufficient to cause damage of host cells . We therefore compared the damage capacities of the cph1Δ/efg1Δ and the Evo strains . The Evo strain exhibited a significantly increased potential to damage both macrophages and epithelial cells compared to the double mutant . The adaptation to macrophages was accompanied by differences in additional traits , such as resistance to cell wall stresses . In the cph1Δ/efg1Δ strain , the higher sensitivity to cell wall disturbing agents , as well as the modified exposure of cell wall components , likely reflect an altered cell wall organization which was restored in the Evo strain . This is supported by findings from a recent study by Zavrel et al . [32] which showed that deletion of EFG1 alone affects cell wall architecture . In our strains , these modifications of the cell wall seemed to be mediated by the kinases Mkc1 and Cek1 . Previous analyses carried out in cek1Δ and mkc1Δ mutants already indicated their direct relationship to cell wall composition and integrity [35] , [65] , [66] . By analyzing the differences in gene expression acquired during co-culture passaging with macrophages , we found that all genes belonging to the core filamentation network [38] were upregulated in the Evo strain . This suggests that during filamentation the Evo strain transcriptionally utilizes the complete filamentation program . The transcription factors Tec1 , Brg1 , Ume6 , Rim101 , Hac1 and Efh1 , which are known to be involved in regulation of filamentation [24] , [67]–[71] , were also upregulated in the Evo strain . Together with Nrg1 , they likely orchestrate filament formation in the Evo strain . For UME6 , it has been shown that its transcription is repressed by Nrg1-Tup1 and that ectopic Ume6 expression in cph1Δ/efg1Δ can rescue the filamentation defect under certain conditions [69] . For the maintenance of hyphal extension , both UME6 and EED1 are central [27] , [31] and both showed an increased expression in the evolved strain . Thus , the mechanisms of hyphal extension seems similar between Evo and wild type cells [27] . Hence , the transcriptional conditions for initiation and maintenance of filamentation , which comprise the release of repression and the upregulation of positive regulators of filamentation , are met in the Evo strain . Furthermore , a considerable number of transcripts specifically up- and downregulated in the Evo strain are both Candida-specific and uncharacterized . It is feasible , therefore , that these uncharacterized transcripts assumed a novel role specifically during filament formation in the Evo strain . This is especially true as the morphological switch is one of the best-investigated characteristics in C . albicans , and genes involved in this process are generally well studied . However , differential regulation of genes not clearly linked to the yeast-to-hyphal switch , including these genes , but also WOR1 and NAT4 ( both involved in the white-opaque switching ) and SST2 ( involved in the mating response pathway ) , could have been caused by the mutated Ssn3 kinase ( see below ) . Finally , it also should be noted that , even in the absence of filamentation , cph1Δ/efg1Δ was able to upregulate certain genes described as hyphae-associated under the condition tested here ( incubation in DMEM+10% FBS at 37°C and 5% CO2 on a plastic surface ) . This is in disagreement with previous data showing that EFG1 is required for expression of several hyphae-associated genes [23] , [72] . It is possible , however , that alternative pathway ( s ) , such as the Rim101 pH response pathway , are involved , as the cells were simultaneously exposed to diverse stimuli for filamentation . However , these genes still showed an increased induction in the Evo strain compared to the cph1Δ/efg1Δ strain , which argues for a further adaptation-induced , filament-associated change in regulation . It has been demonstrated that acquired drug resistance in C . albicans is often accompanied by aneuploidy and/or isochromosome formation [54] , [73] and that several stress conditions can enhance the rates of LOH events likely by mitotic recombination [44] . However , we did not detect any LOH events between the cph1Δ/efg1Δ and the Evo strain . The chromosome 7 trisomy was present initially in the cph1Δ/efg1Δ strain [74] and the Evo strain restored disomy by loss of one copy . The remaining gross genetic difference , an URA3 amplification in the Evo strain can be explained by an insufficient Ura3 expression from the EFG1 locus . An amplification of the gene may have increased fungal fitness during our experiment by ensuring more transcripts and hence more efficient growth . Prior studies suggested that ectopic expression of URA3 influences the phenotypes of a diverse range of mutants [75] , [76] , and duplication of a hisG-URA3-hisG cassette resulted in restored filamentation of an hwp1Δ mutant [77] . We were able to exclude these Ura3 effects as causes for the Evo strain filamentation , as the acquired filamentation phenotype was maintained after removal of the multiple URA3 copies . Furthermore , after re-introduction of a single copy of URA3 , no differences in virulence traits , like adhesion or invasion , were detectable as compared to the multi-copy strains . Overall , these data and the fact that the observed filamentation and other phenotypes persisted even after repassaging in rich ( YPD ) medium , argued for small-scale genomic alterations , rather than epigenetic changes , acquired by cph1Δ/efg1Δ cells adapting to macrophages . Comparative genome sequencing ( by WGS and RNA-Seq ) of cph1Δ/efg1Δ and Evo strains allowed us to pinpoint the microevolutionary changes in the Evo strain at the single nucleotide level . By combining the different approaches , we detected an expressed SNP in SSN3 , which resulted in an Arg-to-Gln change at a highly conserved position within the presumable protein kinase domain . This SNP and thus gain of heterozygosity was found to be central for the yeast-to-filament transition of the Evo strain . Deletion of the mutated SSN3 allele prevented the morphological switch in the Evo strain during growth under filament-inducing conditions and interaction with several types of host cells . Other phenotypes specific to the Evo strain were not affected by the deletion of the mutated SSN3 allele , suggesting that they evolved independently from filamentation . Importantly , introduction of a single mutated allele into an independent efg1Δ/cph1Δ strain fully copied the filamentation and host cell damage phenotype of the Evo strain . This strain contained neither the multiple URA3 copies nor the trisomy of chromosome 7 or any other possible genetic alterations of the original efg1Δ/cph1Δ strain . Hence , the SSN3 mutation alone bypassed the lack of the central transcription factors Cph1 and Efg1 and restored the ability to cause host cell damage in vitro , and likely to induce higher virulence in vivo . Ssn3 itself ( also referred as Srb10 or Cdk8 ) is part of the CDK ( cyclin-dependent kinase ) module ( SRB10/11 ) of the Mediator complex , which is a regulator of RNA-polymerase II ( RNAP II ) activity [78] , [79] . This CDK module phosphorylates the largest subunit of RNAP II , and Ssn3 additionally has roles in both transcriptional activation and repression in response to physiological signals , coordinating gene expression . By regulating the stability of the two important regulators , Ste12 ( ortholog of Cph1 ) and Phd1 , Ssn3 in S . cerevisiae is involved in the differentiation of yeasts into pseudohyphae under nutrient-limiting conditions [51] , [80] . Interestingly , a kinase-deficient Asp290Ala Ssn3 only weakly phosphorylates Ste12 in vitro , and the lack of phosphorylation increases its stability [51] . Moreover , the catalytic activity of Ssn3 contributes to the repression of a subset of Tup1-regulated genes [81]–[83] in S . cerevisiae . Tup1 is recruited to promoters by Nrg1 [84] , a factor which was downregulated in the Evo strain . Although the precise signaling pathway ( s ) controlling Ssn3 remain to be determined , Chang et al . [85] showed that the activity of Srb9 , another subunit of the CDK kinase module , is regulated by the PKA signaling pathway in S . cerevisiae . Based on our data it is tempting to speculate that the activation of the cAMP-PKA pathway results in activation of Ssn3 kinase activity , and the observed filament-specific transcriptional changes may thus depend on either a reduced or absent substrate recognition or on impaired substrate phosphorylation activity due to the Arg352Gln substitution . It is supposable that a loss of the substrate-specific kinase activity increases the stability of positive regulator ( s ) of filamentation by reducing their phosphorylation . Alternatively ( or in addition ) , the impaired kinase activity could lead to a derepression of genes associated with positive regulation of filamentous growth . Importantly , in this model the kinase-deficient Ssn3 remains part of the Mediator complex , and could fulfill any additional function it may have ( e . g . in the structure or recruitment of additional proteins ) . In both models , a decreased kinase activity would reduce inhibitory effects on filamentation , and hence would increase the sensitivity of the filamentation network to external stimuli . This would likely allow to bypass the need for additional Efg1 signaling . Additional genes outside of the immediate filamentation network may also be affected , as this model implies a pleiotropic effect of the Ssn3 mutation , with several transcription factors as possible clients . Thus , it seems that not the disrupted cAMP-PKA signaling pathway itself evolved in our microevolution experiment , but instead a regulatory hub for filamentation which the pathway probably targets in addition to Efg1 . In this hub , even single or few mutations seem to be able to lead to striking phenotypic alterations , as many filament-associated genes are directly or indirectly targeted . Finally , it is interesting to speculate why only one SSN3 allele was mutated , and we did not observe any LOH event to homozygosity at this locus . It seems possible that one mutated allele alone was sufficient to promote filamentation in macrophages , while the other wild type allele , still capable of full phosphorylation activity , was still required for additional functions of Ssn3 . This is somewhat supported by the observation that overexpression of the mutated SSN3 allele in a background with the native SSN3 alleles still in place was sufficient to allow hyphae formation . In our model , the mutated Ssn3 competes with the wild type Ssn3 , and overexpression allows the mutated protein to gain entry into a sufficient number of Mediator complexes . In conclusion , using the nonfilamentous mutant cph1Δ/efg1Δ , we have shown that C . albicans can rescue one of its key virulence traits , the yeast-to-hyphal switch , with a single nucleotide change when put under adequate selection pressure . A mutation in the transcriptional regulator Ssn3 adaptively rewired the transcription network to enable filamentation in response to external cues while bypassing the need for Efg1 and Cph1 . This shows an unexpected robustness of the whole filamentation system even to severe disruptions , and a high degree of adaptability . The selection scenario we used - co-incubation with macrophages - clearly reflects a condition C . albicans encounters in the host and thus might be an evolutionary pressure that can shape the infection biology of this fungus . In fact , this hypothesis is supported by another evolution experiment , which analyzed the adaptation of C . glabrata to macrophages . There , the selection pressure resulted in the appearance of a strain with pseudohyphae-like structures and increased virulence again by a single nucleotide mutation [86] . This demonstrates that during interaction with the host or host cells , significant changes in morphology and virulence are possible on a very short evolutionary time-scale .
All animal experiments were in compliance with the German animal protection law and were approved by the responsible Federal State authority ( Thüringer Landesamt für Lebensmittelsicherheit und Verbraucherschutz ) and ethics committee ( beratende Kommission nach § 15 Abs . 1 Tierschutzgesetz; permit no . 03-007/07 ) . Body surface temperature and body weight were recorded daily and animals were monitored twice a day for disease progression . Mice showing severe signs of illness ( isolation from the group , apathy , hypothermia and drastic weight loss ) were humanely sacrificed by ketamine/xylazine overdose and exsanguination . Candida albicans strains and mutants used in this study are listed in S1 Table . Strains were grown in YPD medium ( 1% peptone , 1% yeast extract , 2% glucose and optionally 2% agar ) or SD medium ( 2% dextrose , 0 . 17% yeast nitrogen base , 0 . 5% ammonium sulfate and optionally 2% agar ) at 30°C . Uridine ( 50 µg/ml ) or nourseothricin ( NAT; 100 µg/ml ) were added as required . If not stated otherwise , stationary phase cells were used in the experiments . Mutants were constructed as described in Protocol S1 . The murine peritoneal macrophage-like cell line J774A . 1 ( DSMZ ) and the human buccal carcinoma epithelial cell line TR-146 ( Cancer Research Technology ) were grown in Dulbecco's Modified Eagle's Medium ( DMEM , PAA ) supplemented with 10% FBS ( PAA ) and routinely cultured until passage 20 . Both cell lines were maintained at 37°C under 5% CO2 . J774A . 1 cells were removed from tissue-culture flasks by gentle scraping , while TR-146 cells were enzymatically harvested by Accutase ( PAA ) treatment . About 8×106 J774A . 1 macrophages were seeded into a 75 cm2 cell culture flask with DMEM supplemented with 10% FBS and 1% Penicillin/Streptomycin ( PAA ) . For the evolution experiment , macrophages were initially infected with 4×106 cells of the cph1Δ/efg1Δ strain . After that , 4×106 re-isolated C . albicans cells were transferred to a fresh macrophage culture . After 24 h of co-incubation , infected macrophages were washed ( 3× with PBS ) and lysed with 2 ml lysis buffer ( 50 mM Tris , 5 mM EDTA , 150 mM NaCl and 0 . 5% Nonidet P40 [Sigma-Aldrich] ) . The lysate was transferred to a 2 ml reaction tube and fungal cells were collected by centrifugation . The C . albicans cells were washed two times with DMEM and counted before infection of fresh macrophages . To verify the absence of EFG1 and CPH1 , Southern blot analysis was performed for the Evo strain as described previously [22] . Briefly , genomic DNA ( gDNA ) was digested with AvaII or KpnI to verify EFG1 or CPH1 deletion , respectively . DIG-labeled probes were generated ( Roche ) using genomic DNA from the strain SC5314 and primers EFG-A/EFG-B and P33/CPH-B ( S1 Table ) . A detailed description of the phenotypic analyses can be found in Protocol S1 . Fungal cells were grown in DMEM+10% FBS on glass coverslips in a 24 well microtiter plate for filipin ( Sigma ) , calcofluor white ( CFW ) and Als3 immunostaining . Flow cytometry was used to quantify mannan and β-1 , 3-glucan exposure on the surface of stationary C . albicans cells after staining with concanavalin A and anti- β-1 , 3-glucan . Piercing and invasion rate were determined by differential staining . All staining procedures are described in Protocol S1 . Epifluorescence ( Leica DM5500B , Leica DFC360 FX ) was used to detect CFW and filipin ( DAPI filter ) , Alexa Fluor 488 ( FITC filter ) and Alexa Fluor 647 ( Cy5 filter ) . Micrographs were taken with a Leica Digital Camera DFC360 FX or a Zeiss AxiCam ICc3 . Two times 105 J774A . 1 macrophages were seeded onto glass cover slips placed in 24 well microtiter plates and allowed to adhere overnight . Non-adherent macrophages were removed by washing with PBS . To monitor intracellular replication , C . albicans cells were labeled with 100 µg/ml fluorescein isothiocyanate ( FITC , Sigma-Aldrich ) in carbonate buffer ( 0 . 1 M Na2CO3 , 0 . 15 M NaCl , pH 9 . 0 ) for 30 min at 37°C and washed 3× with PBS . To quantify piercing rates , cells were washed without prior staining . Two times 105 fungal cells were added to macrophages in DMEM+10% FBS . The plates were incubated for indicated timepoints ( see figure legends ) . Cells were then washed once with PBS and fixed with 4% paraformaldehyde . Intracellular replication was detected by fluorescence microscopy after mounting the samples in ProLong Gold Antifade Reagent with DAPI ( Invitrogen ) . After co-incubation , piercing of macrophages by filaments was quantified by differential staining . The assays were performed in biological triplicates . Two times 105 TR-146 epithelial cells were seeded onto glass cover slips placed in 24 well microtiter plates and cultured for 2–3 days to 95%-100% confluency . Adherence and invasion assays were performed as previously described [17] . Briefly , to determine the adhesion rate , TR-146 monolayers were infected with 1×106 C . albicans cells . After one hour of co-incubation , non-adherent yeast cells were removed by rinsing 3× with PBS . Cells were fixed with 4% paraformaldehyde , permeabilized with 0 . 5% Triton X-100 and adherent C . albicans cells were stained with CFW for fluorescence microscopy . Invasion rates were determined by infecting TR-146 monolayers with 1×105 C . albicans cells . After incubation , cells were fixed and differentially stained for fluorescence microscopy . Both assays were repeated at least three times . Five times 104 host cells ( J774A . 1 or TR-146 ) were seeded in 96 well microtiter plates . J774A . 1 macrophages were cultured for 1 day before use , while TR-146 epithelial cells were cultured for 2 days to 95%–100% confluency . Damage of macrophages and epithelial cells was determined by measuring the release of lactate dehydrogenase ( LDH ) with the Cytotoxicity Detection Kit ( Roche Applied Science ) following 32 h of co-incubation with 5×104 C . albicans cells according to the manufacturer's protocol . The experiments were performed as previously described [17] and repeated at least three times . For survival studies the intravenous challenge model for disseminated C . albicans infection was used . Six to eight weeks old female BALB/c mice ( 18–20 g ) purchased from Charles River were used for the experiments . Mice were challenged intravenously with 5×105 C . albicans cells in 200 µl PBS via the lateral tail vein . All mice surviving to day 20 were humanely sacrificed . For histology , kidneys were collected and fixed with buffered formalin and paraffin-embedded sections were stained with Periodic acid-Schiff ( PAS ) according to standard protocols . To detect phosphorylated Mkc1 and Cek1 as well as α-tubulin , cells of an overnight culture were adjusted to an OD600 of 0 . 5 in SD medium ( control ) or SD medium supplemented with 450 µg/ml congo red , and incubated for 4 hours at 30°C . Cell disruption , protein extraction and western blot analysis using anti-phospho-p44/42 MAP kinase antibody ( Cell Signalling Technology ) and rat anti-α-tubulin antibody ( AbD Serotec ) , respectively , were performed as previously described [87] . C . albicans cells from an overnight culture were diluted to OD600 = 0 . 2 in YPD medium and grown to log-phase for 4 h at 30°C . Cells were collected by centrifugation and a zero time point sample was frozen in liquid nitrogen until RNA extraction . In addition , 1×107 cells were incubated one hour under filament-inducing conditions ( DMEM+10% FBS at 37°C and 5% CO2 in a 75 cm2 cell culture flask ) . For farnesol experiments , 10 µM farnesol was added to the medium just prior to the experiment . After incubation , medium and non-adherent cells were removed and 5 ml ice-cold PBS was added . The cells were collected by scraping and then centrifuged for 5 min at 6 , 000 g at 4°C . Cell pellets were snap frozen in liquid nitrogen . Total RNA was isolated using the Ribopure-Yeast Kit ( Ambion ) and treated with Turbo DNase ( Ambion ) . RNA quality was determined in a Bioanalyzer with an RNA 6000 Nano LabChip Kit ( Agilent Technologies ) according to the manufacturer's protocol . RNA concentration was determined with a Nanodrop ND1000 ( Peqlab ) . Copy number and expression levels of selected genes were analyzed with a my-Budget 5× EvaGreen QPCR Mix II ( Bio&Sell ) in a C1000TM Thermal Cycler ( BioRad ) using gene-specific primers ( S1 Table ) . For expression analysis , 600 ng of total RNA was reversely transcribed with the SuperScript III First-Strand Synthesis Kit ( Invitrogen ) according to the manufacturer's instructions . URA3 gene copy number was determined from 100 ng of gDNA with primers URA3-fw and URA3-re ( S1 Table ) . PCR conditions were as followed: 95°C for 15 min , 40 cycles of each 95°C for 15 s , 60°C for 40 s and 72°C for 15 s . A melting profile was generated to confirm PCR product specificity . Relative gene expression levels were determined by the 2ΔΔCt method [88] with ACT1 , EFB1 and PMA1 as internal controls . URA3 copy number was calculated with ACT1 internal control and gDNA from SC5314 ( containing two copies of URA3 ) as reference . Three independent experiments were performed . PFGE and SNP-RFLP are described in Protocol S1 . In order to use only high quality reads , trimming was performed using Btrim ( window size = 15 , average quality score = 20 ) [89] . For differential gene expression analysis high quality trimmed reads were mapped against the sequence Assembly 21 of strain SC5314 [48] using the spliced read mapper TopHat 2 . 0 . 6 [90] with the “known transcripts” ( -G option ) and uniquely mapped reads were counted using HTSeq [86] . Raw counts for each gene were loaded into R and differentially expressed genes were identified using the packages edgeR and DESeq [91] , [92] and filtered by adjusted p-values ( <0 . 01 ) and RPKM value ( ≥1 ) . Data were deposited at the Gene Expression Omnibus ( GSE56174 ) and can be found in S2 Table . Nrg1 binding sites ( A/C ) ( A/C/G ) C3T in putative promoter regions ( usually −1000 bp/+50 bp ) of all C . albicans genes were determined by SiTaR [93] allowing no mismatch . Fisher's exact test was used to determine if the Nrg1 motif-containing promoters were overrepresented in genes specifically upregulated twofold in the Evo strain , as compared to all remaining genes . For SNP calling quality trimmed reads from all samples of each strain were merged and the protocol of GATK [94] with slight changes was followed ( i . e . reads were mapped using BWA algorithm [95] , duplicates were removed and realignment around indels and base recalibration was performed ) . Next , we used bam-readcount ( www . github . com/genome/bam-readcount ) , which determines the nucleotide distribution at each single base . Heterozygous SNPs were defined as positions where 25% or more of the reads showed an alternative nucleotide . Homozygous SNPs were defined as positions where more than 90% of the reads differed from the reference . Minimum nucleotide sequence depth was 20 . Clustal Omega [96] was used for multiple sequence alignments . Genomic DNA isolated from the cph1Δ/efg1Δ and Evo strains were processed to prepare libraries for Illumina sequencing , and the TruSeq DNA Sample Prep kit ( Illumina ) was used according to the manufacturer's recommendations . DNAs were randomly fragmented by sonication to an average fragment length of 500 bp and Illumina adapters were blunt-end ligated to the fragments . The final libraries were amplified by PCR followed by sequencing on an Illumina Genome Analyzer platform ( Illumina GAII ) . 60 nt single-end reads were aligned to the C . albicans strain SC5314 reference genome [48] downloaded on 02/24/2012 using shore 5 . 0 [97] . Sequencing depth scores were computed for each 1 kb region across the genomes and for ORFs using sequencing depth data for each nucleotide located within the 1 kb region or the ORF . Sequencing depth scores were normalized based on the overall sequencing depth obtained for each genome . Single nucleotide polymorphisms were identified using shore 5 . 0 [97] at positions covered at least 30 times with a minimum quality of 25 . Homozygous SNPs were defined as positions where 90% of the reads meeting these criteria differed from the reference genome . Heterozygous SNPs were defined as positions where 20% or more of the reads showed one allele and 80% or less of the reads showed a second allele . Data were visualized and statistically analyzed using GraphPad Prism version 5 . 00 ( GraphPad Software , USA ) . Statistical analyses were performed by 1-way ANOVA ( mannan and β-1 , 3-glucan exposure ) or 2-way ANOVA ( piercing , adhesion , invasion , damage and gene expression ) followed by a Bonferroni correction . Differences in survival of mice were evaluated by Log-rank ( Mantel-Cox ) test .
|
Pathogenic microbes often evolve complex traits to adapt to their respective hosts , and this evolution is ongoing: for example , microorganisms are developing resistance to antimicrobial compounds in the clinical setting . The ability of the common human pathogenic fungus , Candida albicans , to switch from yeast to hyphal ( filamentous ) growth is considered a central virulence attribute . For example , hyphal formation allows C . albicans to escape from macrophages following phagocytosis . A well-investigated signaling network integrates different environmental cues to induce and maintain hyphal growth . In fact , deletion of two central transcription factors in this network results in a mutant that is both nonfilamentous and avirulent . We used experimental evolution to study the adaptation capability of this mutant by continuous co-incubation within macrophages . We found that this selection regime led to a relatively rapid re-connection of signaling between environmental cues and the hyphal growth program . Indeed , the evolved mutant regained the ability to filament and its virulence in vivo . This bypass of central transcription factors was based on a single nucleotide exchange in a gene encoding a component of the general transcription regulation machinery . Our results show that even a complex regulatory network , such as the transcriptional network which governs hyphal growth , can be remodeled via microevolution .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"gene",
"regulation",
"microbiology",
"mutation",
"fungi",
"model",
"organisms",
"fungal",
"pathogens",
"research",
"and",
"analysis",
"methods",
"mycology",
"medical",
"microbiology",
"gene",
"expression",
"microbial",
"pathogens",
"pathogenesis",
"yeast",
"candida",
"microevolution",
"point",
"mutation",
"host-pathogen",
"interactions",
"genetics",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"evolutionary",
"biology",
"evolutionary",
"processes",
"organisms",
"candida",
"albicans"
] |
2014
|
Microevolution of Candida albicans in Macrophages Restores Filamentation in a Nonfilamentous Mutant
|
The association between species richness and ecosystem energy availability is one of the major geographic trends in biodiversity . It is often explained in terms of energetic constraints , such that coexistence among competing species is limited in low productivity environments . However , it has proven challenging to reject alternative views , including the null hypothesis that species richness has simply had more time to accumulate in productive regions , and thus the role of energetic constraints in limiting coexistence remains largely unknown . We use the phylogenetic relationships and geographic ranges of sister species ( pairs of lineages who are each other’s closest extant relatives ) to examine the association between energy availability and coexistence across an entire vertebrate class ( Aves ) . We show that the incidence of coexistence among sister species increases with overall species richness and is elevated in more productive ecosystems , even when accounting for differences in the evolutionary time available for coexistence to occur . Our results indicate that energy availability promotes species coexistence in closely related lineages , providing a key step toward a more mechanistic understanding of the productivity–richness relationship underlying global gradients in biodiversity .
The relationship between species richness and energy availability—often described in terms of ecosystem productivity—is widespread yet poorly understood [1–5] . A link between the flux of energy through an ecosystem and the number of species it contains has long been recognised [6] , but the exact form of the relationship and its scale-dependence have traditionally been the focus of much debate [4 , 7–11] . Recent analyses have established that , when measured over large geographic and taxonomic scales ( >50 km grain size across continental or global study regions ) , species richness increases strongly with the availability of potential energy ( e . g . , net primary productivity or associated climatic proxies [12] ) , and that this relationship explains much of the spatial variation in biodiversity [13–17] . Similar patterns are repeated across a variety of life forms and regions with contrasting evolutionary histories , implying that energy availability may offer a universal explanation for Earth’s major gradients in biodiversity [1 , 18] . However , resolving the processes driving this relationship has proven far more challenging [2 , 3 , 19 , 20] . The predominant explanation for the positive relationship between energy availability and species richness ( the energy–richness relationship ) is that the amount of energy flowing through an ecosystem places a fundamental constraint on the number of species coexisting at any single point in space ( alpha-diversity ) [2 , 3 , 6] . Higher energy fluxes and the corresponding expansion in the breadth and availability of useable resources are expected to reduce the incidence of stochastic population extinction by sustaining a larger total number of individual organisms ( the “more individuals” hypothesis ) [18 , 21] and to increase the potential for local niche partitioning required for stable coexistence [22–24] . However , direct support for these hypotheses is very scant [2] , suggesting that the energy-richness relationship may have arisen through an alternative process [19 , 25] . For instance , species may have diversified more rapidly in productive environments because higher temperatures [26] , increased solar radiation [27] , larger population sizes [21] , or reduced dispersal ability [28] drive accelerated rates of speciation ( the “speciation rates” hypothesis ) [2 , 29–31] . Alternatively , if most clades originated in the humid tropics , then species richness may simply have had a longer time to accumulate in regions with high energy availability , while strongly conserved physiological constraints have prevented the expansion of species or clades into colder or drier environments with lower productivity ( the “niche conservatism” hypothesis ) [11 , 14 , 25 , 32–34] . Thus , rather than reflecting energetic constraints on community assembly , more species may coexist in productive ecosystems for largely historical reasons [2 , 35] . These contrasting historical explanations have not been ruled out by previous studies because standard methods for testing the relationship between energy availability and limits to species coexistence are largely indirect , including correlations between energy availability and assemblage biomass [36 , 37] , population density [24 , 38 , 39] , and rates of local extinction [40] . While these correlations are suggestive of a link between energy availability and community assembly dynamics , they do not conclusively establish whether coexistence is subject to energetic constraints , nor whether such constraints drive broad-scale gradients in species richness [2 , 3] . An alternative approach has been to examine the relationship between energy availability and the richness of entire assemblages , while statistically accounting for differences in regional diversity [14 , 19 , 41–44] . However , this method is problematic because it lumps together numerous unrelated species spanning a vast array of lifestyles and diets , with only minor ecological overlap , and ignores the possibility that regional species diversification may itself ultimately be regulated by local limits to coexistence [45 , 46] . Understanding the role of energetic constraints on biodiversity therefore requires an approach that focuses on species with the broadest overlap in ecological niches , while robustly accounting for processes playing out over evolutionary time . Here , we address these issues using a comparative framework to explore patterns of geographic range overlap among sister species , relatively young lineages for which energetic constraints on coexistence resulting from similarity in resource use are expected to be most pronounced [47 , 48] . Our analyses include data from 1 , 021 sister pairs , distributed across the avian phylogenetic tree and the world’s major landmasses , and representing 30% of all species for which genetic sequence data are available ( S1 Fig ) [49] . The breadth of this dataset and the evolutionary context provided by the phylogenetic relationships among lineages allows us to test the extent to which energy availability explains species coexistence in birds as well as the relationship between coexistence and contemporary gradients in species richness . We first quantify patterns of coexistence on the basis of overlap in the breeding distributions of sister species and test whether energy availability explains the probability of coexistence , both across species pairs and geographic space . Our analysis accounts for the potentially confounding effects of other abiotic variables and the phylogenetic nonindependence of species . We then take advantage of the fact that almost all speciation events in birds have involved a phase of geographic isolation [46 , 50] , using estimated divergence times to test how ecosystem productivity regulates the temporal dynamics of coexistence following divergence in allopatry [47 , 51] . We compare a model in which energy availability predicts either the initial rate at which coexistence is attained , or its temporal duration , to a null model in which variation in the incidence of coexistence is explained purely by differences in the time elapsed since speciation [47] . Throughout , we account for uncertainty in phylogenetic relationships and estimates of divergence times by fitting our models across multiple trees sampled at random from the posterior distribution [49] . In a final set of analyses , we assess the relationship between patterns of species coexistence and the total richness of avian assemblages using the geographic distributions of all 9 , 993 bird species . By combining these approaches , we aim to clarify the role of energetic constraints in both the establishment and maintenance of species coexistence as well as the contribution of these dynamics to global-scale gradients in species richness .
Across our dataset , 28% of sister species coexist , with the rest having either geographically isolated distributions or exhibiting only marginal overlap along narrow contact zones ( area of range overlap <20% of the smaller species range; see Materials and Methods ) . To examine the incidence of local coexistence and how this varies across geographic space , we quantified the percentage of sister species pairs coexisting within equal area quadrats ( resolution of 110 km x 110 km , ≈ 1° at the equator ) . Few coexisting sister species are sympatric over the entirety of their geographic range ( mean range overlap = 66% of the smaller species range and 22% of the total geographic range of both sister species combined ) , resulting in a low average incidence of coexistence among sister species at the local scale ( mean percent of sister pairs in a cell where both species are locally present = 7% ) . However , levels of local coexistence exhibit substantial variation across geographic space ( 0%–34%; Fig 1A ) . Areas containing a particularly high concentration of coexisting lineages occur throughout the wet tropics , including the eastern slope of the Andes and Amazonia , the Congo basin , New Guinea , the eastern Himalayas , and the Malay Archipelago . Beyond the tropics , additional regions of high coexistence occur along the eastern coast of Australia and throughout the northern Nearctic ( Fig 1A ) . We used sister species pairs to assess patterns of coexistence , thus avoiding the pseudoreplication introduced when analysing assemblage level patterns . The energy available for each sister pair was quantified by averaging mean annual net primary productivity ( NPP; Fig 1B ) across their combined geographic distribution ( for allopatric pairs ) or those grid cells where both species coexist ( for sympatric pairs ) ( Materials and Methods ) . We found that the probability of coexistence between sister species increases strongly with local NPP ( generalised linear model [GLM] , slope = 0 . 22 , p < 0 . 01; Fig 2A ) . This significant positive association was evident regardless of the degree of geographic range overlap used to define coexistence ( 5% to 80% ) , but became increasingly steep when considering more stringent overlap thresholds ( Fig 2B ) . NPP is the most appropriate metric for testing energetic constraints in heterotrophic organisms , but we also detected the same significant trend using actual evapotranspiration ( AET ) to quantify potential energy ( GLM , slope = 0 . 15 , p < 0 . 05 ) . In both cases , the effect of energy availability on coexistence was nonlinear , with the inclusion of a positive quadratic term leading to a substantial improvement in model fit ( GLM , slope = 0 . 21 , quadratic = 0 . 21 , p < 0 . 01 , Δ Akaike information criterion [AIC] = 8 . 2 in favour of a model including a quadratic term ) . Thus , while it has long been debated whether increased ecosystem productivity may elevate the intensity of competition , thereby reducing coexistence under conditions of high resource availability [4 , 22] , we find the opposite pattern , wherein the probability of coexistence is a positive accelerating function of energy availability . If the incidence of coexistence between sister species shows a strong phylogenetic signal , then the positive effects of energy availability could be driven by one or a few individual clades . To evaluate this possibility , we calculated the D statistic , which provides an estimate of the phylogenetic signal in a binary character relative to both a phylogenetically random distribution ( expected D = 1 ) and a Brownian motion model of trait evolution ( expected D = 0 ) [53] . We found that while the incidence of coexistence was not randomly distributed across the avian tree ( p[D = 1] < 0 . 01 ) , phylogenetic signal was low ( D = 0 . 88 ) , indicating that any tendency for coexistence to be clustered in particular clades is weak ( Materials and Methods; S1 Fig ) . Furthermore , when we included the phylogenetic covariance between species as a random effect in our models , the significant positive association between energy availability and coexistence remained ( Fig 2B and S4 Table ) . Another possibility is that the relationship between species coexistence and energy availability could arise due to covariation with other environmental factors . For instance , it has been proposed that resource specialisation leading to coexistence may be precluded in more seasonal environments in which resource abundance undergoes larger short-term temporal fluctuations [2] . Fluctuations in climate and resources may also play a role over longer time-frames , and it is variously predicted that coexistence could be promoted [54] or reduced [55] in regions covered by ice sheets during recent glacial maxima . Other physical attributes of the environment are thought to promote coexistence , including topographical heterogeneity [56] and ambient temperature [57] . To address these possibilities , we assessed the role of energy availability on coexistence relative to a suite of abiotic variables in both single and multipredictor models ( all variables were normalised to allow effect sizes to be directly compared ) . In addition to the effects of energy availability , our models highlighted a number of significant predictors of species coexistence ( Fig 3 , S1–S4 Tables ) . When assessed in isolation , we found that coexistence was significantly reduced in areas experiencing greater environmental seasonality . However , seasonality was not significant in a multipredictor model , suggesting that these effects arise from covariation with other abiotic predictors . We also detected negative relationships between coexistence and both the change in temperature since the last glacial maximum and topographic heterogeneity . In both cases , the relationship is distinctly U-shaped , with coexistence first declining and then increasing . These inconsistent slopes suggest that neither long-term climatic variability nor topographic heterogeneity have a general or direct mechanistic effect on coexistence . In particular , our results suggest that the effect of topographic heterogeneity may be sensitive to the inaccuracies in broad scale distribution maps in mountainous regions [58] . This is because , when we used more stringent overlap thresholds to define coexistence , the slope of the relationship between topographic heterogeneity and coexistence shifted to become increasingly negative and monotonic ( Fig 3 , S1–S3 Tables ) . Finally , contrary to the well-established positive relationship between species richness and temperature in birds [59 , 60] , we found that ambient temperature was unrelated to the probability of coexistence . This finding suggests that temperature may be associated with richness for historical reasons ( i . e . , tropical niche conservatism ) rather than because of any direct mechanistic link with the maintenance of species diversity . Importantly , we found that when we statistically accounted for all these additional abiotic variables , the positive effect of NPP on coexistence was strengthened ( Fig 3 , S1–S4 Tables ) . Because the vast majority of speciation events in birds are thought to require a period of geographic isolation , we can assume that coexistence between species only arises at a later stage following the expansion of species geographic ranges [61] . The median age of sister species varies geographically , being highest in Australasia and the Old World tropics and decreasing toward both high northern latitudes and the New World ( Fig 1C ) . Because of variation in the time available for range expansions [47 , 62 , 63] , these gradients in sister species age may contribute to geographical differences in the incidence of coexistence and its environmental correlates . Furthermore , rates of geographic range expansion may vary across species due to differences in intrinsic dispersal ability . For instance , recent evidence from New World birds [51] reveals that the rate at which sympatry is attained increases with species hand-wing index ( HWI ) , a measure of wing-shape correlated strongly with long-distance flight ability [64] . Robustly establishing the role of energy availability in limiting coexistence thus requires accounting for these effects . To address this , we modelled the probability of coexistence as a function of both species age and HWI in isolation and alongside energy availability and other abiotic predictors ( Materials and Methods ) . As expected , the probability of coexistence increases strongly with species age because sympatric sister pairs are , on average , >1 million years older ( 4 . 68 million years [Ma] ) than allopatric pairs ( 3 . 51 Ma ) ( Fig 3 ) . Having accounted for both species phylogenetic relatedness and abiotic factors , we found evidence that coexistence is promoted by high dispersal ability ( S4 Table ) . However , this effect was relatively weak and was not significant in our standard nonphylogenetic analysis ( Fig 3 , S1–S3 Tables ) . Overall , while our analysis highlights the important contribution of dispersal to current patterns of coexistence , this is unlikely to explain our results , because energy availability retained its independent effects even after accounting for these historical factors ( Fig 3 , S1–S4 Tables ) . We note that all these results were robust to phylogenetic uncertainty in sister species relationships and divergence times and held regardless of whether we fitted our model across either the Bayesian posterior distribution of trees or the single most credible tree ( S1–S4 Tables ) . To further explore the effects of energy availability , we extended our comparative framework to test whether energy also predicts the geographical patterns of coexistence within species pairs ( S3 Fig; Materials and Methods ) . This test is more conservative as it controls for differences in time since divergence [65 , 66] , species traits ( e . g . , dispersal ability ) , potential geographic variation in taxonomic practices and species description , and factors that may co-vary with energy availability but that are not easily accounted for in comparisons across species pairs ( e . g . , variation in rates of ecological divergence ) [47] . Within-species pair analyses confirmed that energy availability is higher in grid cells where both sister species coexist than in cells where only one sister is present ( GLM , slope = 0 . 31 , quadratic = 0 . 33 , p < 0 . 01 , n = 187 , S1 Table ) , a result that was further strengthened when we statistically accounted for other abiotic variables ( Fig 3 and S1 Table ) . These results strongly reject the possibility that patterns of coexistence arise simply due to historical factors , including differences in species age or rates of niche evolution . Looking beyond deterministic associations , it is also important to consider whether our results may be explained by stochastic effects . Although highly controversial [17 , 67] , it has been argued that broad-scale gradients in both species richness and its environmental associations may be driven by random range dynamics within a bounded geographic domain , the so-called “mid-domain effect” [68 , 69] . However , our findings are inconsistent with this stochastic model because we show that current energy availability predicts not only the broad-scale variation in range overlap , but also the particular locations of coexistence and allopatry within sister pairs ( Fig 3 and S1 Table ) . These results confirm previous evidence demonstrating that random geographic range expansion cannot explain species distributions of birds [70] . When viewed from a historical perspective , the positive relationship between coexistence and energy availability may be generated either because more productive ecosystems facilitate the initial transition to sympatry following speciation [46 , 71] , or because they prolong the duration of coexistence by reducing rates of local extinction [40] . To examine these possible mechanisms , we applied a stochastic approach to model the dynamics of coexistence between species over evolutionary time ( Materials and Methods ) [47] . In this model , we assumed that sister species are spatially isolated at the time of population divergence and then transition to a state of coexistence at a constant rate , σ . Because local extinction may result in coexisting species returning to a state of spatial segregation , we incorporate this process in our model by allowing reverse transitions back to allopatry at rate ε . Based on this model , we obtained the likelihood of observing sister species pairs in their current geographic state ( allopatric/parapatric or sympatric ) given the empirical distribution of sister species ages . We then used likelihood optimisation to estimate the transition rates to ( σ ) and from ( ε ) coexistence , from which the expected waiting time to sympatry following speciation ( 1/σ ) and the subsequent expected duration of coexistence ( 1/ε ) can be calculated ( Materials and Methods ) . By comparing AIC scores , we evaluated the relative fit of an “energy-availability dependent” ( EAD ) model , in which either σ , ε , or both are allowed to vary as a log-linear function of NPP , to a null model , in which transition rates between geographic states are equivalent across species . Using this approach , we were able to provide estimates of coexistence dynamics that are independent of any geographic gradient in sister species ages ( S1 Text and S4 Fig ) . According to our transition model , the mean expected waiting time to coexistence ( i . e . , 1/σ ) following speciation is 5 . 56 Ma , confirming our previous results highlighting the importance of time in the build-up of species coexistence ( S5 Table ) . Furthermore , in accordance with our standard statistical models , we found that an EAD model fits the data best , rejecting the null hypothesis that the probability of coexistence depends only on sister species age and , thus , the time available for range expansion ( S5 Table ) . A model in which ε decreases with ecosystem productivity ( ΔAIC = 7 . 92 in favour of the EAD model ) outperforms a model in which productivity influences σ ( ΔAIC = 5 . 72 relative to the null model ) , and there was no further improvement in model fit when combining the effects of energy availability on both σ and ε ( ΔAIC = 6 . 05 relative to the null model; S5 Table ) . The effects of energy availability on ε inferred by our model are substantial , with the mean expected duration of coexistence ( i . e . , 1/ε ) being two times longer in high ( 4 . 15 Ma; first NPP quartile ) compared to low energy environments ( 2 . 04 Ma; fourth NPP quartile ) . To examine whether these inferred coexistence dynamics are consistent with the patterns observed across sister species , we plotted how the probability of coexistence predicted by our model increases as a function of both age and NPP ( Fig 4A and 4B ) . In contrast to the poor fit of the null model ( Fig 4A ) , we find that an EAD model accounting for the effects of productivity on coexistence duration ( Fig 4B ) better captures the observed variation in the incidence of coexistence across the global gradient in both species age and energy availability ( Fig 4C and 4D; Materials and Methods ) . In particular , this model explains both the similar levels of coexistence observed among recently diverged species , regardless of local energy availability , and the apparent increase in the effect of energy availability over time as differences in the duration of coexistence are realised . Thus , our results suggest that the primary effect of energy availability is not brought about by increasing the rate at which coexistence is attained following speciation , but rather by extending the duration of coexistence , thus allowing a greater accumulation of sympatric diversity . Recent evidence based on the age of sympatric sister lineages of New World birds suggested that sympatry is attained more rapidly at high latitudes compared to the tropics [54] . This pattern has been explained in terms of the large-scale shifts in habitats following the retreat of Northern Hemisphere ice sheets , with vacant ecological niche space facilitating geographic range expansions [54] . Given the general decline in ecosystem productivity away from the equator ( Fig 1B ) , such a latitudinal increase in rates of secondary sympatry would appear to be at odds with the strong positive effect of energy availability on coexistence reported here . To resolve this , we used our analytical approach to fit a “latitudinal dependent” ( LD ) model and examined how the dynamics of coexistence varies with absolute geographic latitude . We fitted this model both globally ( n = 1 , 021 sister pairs ) and within the New World ( i . e . , the Nearctic and Neotropical realms described by Olson et al . [72]; n = 492 sister pairs ) . Our results reaffirmed a positive effect of latitude on the transition rate to sympatry in New World birds ( ΔAIC = 2 . 19 in favour of the LD model; S6 Table ) , likely contributing to the high levels of coexistence found across the northern Nearctic ( Fig 1A ) . However , when we extended our analysis globally we found no effect of latitude on the dynamics of coexistence ( ΔAIC = 1 . 63 in favour of the null model; S6 Table ) . Our results thus indicate that a latitudinal gradient in the rate of secondary sympatry is not a general trend but only a regional phenomenon , and that this does not override the positive effect of energy availability on the duration of coexistence at global scales . By demonstrating that high energy availability enhances species coexistence , our results provide a long-sought mechanistic link between current environmental conditions and broad-scale gradients in species richness . However , sister species pairs typically comprise only a fraction of all species within an assemblage , and the extent to which energetic constraints on coexistence contribute to variation in species richness remains unclear . To explore this , we examined how the incidence of coexistence across grid cells is related to the assemblage richness of the 2 , 042 sister species analysed and to the total richness of all 9 , 993 bird species ( Fig 1D ) . The results of these approaches revealed that , for a given level of coexistence , species richness is highly variable . Thus , we found a positive but relatively weak association between coexistence and both sister species richness ( Spearman’s ρ = 0 . 36 , p <0 . 001 ) and total avian richness ( Spearman’s ρ = 0 . 34 , p <0 . 001 ) ( S2 Fig ) . However , the relationship between richness and coexistence is triangular with a clear upper boundary , so that maximum species richness increases strongly with the percentage of coexisting sister species ( Fig 5 ) . In other words , while high levels of coexistence can be found regardless of local richness , species-rich locations are uniquely those supporting high levels of coexistence rather than simply a large number of allopatric lineages ( Figs 5 and S2 ) . To examine this relationship in more detail , we divided the earth’s land surface into six biogeographic realms [72] , each of which has a largely independent evolutionary history and contrasting average levels of species richness ( Fig 5 ) . We found that the positive relationship between coexistence and richness was replicated within realms . Nonetheless , there was also evidence of significant interrealm variation in both model slopes and intercepts , potentially reflecting historically driven differences in the richness of regional species pools ( Fig 5 and S7 Table ) [14] . Having accounted for this between-region variation , correlations between coexistence and richness were substantially strengthened compared to the global model ( Afrotropics: ρ = 0 . 54 , Australasia: ρ = 0 . 80 , Indomalaysia: ρ = 0 . 72 , Neotropics: ρ = 0 . 53 , Palearctic: ρ = 0 . 34 , p < 0 . 001 in all cases ) . The sole exception to this pattern was the Nearctic , where coexistence was negatively correlated with richness ( ρ = -0 . 11 ) . The consistent positive effect of sister species coexistence on assemblage richness is at odds with purely historical explanations for richness gradients based solely on differences in opportunities for species diversification [11 , 29] and also challenges the idea that high levels of coexistence among sister species simply reflects a lack of community saturation in more depauperate biotas [54 , 73] . Instead , our analysis confirms the significant role of range expansions in establishing broad-scale gradients in species richness . A robust demonstration of the fundamental relationships linking energy availability , coexistence , and assemblage richness has hitherto been lacking because of the difficulties in accounting for purely historical processes , including variation in the size of regional species pools or differences in the evolutionary time available for speciation and range expansion [14 , 25 , 74] . By focusing on interactions between avian sister species of known evolutionary age , we have shown that the probability of coexistence increases with energy availability and that this effect cannot be explained by such historical artefacts . We further demonstrate that the geographical variation in levels of coexistence among closely related lineages is strongly aligned with observed gradients in assemblage richness , supporting a mechanistic link between the global-scale increase in species richness with energy availability . The increasing application of molecular phylogenetic data to understanding macroecological patterns has often highlighted the importance of evolutionary history in the origin of broad-scale richness gradients , with many studies supporting a model in which the increase in richness with energy availability arises largely as a byproduct of accelerated rates of species diversification or the greater age and area of tropical biomes [11 , 14 , 25 , 29 , 30 , 32] . Our phylogenetic analysis of coexistence dynamics is at least partially consistent with this body of work by identifying the critical importance of evolutionary time in enabling the accumulation of sympatry following the generation of species in allopatry . However , our results also reveal that these historical effects are not sufficient on their own to explain patterns of coexistence and that the formation of species richness gradients thus depends on how energy availability determines the assembly of species into communities . While resolving the historical dynamics of coexistence based on current species distributions is challenging , our analyses suggest that energy availability has relatively little influence on the rate at which coexistence is established following speciation , and that the predominant effect of energy availability is to maintain coexistence over longer periods of time . This effect of productivity on the duration of coexistence suggests that the key factor is not an accelerated transition into sympatry as a result of weaker or more diffuse species interactions or by faster rates of character displacement [61 , 75] . Indeed , it has previously been shown that negative species interactions can constrain the establishment of coexistence following speciation in vertebrates , even in highly productive tropical regions [47 , 48 , 54] . Instead , our results are consistent with the theory that higher energy availability , acting either directly or indirectly on population dynamics [18] and niche partitioning [24] , reduces the rate of local extinction , ultimately allowing more species to be “packed” into productive tropical ecosystems [23] . The relative importance of ecological mechanisms linking energy availability to coexistence remains to be resolved . In particular , while the total biomass and numerical abundance of avian communities appears to generally increase with ecosystem productivity , implying a reduction in local extinction , whether this can account for the magnitude of observed differences in coexistence is unclear [3 , 36 , 76] . The increased vegetation complexity supported by higher energy environments seems a prime candidate for facilitating the extended coexistence of ecologically similar bird species [24 , 77] , whereas other energy-related processes may exert a similar influence by facilitating local adaptation [78] . It also seems likely that processes enhancing coexistence will interact synergistically with macroevolutionary diversification , the main alternative explanation for the accumulation of higher species richness in productive regions [14 , 25 , 29 , 32] . Indeed , according to models of adaptive radiation , a greater capacity for local coexistence is expected to elevate both rates of diversification and species-carrying capacities at the regional scale [1 , 46] . Ultimately , how greater fluxes of energy and the concomitant increases in resource availability influence species coexistence will likely depend on context , region , scale , and clade . Nevertheless , our results suggest that energetic constraints on coexistence play a fundamental role in shaping contemporary gradients in species richness , and form a vital component of any mechanistic explanation for global patterns in biodiversity .
Avian sister pairs and their estimated divergence times ( Ma ) were extracted from the Jetz et al . [14] time-calibrated phylogeny , pruned to contain only those species represented by genetic data ( n = 6 , 670 ) and based on the primary backbone topology proposed by Hackett et al . [79] . To account for phylogenetic uncertainty in sister pairings and divergence times , we repeated our analysis across 100 trees drawn at random from the posterior distribution . All reported values and results represent the median across trees ( trees can be downloaded from http://birdtree . org ) . We also conducted our analysis across a single maximum clade credibility ( MCC ) tree generated using TREEANNOTATOR ( included in BEAST v . 1 . 6 . 1 ) [80] . Results based on the MCC tree were highly concordant with our main analysis and are presented in S2 Table . From our dataset of sister species pairs ( n = 1 , 817 ) , we excluded ( i ) very recently diverged species ( <0 . 75 Ma , n = 236 ) in which ongoing introgression and ancestral polymorphism may confound reliable estimates of splitting events [81]; and ( ii ) species from poorly sampled genera ( sampling <70% , n = 638 ) in which pairs are unlikely to represent true sister species . For the remaining sister species , we quantified coexistence by combining information on overlap of species breeding distributions and broad-scale habitat occupancy . For each sister pair , we estimated the area of distributional overlap from rasterised ( 1 km resolution ) expert opinion maps of extent of occurrence ( available to view at http://mol . org ) [82] . We quantified percentage of range overlap between species according to the Szymkiewicz-Simpson coefficient [AreaOverlap/min ( AreaSister1 , AreaSister2 ) ] [61 , 83 , 84] . Sisters with abutting distributions or overlapping only marginally along narrow contact zones do not represent true coexistence and are thus sensitive to errors of commission , which increase strongly with spatial map resolution below ca . 150 km grain [85] . To ensure that our results are robust to these mapping errors , we repeated analyses using a range of different overlap thresholds ( 5% , 20% , 50% , 80% ) to define coexistence . Unless otherwise stated , results presented in the main text refer to those based on the 20% overlap threshold commonly used in studies of species sympatry [61 , 83 , 84] . Species with overlapping ( i . e . , sympatric ) breeding distributions may occupy the same ( syntopic ) or different ( allotopic ) habitats , but only where species are syntopic is energy availability expected to constrain coexistence [86] . We therefore used information on species altitudinal and habitat preferences to identify sister species pairs occupying distinct major habitat types or elevation zones ( n = 127 ) . Sister species occupying nonoverlapping elevation zones ( in accordance with polygon range data defined as <20% overlap in elevation range ) were identified using data on minimum and maximum elevation ranges compiled from a variety of published sources subjected to thorough cross-checking and updated to match current taxonomy [87–89] . Habitats were classified as forest , shrubland , bare ground , wetland , and “other” based on published information [89] . Finally , we excluded species pairs for which estimates of energy availability were either unavailable in areas of sympatry or unlikely to represent foraging areas ( e . g . , species breeding on islands but predominantly foraging at sea ) . In total , n = 1 , 021 sister pairs were included in our main analysis . To map and test the environmental predictors of coexistence , we extracted polygon range maps onto an equal area grid ( resolution of 110 km ≈ 1° at the equator ) [82] . The incidence of coexistence was mapped as the proportion of sister species pairs coexisting within each grid cell . Because sister species often coexist over only part of their geographic range , we ensured that species pairs only contribute to positive cases of coexistence where they both occur; that is , cells occupied by only a single sister contribute a value of zero to the incidence of coexistence in those cells even if those species coexist in other areas , whereas cells occupied by both species contribute a value of one . Sister species age was mapped as the median age across pairs present within each cell . Maps in the main text show the median cell values from across the posterior distribution of trees . We used the same equal area grid for extracting species distributions to sample environmental and geographical data for each sister pair , focusing on two standard global layers representing alternative metrics of energy availability [90] . First , we used consensus estimates from a large model intercomparison ( Fig 1B , [91] ) to estimate mean annual energy available to heterotrophs or NPP ( gCM-2 , 30′ resolution , reflected and square-root-transformed ) . Annual , rather than seasonal , estimates were used because , in addition to the direct effects of resource abundance on individual growth rates and population density during the breeding season , productivity is expected to influence coexistence year-round through its effects on vegetation structural complexity and resource variety [24] . Second , we used the layers of Ahn and Tateishi [92] as estimates of global variation in actual evapotranspiration . To account for possible covariance with other environmental factors , we assessed a number of putative predictors of species coexistence: mean annual temperature [57] ( temperature: data from 1961 to 1990 with 10′ original resolution , reflected and LN-transformed ) [93]; temperature and precipitation seasonality [2 , 94] ( calculated as the average three-month intra-annual variance , based on the same sources as in Jetz and Rubenstein [95] , LN-transformed ) ; topographic heterogeneity [56] ( elevation range: GTOPO30 ( USGS 1996 ) range in elevation with 30′ resolution , square-root-transformed ) ; and the difference in mean annual temperature between the Last Glacial Maximum ( LGM; 21 kya ) and the present day ( LGM temperature anomaly , an index of long-term climatic variability , LN-transformed ) [54 , 55] . Estimates of past climate were obtained from the Paleoclimate Modelling Intercomparison Project Phase II ( MIROC3 . 2 coupled ocean–atmosphere model , originally in 2 . 5′ resolution ) [96] . Although estimates of climatic variability would ideally be integrated over time , we note that the median age of sister species pairs in our dataset is 3 . 81 Ma , and , thus , the LGM temperature anomaly is likely to represent the spatial patterns of climatic variability over timescales relevant to our analysis ( see [97] ) . To determine the position of species pairs across the global gradient in each environmental variable , we calculated the mean conditions across the combined geographic range of both species ( for allopatric sister pairs ) or those cells where both species were present ( for sympatric sister pairs ) . Coexistence between species is generally restricted to particular spatial locations , and by only calculating conditions in areas where both species are present , we were able to directly match the incidence of coexistence to the local environment ( S3 Fig ) . Sister species pairs were assigned the biogeographic realm containing the majority of their combined geographic range . We quantified relative dispersal ability using the hand-wing index ( HWI ) [51 , 98] , a measure of the wing aspect-ratio that is a strong determinant of long-distance flight efficiency and both natal and migratory dispersal distances [64 , 99] . Following Claramunt [98] , the HWI was calculated as HWI=100×Kipp’sdistancewingchord where wing chord is the distance from the carpal joint ( wrist ) to the tip of the longest primary , and Kipp’s distance is the distance between the tips of the longest primary feather and the first secondary feather , both measured on the closed wing . Measurements were obtained from museum specimens , with a mean of five individuals per species ( we aimed for a minimum of two individuals of each sex ) . We excluded two sister pairs for which wing data were unavailable and used the average HWI of each sister pair ( square-root-transformed ) in our analysis . Kipp’s distances for flightless species of the genus Apteryx , which retain only a vestigial wing , could not be measured , and so these species were assigned the minimum HWI observed across the dataset . We examined the relationship between species coexistence and energy availability using two different modelling frameworks . First , we treated coexistence as a binary trait and tested its predictors in a generalised linear model with a binomial error structure . To account for the possible effects of other variables that may co-vary with energy availability , we fitted multipredictor models including each environmental variable as a main effect . We included quadratic terms to account for potential nonlinearity in the relationship between each variable and coexistence probability . To allow comparison among effect sizes , we normalised variables to unit variance . Temperature seasonality was highly collinear with energy availability ( Pearson’s r = −0 . 83 ) , and so we excluded this variable from our model . We note that results remained qualitatively unchanged when including temperature seasonality ( S3 and S4 Tables ) . In a further analysis , we focused only on coexisting species sisters ( n = 187 ) and used a paired design to compare mean environmental conditions in zones of allopatry ( i . e . , nonoverlap ) and coexistence ( i . e . , overlap ) . If coexistence is determined by phylogenetically inherited traits , then treating sister pairs as independent may overestimate the significance of any association between the incidence of coexistence and local environment conditions . To evaluate this possibility , we calculated the phylogenetic signal in coexistence using the D statistic in the R package Caper [53 , 100] . This metric compares the distribution of a binary trait ( 1 , 0 ) across the tips of the tree to two null models: ( i ) a Brownian motion model of trait evolution and ( ii ) a random trait distribution generated by shuffling species tip values . A value of D = 1 indicates a random distribution , while a value of D = 0 is the expectation under Brownian motion . The significance of the departure of the observed patterns from these two expectations is assessed through simulation ( n = 1 , 000 ) . Values of D can also extend beyond the range of 0–1 . In these cases , D > 1 indicates greater overdispersion compared to a phylogenetically random pattern , while D < 0 indicates greater conservatism than expected under Brownian motion [53] . We found that the phylogenetic signal in coexistence is low but detectable ( D = 0 . 88 ) . Thus , to ensure that our results were robust to phylogenetic nonindependence , we repeated our analysis using phylogenetic mixed models fitted using Bayesian Markov Chain Monte Carlo methods in the R package MCMCglmm [101] . We included the phylogenetic covariance between species pairs as a random effect and a probit link function . Because MCMCglmm assumes an ultrametric tree , we modelled covariance among sister pairs using the evolutionary distances at the present day rather than the time at which sister species diverged . However , this is unlikely to influence our results , because the median age of sister pairs ( 3 . 81 Ma ) is young compared to the age of the tree ( 98 Ma ) . We ran all models for 1 million iterations with a burn-in of 50 , 000 iterations and a thinning interval of 100 iterations . We set flat noninformative priors with a low degree of belief across all variables . Second , we modelled the dynamics of species coexistence over time as a constant-rate Markov process and examined the effects of energy availability on both the waiting time to coexistence and the duration of coexistence . In this model , we assumed that at the time of population divergence , sister species have allopatric distributions [46 , 50 , 102] . Given the observed time since divergence ( Ma ) and the current geographical relationship of each sister pair ( allopatric or coexisting ) , we then used maximum likelihood to estimate the per-lineage rate of transition from allopatry to coexistence ( σ ) and the reverse transition back to allopatry ( ε ) [47 , 51] . Using this approach , we compared the fit of a null model in which transition rates were equivalent across species ( nparameters = 2 ) to an “energy-availability dependent” ( EAD ) model in which NPP was included as a covariate on either σ or ε ( nparameters = 3 ) . Finally , we fitted an EAD model in which NPP had independent effects on both σ and ε ( nparameters = 4 ) . In each case , we tested for an improvement in model fit using AIC . We assessed the relative ability of these models to accurately explain patterns of coexistence by using our model parameter estimates to predict the incidence of coexistence as a continuous function of both species age and local productivity . Models were fitted in the R package msm [103] . Simulation tests demonstrating that σ and ε can be reliably inferred given present day information on sister species coexistence and divergence times are described in S1 Text and S4 Fig .
|
The increase in the number of species with the availability of energy in the environment is one of the most general but least understood patterns in global biodiversity . The finite amount of energy flowing through an ecosystem has long been suspected to place a fundamental constraint on the ability of species to subdivide ecological resources , with greater energy availability—and , thus , ecosystem productivity—in the tropics , potentially facilitating increased levels of coexistence . However , empirical support for this hypothesis has been lacking , raising the possibility that richness is higher in productive ecosystems for largely historical reasons , including the greater geological age and area of tropical biomes , which increases the period of time available for diversity to accumulate . By combining phylogenetic and geographic data from across the world’s bird species , we show that greater ecosystem productivity is associated with an increased probability of coexistence among closely related lineages and that this pattern contributes to the higher species richness in the tropics . Our results confirm that contemporary gradients in species richness are fundamentally shaped by energetic constraints on coexistence .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"taxonomy",
"biogeography",
"ecology",
"and",
"environmental",
"sciences",
"population",
"genetics",
"vertebrates",
"animals",
"animal",
"phylogenetics",
"phylogenetics",
"data",
"management",
"phylogenetic",
"analysis",
"speciation",
"molecular",
"biology",
"techniques",
"population",
"biology",
"zoology",
"research",
"and",
"analysis",
"methods",
"ecological",
"metrics",
"computer",
"and",
"information",
"sciences",
"geography",
"birds",
"ecosystems",
"phylogeography",
"evolutionary",
"systematics",
"molecular",
"biology",
"species",
"diversity",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"ecology",
"earth",
"sciences",
"genetics",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"evolutionary",
"processes",
"organisms"
] |
2016
|
Energetic Constraints on Species Coexistence in Birds
|
The lack of an effective diagnostic tool for Carrion’s disease leads to misdiagnosis , wrong treatments and perpetuation of asymptomatic carriers living in endemic areas . Conventional PCR approaches have been reported as a diagnostic technique . However , the detection limit of these techniques is not clear as well as if its usefulness in low bacteriemia cases . The aim of this study was to evaluate the detection limit of 3 PCR approaches . We determined the detection limit of 3 different PCR approaches: Bartonella-specific 16S rRNA , fla and its genes . We also evaluated the viability of dry blood spots to be used as a sample transport system . Our results show that 16S rRNA PCR is the approach with a lowest detection limit , 5 CFU/μL , and thus , the best diagnostic PCR tool studied . Dry blood spots diminish the sensitivity of the assay . From the tested PCRs , the 16S rRNA PCR-approach is the best to be used in the direct blood detection of acute cases of Carrion’s disease . However its use in samples from dry blood spots results in easier management of transport samples in rural areas , a slight decrease in the sensitivity was observed . The usefulness to detect by PCR the presence of low-bacteriemic or asymptomatic carriers is doubtful , showing the need to search for new more sensible techniques .
Bartonella bacilliformis is the etiological agent of Carrion’s disease , an overlooked illness with a lethal febrile stage and a warty phase . Its endemicity is restricted to Peru , Ecuador and Colombia , with some cases having been described in Bolivia and Chile . The transmission is by a sandfly of the genera Luyzomyia , mostly Lutzomyia verrucarum [1] . The human is the only reservoir known , and in endemic areas about 40% of asymptomatic carriers have been described [2] . In addition , Carrion’s disease-like syndromes have been related to two other Bartonella species: Bartonella rochalimae and Bartonella ancashensis [3–5] . Although its relevance remains uncertain , these species may be an explanation for the Carrion’s disease cases sporadically reported in distant areas such as Guatemala or Thailand [1] . In fact , B . rochalimae has been isolated worldwide [6 , 7] . Although the warty phase is easy to diagnose by the clinic manifestations , the initial febrile stage as well as asymptomatic carriers , are often misdiagnosed or non-diagnosed leading to perpetuation of the illness . Correct diagnosis of both acute and asymptomatic carriers is extremely important and adequate treatment is imperative to save lives . In endemic areas the diagnosis is usually made by thin blood smear and/or by clinical data . Despite having a specificity of microscopy of 96% , a low sensitivity of 36% has been described [8] . Moreover , other diseases such as malaria , dengue or tuberculosis that are also present , should be taken into account , since the first symptoms are common and may lead to misdiagnosis and erroneous treatments . All these factors are of enormous relevance since the mortality rates of Carrion’s disease are of 40–85% without treatment [9] . Furthermore , even despite receiving correct treatment the mortality rate is of 10% [10] . A more reliable method is blood culture but this is cumbersome , time-consuming and contaminations have been described in the 7–20% of the cultures [11] . Serologic tests have also been described and show a higher specificity of about 85% for both IgM ELISA and indirect fluorescence antibody test , but are difficult for routine practice [1] . Molecular diagnosis by PCR is probably the easiest way to achieve a more accurate diagnosis in endemic areas , as the equipment required is not as sophisticated or expensive , may be installed in different Health Regional Centers which may provide diagnosis to more peripheral patients , and the personnel may be easily trained in technique management . Several PCR approaches have been described in the literature in the last years [12 , 13] . However , these studies do not generally involve a large number of samples and additionally , as occurs with the remaining diagnostic tools , they are hampered by the lack of a standard case definition . In any case , PCR approaches have been showed as more effective that optical microscopic [12] , being able to diagnostic Carrion's disease patients in acute phase previously classified as negatives by thin blood smear . Nonetheless , a critical issue is the detection limit of these techniques , raising doubts about its usefulness in the detection of low-bacteraemia carriers . Dried blood spot ( DBS ) is used for the diagnosis of several infectious diseases [14 , 15] , and has been proposed for use as easy method to transfer blood samples from endemic areas to reference centers in order to carry out molecular techniques for the diagnosis of Carrion’s disease [13] . Therefore , since this illness principally affects children in rural areas , DBS may be an easy solution to both the transportation of samples and for small blood volume collection in the pediatric setting . The aim of this study was to evaluate the detection limit of three PCR approaches designed to detect B . bacilliformis , both in blood and filter papers to test their potential use for transferring samples from endemic areas to reference centers .
We used a collection strain of B . bacilliformis from the Institute Pasteur , CIP 57 . 20 ( NCTC 12136 ) . The strain was grown on blood agar ( BD , Germany ) at 28°C and 5% CO2 until confluent growth . To accurately quantify the amount of B . bacilliformis we used flow cytometry from the Citomics core facility of the Institut d’Investigacions Biomèdiques August Pi i Sunyer ( IDIBAPS ) . For this , one grown agar plate was diluted in appropriate buffer and Perfect-count microspheres were used . Serial dilutions ( 106 CFU/mL—10 CFU/mL ) were made in whole blood provided by the blood bank of the Hospital Clinic . One-hundred μL of the above mentioned bacterial serial dilutions were transfered to Whatmann 903 filter papers and let dry at least one week at room temperature to mimic the sample transfer conditions in a real scenario . DNA extraction was done from 100 μL blood and from dry blood spots with the Qiamp DNA Mini Kit ( Qiagen , Germany ) , according to the manufacturer’s instructions except that the final elution volume was 100 μL . Fragments of Bartonella-specific 16S rRNA , flagellin ( fla ) genes as well as the variable-intergenic region ( its ) were amplified . The primers used were 5’-CCTTCAGTTMGGCTGGATC-3’ and 5’-GCCYCCTTGCGGTTAGCACA-3’ for 16S rRNA [16] , 5’-ATAGAAAGAGCCTGAATACC-3’ and 5`-TGATGAAGCATGACAGTAACAC-3’ for flagellin and 5’-AGATGATGATCCCAAGCCTTCTGG-3’ [17] , and 5’-CTTCTCTTCACAATTTCAAT-3’ [18] for the amplification of variable-intergenic region . The PCRs were performed in a 25-μL total reaction volume with 500 nM forward primer , 500 nM reverse primer , 9 , 75 μL H2O and 5 μL of DNA following the conditions: 30 seconds at 94°C , 30 seconds at 55°C and 2 minutes at 72°C for 30 cycles . A 2% agarose gel stained with Sybr Safe was performed , and the results were visualized with an ImageQuant LAS4000 transiluminator ( GE Healthcare Europe GmbH , Barcelona , Spain ) . The detection limit was considered as the lowest dilution at which a positive result was obtained and considering the number of copies of each gene in the B . bacilliformis genome . All the above mentioned experiments were done in duplicate intra-assay and at two different times . The specificity was tested by doing the same PCR approaches to other member of the Bartonella genus both in vitro: Bartonella elizabethae ( strain 30455 ) , Bartonella grahamii ( strain 50771 ) , Bartonella henselae , Bartonella koehlerae ( strain 30773 ) , Bartonella tamiae ( Strain Th307 ) , and Bartonella vinsonii subsp . vinsonii ( strain 30453 ) , and in silico for the remaining 25 recognized species plus B . ancashensis . In addition other plate-grown bacteremia microorganisms such as Escherichia coli , Pseudomonas spp . , Shigella spp . , Klebsiella spp . , Haemophilus spp . , Staphylococcus aureus and Streptococcus spp . , as well as an intracellular microorganism such as Ricketsia spp . and Brucella melitensis were also tested .
When DNA was directly extracted from the blood , the detection limit was 5 CFU/μL for both the Bartonella-specific 16S rRNA and the fla genes . Meanwhile , a limit of 500 CFU/μL was obtained on amplification of the its region . In the case of DBS , the Bartonella-specific 16S rRNA PCR approach showed the lowest detection limit , which was also of 5 CFU/μL . Concerning dry blood , despite the detection limit being the same for 16S rRNA and its , the sensitivity decreased for fla when the detection limit dropped to 500 CFU/μL compared with 5 CFU/μL obtained directly from blood ( Table 1 ) . It was of note that fainter bands were always obtained with DBS . Regarding specificity , the 16S rRNA gene amplifies for all Bartonella species ( either in vivo or in silico ) but a positive result was also obtained when tested B . melitensis . The its amplification assay was specific for Bartonella spp . , and no other of the tested microorganisms had a positive PCR . Moreover , the its scheme might allow to distinguish between different Bartonella spp . by the different amplified size . The fla gene amplification was also specific for Bartonella species and differentiates between Bartonella spp . causing Carrion’s disease ( B . bacilliformis , B . rochalimae and B . ancashensis ) and the remaining Bartonella causing human disease ( Table 2 ) once no amplification was obtained or predicted for the last ones .
Carrion’s disease is an overlooked and restricted disease that affects the poorest populations living in remote rural areas , which badly communicated , without equipped laboratories , and with many other illnesses with a common symptomatology [1] . Thus , correct diagnosis of Carrion’s disease is essential , particularly since misdiagnosis is frequent [12 , 19] . PCR techniques rank among the most rapid techniques to diagnose B . bacilliformis . For this reason , the determination of the detection limit of these techniques is extremely important . For this study we have chosen three approaches , the amplification of 16S rRNA , the hypervariable intergenic transcribed spacer 16S-23S rRNA and the fla gene which codes for the flagelin protein of B . bacilliformis . The amplification of 16S rRNA has been proposed for Carrion’s disease diagnostic in Peru [12] . All tested Bartonella had an amplified product of 438 bp . Moreover , the in silico analysis showed that these primers are able to amplify all Bartonella spp . Then , this PCR approach may be also useful in other environments to detect and identify other Bartonella spp . either combining with sequencing or RFLP . The its amplification permits to differentiate between B . bacilliformis and B . ancashensis from the main pathogenic Bartonella spp [20] . In fact the its region has been used in different studies of Bartonella spp [7] . Regarding fla , this gene permit to distinguish between the three Bartonella causing Carrion’s disease: B . bacilliformis ( 940 pb ) , B . rochalimae ( 974 bp ) and B . ancashensis ( 937 bp ) from the remaining Bartonella spp . with clinical interest . However , one exception is B . clarridgeiae ( 997 bp ) . Additionally , and in silico analysis showed that B . schoenbuchensis will also results in a positive fragment of 1008 bp . Our results show that the Bartonella-specific 16S rRNA PCR seems to be the best of the techniques analyzed to detect the presence of B . bacilliformis in blood samples ( 5 CFU/μL ) since the lowest detection limit was achieved on comparison with fla and its PCRs . These results are in accordance with Angkasekwinai et al . [21] , who reported a detection limit of 1 and 10 copies/μL in a loop-mediated isothermal amplification when the detection limit was determined using bacterial genomic DNA alone or in the presence of human plasma respectively . This sensitivity might allow diagnosing the acute cases of Carrion’s disease , in which the mean percentage of infected RBCs is 61% ( ranging from 2 to 100% ) [22] . Nonetheless , the concomitant use of these PCR approaches will provide information about other Bartonella spp . infections . Filter paper may be an alternative for easy transportation of samples from endemic areas to reference laboratories but the decreasing sensitivity of the results must been taken into account which may lead to the non-detection of cases with a low bacteremia . Although the same detection limit was obtained for 16S rRNA PCR both directly from blood and filter papers , the bands were fainter in the latter . It is true that 1 week delay in the sample processing could affect the PCR by increasing the detection limit . Nonetheless , in rural settings the transfer of samples to reference centers is associated with bad communications ways , resulting in some days from sample collection to molecular determinations . None of the non-Bartonella microorganisms included in the study were positive when its or fla PCRs were performed . Nonetheless , when Brucella spp . was tested , amplification was obtained to 16S rRNA PCR . Although this is a limitation , it is need to take into account that a diagnostic should to be performed both in the adequate clinic context and in parallel with other diagnostic tools such as differential PCR for Brucella diagnostic when needed [23] . The prevalence of asymptomatic people in endemic areas has already been described by PCR being 0 . 5% [1] . However , the number of inhabitants previous exposed increases to around 40% when serologic techniques like ELISA or IFA are performed [1] . It is need to take into account that B . bacilliformis possess tropism for both erythrocytes and endothelial cells , being then present a non-blood circulating bacterial . In the chronic illness stage ( verrucuous patients ) the sensitivity of the microscopical techniques decreases from the 36% described in the acute phase to less than 10% [24] , highlighting the lower blood bacterial carriage and a possible transient bacteremia . Those facts might results in false PCR-negative when the technique is applied in the detection of both verrucous patients and asymptomatic carriers . It is important to remark that in the last years 2 more sensitive PCR techniques have been described in the literature: qPCR [13] in which 24 . 6% of DBS samples are positive , as well as a loop-mediated isothermal amplification [21] that achieves good results on analysing Lutzomyia samples . However , qPCR requires the expertise of trained personnel and is more expensive and difficult to be implemented . Meanwhile the usefulness of loop-mediated isothermal amplification remains to be validated to detect the presence of B . bacilliformis in human clinical samples . Enrichment of the sample before conventional PCR has been proposed to increase the positivity by 55% when compared with the original blood samples [25] . However , this enrichment technique results in a 14-days delay in sample processing thereby making it unaffordable for diagnostic purposes . To conclude , here we show that 16S rRNA PCR have low cfu detection limit and should be used with special attention to test samples from individuals with clinical suspicion of Carrion’s disease since the applicability to detect healthy carriers is not clear . The use of DBS could facilitate the transfer of samples from rural endemic areas to health facilities , despite the possibility of a small decrease in positivity . It is critical to develop rapid , sensitive and specific techniques which may be applied in endemic rural areas to avoid misdiagnosis and to facilitate the detection of asymptomatic carriers and thereby the decrease the number of B . bacilliformis cases .
|
Carrion’s disease is an endemic illness in the Andean valleys of Peru that achieves high mortality rates in the absence of antibiotic treatment . There are three clinical manifestations , febrile acute patients , chronic patients as well as asymptomatic carriers . No effective diagnostic tool exists nowadays leading to misdiagnosis and the perpetuation of the illness . The objective of this study was to determine the detection limit of three PCR approaches both from blood samples as well as from filter papers . Furthermore , the specificity was also accessed . We found that the best PCR approach studied was the amplification of the 16S rRNA from blood samples with a detection limit of 5 CFU/μL , the same when using dry blood in filter paper , although the obtained bands were not so evident . Present results highlight the need to develop more sensitive techniques able to be used both in rural areas and in the detection of asymptomatic carriers .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"laboratory",
"equipment",
"engineering",
"and",
"technology",
"pathogens",
"microbiology",
"organisms",
"bacterial",
"diseases",
"non-coding",
"rna",
"filter",
"paper",
"molecular",
"biology",
"techniques",
"cellular",
"structures",
"and",
"organelles",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"artificial",
"gene",
"amplification",
"and",
"extension",
"medical",
"microbiology",
"microbial",
"pathogens",
"molecular",
"biology",
"hematology",
"ribosomes",
"bartonella",
"biochemistry",
"rna",
"diagnostic",
"medicine",
"ribosomal",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"anatomy",
"blood",
"equipment",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"carrion's",
"disease",
"gene",
"amplification",
"polymerase",
"chain",
"reaction"
] |
2016
|
Evaluation of PCR Approaches for Detection of Bartonella bacilliformis in Blood Samples
|
A common biological pathway reconstruction approach—as implemented by many automatic biological pathway services ( such as the KAAS and RAST servers ) and the functional annotation of metagenomic sequences—starts with the identification of protein functions or families ( e . g . , KO families for the KEGG database and the FIG families for the SEED database ) in the query sequences , followed by a direct mapping of the identified protein families onto pathways . Given a predicted patchwork of individual biochemical steps , some metric must be applied in deciding what pathways actually exist in the genome or metagenome represented by the sequences . Commonly , and straightforwardly , a complete biological pathway can be identified in a dataset if at least one of the steps associated with the pathway is found . We report , however , that this naïve mapping approach leads to an inflated estimate of biological pathways , and thus overestimates the functional diversity of the sample from which the DNA sequences are derived . We developed a parsimony approach , called MinPath ( Minimal set of Pathways ) , for biological pathway reconstructions using protein family predictions , which yields a more conservative , yet more faithful , estimation of the biological pathways for a query dataset . MinPath identified far fewer pathways for the genomes collected in the KEGG database—as compared to the naïve mapping approach—eliminating some obviously spurious pathway annotations . Results from applying MinPath to several metagenomes indicate that the common methods used for metagenome annotation may significantly overestimate the biological pathways encoded by microbial communities .
Microbial whole genome sequencing has become a routine practice in recent years , because of the rapid advances of DNA sequencing technologies [1] . One of the first analyses that biologists attempt , once they obtain a complete genome sequence , is to reconstruct the biological pathways encoded by the organism , which is usually accomplished in silico by mapping the protein coding genes onto reference pathway collections , such as KEGG [2] or SEED [3] , based on their homology to reference genes with previously characterized functions . For example , KAAS , the pathway annotation system based on the KEGG database [4] , first annotates K numbers ( each K number represents an ortholog group of genes , and is directly linked to an object ( a biochemical step ) in the KEGG pathway map ) , and then reconstructs pathways based on the assigned K numbers . Similarly , the RAST server ( and MG-RAST ) first annotates FIG families and then maps the identified FIG families onto the SEED subsystems [5] , [6] . These automatic methods are promising for the analysis of most genomes , although they may leave “holes” in the reconstructed pathways , due to either missing genes ( i . e . the genes are non-homologous to reference genes of the same specific functions , and thus cannot be identified by a homology-based method , or were simply not annotated as ORFs by annotation pipelines ) [7] , or alternative and novel pathways ( i . e . the target organism adopts variant pathways , which are different from the reference pathway , to accommodate a specific niche or lifestyle ) [8] . After all , many bacterial genomes have fewer than 60% of their genes assigned to a proposed function [9] , [10] . We note that pathway reconstruction is essential for understanding the biological functions that a newly sequenced genome encodes . For instance , in a recently published report , the coupling of N2 fixation to cellulolysis was revealed within protist cells in the termite gut , based solely on the in silico pathway reconstruction of the complete genome sequence of a bacterial endosymbiont [11] . Moreover , pathway reconstruction based on some new high throughput techniques must provide conclusions from explicitly incomplete information , which poses fresh challenges . For example , in a typical proteomics experiment , the proteins represent a particular biological sample collected under a specific physiological condition or from a specific tissue ( e . g . from yeast cells after the heat shock ) , which are in high enough abundance to be identified by tandem mass spectrometry [12] , [13] . Based on these data , one may ask , what biological pathways were activated ( or suppressed ) under the physiological condition ? A similar , but more complicated case is pathway analysis of metagenomic data , to characterize the aggregate metabolic processes of microbial communities in a given environment [14] . Metagenomic profiling data can be viewed as a sampling of the genomic sequences from many kinds of microbes living in a specific environment . Again , the incompleteness of the data makes it difficult to reconstruct the entire pathways encoded by a metagenome . Nevertheless , it is becoming routine to “reconstruct” pathways for proteomic [15] and metagenomic data [16] , [17] , by best similarity matches ( often derived from BLAST searches ) : a pathway is inferred to be absent or present in a dataset if highly confident homolog protein hits identify one or more of the protein functions associated with the pathway in other organisms . In addition to the problems that arise from incomplete data , existing methods of pathway reconstruction or inference may over-estimate the number of pathways because of redundancy in the protein-pathway , at four levels . First , different pathways may share the same biological functions . The partition of pathways ( as the entire cellular network is partitioned into several hundreds of biological pathway entities in KEGG database ) is extremely important for understanding of biological processes , even though there is only a single large biological network within any cell and all pathways are to some extent connected [18] . It is not surprising that many pathways defined in the pathway databases are overlapping . Second , some proteins carry out multiple biological functions [19] , e . g . through different protein domains , active sites , or substrate specificities . Third , neither organisms nor communities are closed boxes , and the products or intermediates of pathways may be exogenously supplied . Finally , homology-based protein searching may map one protein to multiple homologous proteins with different biological functions ( i . e . paralogous proteins ) . In summary , it cannot be safely concluded that a pathway is present , even if one or more proteins are mapped to it . Even for single complete genomes , pathway reconstruction does not always give a clear picture of the biological functions in an organism , and human curation and experimental verification is often needed [20] , [21] . We illustrate this by a rather extreme example found in the pathway analysis of the human genome . The KEGG pathway annotation of the human genome includes the reductive carboxylate cycle , with proteins annotated to 6 steps in this pathway ( http://www . genome . jp/kegg-bin/show_organism ? menu_type=pathway_maps&org=hsa ) ( as of July 2nd , 2009 ) . The Calvin cycle is the most common method of carbon fixation , while the reductive carboxylate cycle is an alternative carbon fixation pathway , currently found only in certain autotrophic microorganisms . In fact , the reductive carboxylate cycle is essentially the reverse of the Krebs cycle ( citric acid or tricarboxylic acid cycle ) , the final common pathway in aerobic metabolism for the oxidation of carbohydrates , fatty acids and amino acids , so they share reactions and functional roles . For this reason , the proteins responsible for the normal function of the Krebs cycle can be mistakenly taken as evidence for the existence of a reductive carboxylate cycle in the human genome . Here we propose a pathway reconstruction/inference method in which we do not attempt to reconstruct entire pathways from a given set of protein sequences ( e . g . identified in a proteomics experiment , or encoded by the sequences sampled in a metagenomic project ) , but to determine the minimal set of biological pathways that must exist in the biological system to explain the input protein sequences sampled from it . In this context , we note pathway inference might be a more suitable terminology than pathway reconstruction . However , considering that pathway inference has been used in a different context to infer networks or pathways from gene express data [22] , and pathway reconstruction is commonly used in the field , we use both pathway inference and pathway reconstruction in this paper . To address the issues of both incomplete data , and pathway redundancy , we formulate a parsimony version of the pathway reconstruction/inference problem , called MinPath ( Minimal set of Pathways ) , which can be roughly described as the following: given a set of reference pathways and a set of proteins ( and their predicted functions ) that can be mapped to one or more pathways , we attempt to find the minimum number of pathways that can explain all proteins ( functions ) ( see Fig . 1 ) . Although this problem is NP-hard in general , we provide an integer programming ( IP ) framework to solve it . We focus on analyzing complete genomes in this study because there is a relatively good understanding of the pathways that actually exist in organisms with completely sequenced genomes ( as compared to the emerging metagenomes ) , making this analysis a good test of our method . Besides , the pathway annotations of these genomes are still far from perfection , as in the example of a carbon fixation pathway in the human genome ( as well as chickens , mosquitoes , etc ) . We also applied MinPath to the analyses of several metagenomic datasets , to demonstrate the potential applications of MinPath in metagenome annotation .
We used MinPath to re-analyze the biological pathways of several metagenomes [17] , which were previously analyzed by a naïve mapping approach . The results are summarized in Table 3 . We used both the KEGG and SEED databases in this experiment . For KEGG pathways , we did local BLAST searches , using the criteria as shown in [16] for KO family identification . For SEED subsystems , the FIG annotations were downloaded from the MG-RAST server ( http://metagenomics . theseed . org/ ) . For all the datasets we tested , MinPath reduced the total number of annotated pathways ( or subsystems ) significantly ( as shown in Table 3 ) . For example , for the metagenome sampled from a coral microbial community ( Coral-Mic ) , there are in total 232 KEGG biological pathways annotated in at least one of the 7 sequencing datasets . Based on MinPath , however , only 160 KEGG biological pathways are sufficient to explain all the functions predicted for these datasets . These results indicate that the naïve mapping of the biological pathways from predicted functions may overestimate the biological pathways ( so the functional diversity ) of those microbial communities , and we need to be cautious when interpreting the results from such an analysis [16] , [17] . We also show the details of pathway reconstruction for a single sequence dataset from the coral biome ( 4440319 . 3 . dna . fa ) . The naïve mapping approach identified 224 KEGG pathways , whereas MinPath identified only 143 KEGG pathways . The pathways eliminated by MinPath include the inositol metabolism pathway , the androgen and estrogen metabolism pathway , the caffeine metabolism pathway , etc ( see more examples at the supplementary website ) . Obviously , comparisons of microbial communities or other biomes will be more telling if spurious pathways are eliminated , and our results suggest that as many as 40% of the 224 pathways could be wrong .
We have developed the MinPath approach to provide more conservative—but more reliable—estimations of biological pathways from a sequence dataset , and applied this approach to revisit the biological pathway reconstruction problem for genomes as well as metagenomes . Our results show that without further post-processing of the reconstructed pathways , the naïve mapping strategy may overestimate the biological pathways that are encoded by a genome or metagenome , which could jeopardize any conclusions drawn from the constructed biological pathways ( such as the metabolic diversity/capacity of an environmental microbial or viral community , as measured by the Shannon Index ) [16] , [17] , or other downstream analysis based on constructed pathways [23] . It was noted in [16] that most of the microbial communities in that study were approaching saturation for known pathways: more conservative estimates of pathways for each environment may allow real functional differences between the samples to be detected . Note that MinPath is not designed to directly improve the still imperfect definition of pathways and/or functions in databases such as KEGG or SEED . For example , as a result of how some pathways are grouped in the KEGG database , peptidoglycan biosynthesis is listed for the human genome by KEGG annotation and MinPath does not eliminate this pathway from the list of annotated pathways from human genome . In this sense , efforts are still needed to improve the elucidation and annotation of extent biochemical pathways . But given a database of reference pathways , we feel that MinPath provides a sensible method for inferring the pathways represented in biological sequence samples .
Pathway reconstruction has become routine in functional annotation of genomes and metagenomes , in which KEGG pathways ( or other biological pathways such as SEED subsystems ) are reconstructed based on homology . KEGG and SEED databases collect pathways ( or subsystems ) curated by experts , each pathway/subsystem consisting of a series of functional roles ( enzymes , transporters , etc ) . Pathway reconstruction consists of two key steps: ( 1 ) predicting the functions ( represented by protein families ) of proteins encoded by the DNA sequences , which is often achieved by similarity searching of the predicted proteins against reference proteins from previously characterized genomes; and ( 2 ) predicting the presence or absence of pathways in the query dataset , based on the identified functions associated to the pathways . Conventional pathway reconstruction usually adopts simple criterion in this second step ( herein referred to as the naïve mapping approach ) , i . e . , a pathway is considered to be present if one or more functions in the pathway are identified in the first step . We have shown in this paper that this approach may lead to the identification of spurious pathways and an overestimation of functional ability , which motivated us to develop a novel approach to pathway reconstruction based on the parsimony principle presented below . We define the minimal pathway reconstruction problem as the following: given a list of functions annotated for a set of genes ( which can be an incomplete set , as we encounter in metagenomic analysis , or a nearly complete set , as in complete genome analysis ) , find the minimal set of pathways that include all given functions ( see Fig 1 ) . Note that this formulation is different from the conventional formulation of the pathway reconstruction problem , which attempts either to reconstruct the complete pathways encoded by a given genomic dataset ( in a sense , the pathway holes should to be minimized ) , or to identify the set of pathways that have at least one associated function annotated ( i . e . , the naïve mapping approach ) . We use integer programming to solve the minimal pathway reconstruction problem . Linear programming ( LP ) is an algorithm for finding the maximum or minimum of a linear function of variables ( objective function ) that are subject to linear constraints [24] . Simplex and interior point methods are widely used for solving LP problems . The related problem of integer programming ( IP ) requires some or all of the variables to take integer ( whole number ) values . Some of the most powerful algorithms for finding exact solutions of combinatorial optimization problems [25] are based on IP . LP and IP have been applied to many fields in the biological sciences , such as the maximum contact map overlap problem for protein structure comparison [26] , optimal protein threading [27] , probe design for microarray experiments [28] , and the pathway variant problem [8] . Here we transform the minimal pathway reconstruction problem to an integer programming problem: Denote the number of functions ( protein families ) that are annotated in a dataset as n . Let the total number of putative pathways which have at least one component function annotated be p . Denote the mapping of protein functions to the pathways as M , where Mij = 1 if function i is involved in pathway j , otherwise 0 ( note one function may map to multiple pathways or subsystems ) . Denote if a pathway j is selected in the final list or not as Pj , with Pj = 1 if selected , Pj = 0 otherwise . The set of pathways with Pi = 1 composes the minimal set of pathways that can explain all the functions that are annotated for a dataset . The objective function for integer programming is , i . e . , our goal is to find the minimum number of pathways that can explain all the functions carried by at least one protein from a dataset . We use the KO and FIG protein families defined in the KEGG database and the SEED subsystems , respectively , for this study . Many of the mappings of KO families to KEGG pathways were done manually in the KEGG database . These families are the basic units for pathway reconstruction ( or subsystem reconstruction in SEED ) , in which a pathway ( or a subsystem ) is composed of a list of functional roles . We use the GLPK package ( GNU Linear Programming Kit; http://www . gnu . org/software/glpk/glpk . html ) for solving the integer-programming problem; all the other functions are implemented in Python . The input for MinPath is a list of protein families ( e . g . , KO and FIG families ) annotated in a given dataset of genes ( from a genome , or a metagenome ) , and the output is the list of pathways reconstructed/inferred for the dataset . Note that in some cases two pathways may share most of their functional roles ( for example , the biosynthesis and degradation pathway of the same biological molecule , such as the lysine biosynthesis and degradation pathways ) . MinPath will keep one of these pathways , because that is sufficient to explain the functional roles identified . We added a post-processing step here to add those pathways that have more than 50% of their functional roles identified back to the pathway pool , even when these functional roles appear in another pathway that is already predicted by MinPath . We revisited the pathway reconstruction for the 854 genomes in the KEGG database ( as of December , 2008 ) that have at least 20 KEGG pathways annotated for each of these genomes . For these genomes , the function ( or protein families ) annotations were downloaded from the KEGG database ( ftp://ftp . genome . jp/pub/kegg/release/current/ ) . We also applied MinPath to reanalyze the pathways for nine biome metagenomic datasets [17] . The FIG family annotations for the metagenomic sequences were downloaded from the MG-RAST server ( http://metagenomics . theseed . org/ ) . We conducted the KO family annotations of the sequences based on the best blast hits with E-value cutoff of 1e-5 , a typical E-value cutoff used for KEGG pathway reconstruction in metagenomes [16] . MinPath is available as a server and the source codes are available for downloading at MinPath webpage , http://omics . informatics . indiana . edu/MinPath/ . Supplementary material is also available at the MinPath website .
|
Even though there is only a single large biological network within any cell and all pathways are to some extent connected , the partition of the entire cellular network into smaller units ( e . g . , KEGG pathways ) is extremely important for understanding biological processes . Biological pathway reconstruction , therefore , is essential for understanding the biological functions that a newly sequenced genome encodes and recently for studying the functionality of a natural environment via metagenomics . The common practice of pathway reconstruction in metagenomics first identifies functions encoded by the metagenomic sequences and then reconstructs pathways from the annotated functions by mapping the functions to reference pathways . To address the issues of both incomplete data ( e . g . , metagenomes , unlike individual genomes , are most likely incomplete ) and pathway redundancy ( e . g . , the same function is involved in multiple pathway units ) , we formulate a parsimony version of the pathway reconstruction/inference problem , called MinPath ( Minimal set of Pathways ) : given a set of reference pathways and a set of functions that can be mapped to one or more pathways , MinPath aims at finding a minimum number of pathways that can explain all functions . MinPath achieves a more conservative , yet more faithful , estimation of the biological pathways encoded by genomes and metagenomes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"computational",
"biology/metagenomics",
"computational",
"biology/metabolic",
"networks",
"computational",
"biology/genomics"
] |
2009
|
A Parsimony Approach to Biological Pathway Reconstruction/Inference for Genomes and Metagenomes
|
Statistical analyses of genomic data from diverse human populations have demonstrated that archaic hominins , such as Neanderthals and Denisovans , interbred or admixed with the ancestors of present-day humans . Central to these analyses are methods for inferring archaic ancestry along the genomes of present-day individuals ( archaic local ancestry ) . Methods for archaic local ancestry inference rely on the availability of reference genomes from the ancestral archaic populations for accurate inference . However , several instances of archaic admixture lack reference archaic genomes , making it difficult to characterize these events . We present a statistical method that combines diverse population genetic summary statistics to infer archaic local ancestry without access to an archaic reference genome . We validate the accuracy and robustness of our method in simulations . When applied to genomes of European individuals , our method recovers segments that are substantially enriched for Neanderthal ancestry , even though our method did not have access to any Neanderthal reference genomes .
Admixture , the exchange of genes among previously isolated populations , is increasingly being recognized as an important force in shaping genetic variation in natural populations . Analyses of large collections of genome sequences have shown that admixture events have been prevalent throughout human history [1] . These studies have shown that modern human populations outside of Africa trace a small percentage of their ancestry to admixture events from populations related to archaic hominins like Neanderthals and Denisovans [1 , 2 , 3] . Further , studies of the functional impact of archaic ancestry have suggested that Neanderthal DNA contributes to phenotypic variation in modern humans [4 , 5] . Central to these studies is the problem of archaic local ancestry inference—the pinpointing of segments of an individual genome that trace their ancestry to archaic hominin populations . Methods for archaic local ancestry inference leverage various summary statistics computed from modern and ancient genomes . For example , at a given genomic locus , individuals with archaic ancestry are expected to have low sequence divergence to an archaic genome [6] . A number of summary statistics [7 , 8 , 9] as well as statistical models that combine these statistics [2 , 10 , 11 , 12] to infer archaic local ancestry have been proposed . These methods are most effective in settings where reference genomes that represent genetic variation in the archaic population are available . For example , the analyses of Neanderthal [6 , 10] and Denisovan admixture events [13] relied on the genome sequences from the respective archaic populations . In a number of instances , however , the archaic population is either unknown or lacks suitable reference genomes . Several recent studies have found evidence for archaic introgression in present-day African populations from an unknown archaic hominin [14 , 15 , 16] while analysis of the high-coverage Denisovan genome has suggested that the sequenced individual traces a small proportion of its ancestry to a highly-diverged unknown archaic hominin [10] . One of the most widely used statistics for identifying archaic ancestry is the S*-statistic [9] , which identifies highly diverged SNPs that are in high linkage disequilibrium ( LD ) with each other in the present-day population as likely to be introgressed . The S*-statistic is attractive as it can be applied even where no reference genome is available . However , the power of the S*-statistic tends to be low in the reference-free setting [3] and its accuracy depends on a number of parameters that need to be fixed in advance . Here , we introduce a new statistical method , ARCHaic Introgression Explorer ( ArchIE ) , that combines several population genetic summary statistics to accurately infer archaic local ancestry without the need for a reference genome . ArchIE is based on a logistic regression model that predicts the probability of archaic ancestry for each window along an individual genome . The parameters of ArchIE are estimated from training data generated using coalescent simulations . Our proposed method has several advantages . First , the model can incorporate a variety of statistics that are potentially informative of archaic ancestry . This flexibility allows the model to be applied to the reference-free setting ( the setting that is the focus in this paper ) . However , the model can be extended to also incorporate reference genomes when available , even when these reference genomes might be from distant representatives [10] or from low-coverage samples [17 , 18] . Second , our use of a statistical model allows us to efficiently estimate model parameters that optimize desired objective functions such as the likelihood . This property allows the model to be adapted to admixture events with different time depths or admixture fractions as well as to infer other population genetic parameters of interest . Indeed , recent studies have shown that statistical predictors that combine weakly-informative summary statistics can substantially improve a number of population genetic inference problems [19 , 20 , 21] . We show that ArchIE obtains improved accuracy in simulations over the S*-statistic ( as well as the recently proposed S’ method [22] ) while being robust to demographic model misspecifications that can cause the distribution of features and archaic ancestry labels in the training data to differ from the test data . We apply ArchIE to Western European ( CEU ) genomes from the 1000 Genomes project and show that the segments inferred to harbor archaic ancestry have an increased likelihood of being introgressed from Neanderthals even though no Neanderthal genome was used in the inference . These segments recover previously observed features of introgressed Neanderthal ancestry: we observe a decreased frequency of these segments in regions of the genome with stronger selective constraint [23] as well as elevated frequency at the BNC2 and OAS loci that have previously been reported to harbor elevated frequencies of Neanderthal ancestry [2 , 3] .
Our method , ArchIE , aims to predict the archaic local ancestry state in a given window along an individual haploid genome . This prediction is performed using a binary logistic regression model given a set of features computed within this window . Estimating the parameters of this model requires labeled training data i . e . , a dataset containing pairs of features and the archaic local ancestry state for a given window along an individual genome . To obtain labeled training data , we simulate data under a demographic model that includes archaic introgression , label windows as archaic or not , compute features that are potentially informative of introgression , and estimate the parameters of our predictor on the resulting training data ( Fig 1A , Methods ) . While our method is general enough to be applicable to non-human populations , we describe the demographic model in terms of a modern human-archaic human demographic history . We simulate training data using a modified version of the coalescent simulator , ms [24] , which allows us to track each individual’s ancestry . We use the demographic model from Sankararaman et al . 2014 [2] ( See Table 1 ) . In this model , an ancestral population splits T0 generations before present ( B . P . ) forming two populations ( archaic and modern human in the case of the Neanderthal-human demography ) . The modern human population subsequently splits into two populations Ts generations B . P . , one of which then interbreeds with the archaic population ( referred to as the target population ) while the other does not ( the reference population ) . We simulate one haploid genome ( haplotype ) in the archaic population , 100 haplotypes in the target population and 100 haplotypes in the reference population ( thus , a target population consists of 50 diploid individuals ) . We sample the archaic haplotype at the same time as the modern human haplotypes , but the statistics we calculate do not rely on features of the archaic genome . We simulate 10 , 000 replicates of 50 , 000 base pairs each ( bp ) , resulting in 1 , 000 , 000 training examples . We use a window of length 50 Kb because that is the mean length of the introgressed archaic haplotype after Ta = 2 , 000 generations based on the recombination rate assumed in our simulations . We summarize the training data using features that are likely to be informative of archaic admixture . Since we are interested in the probability of archaic ancestry for a given focal haplotype , we compute features that are specific for the focal haplotype . First , for the focal haplotype , we calculate an individual frequency spectrum ( IFS ) , which is a vector of length n , the haploid sample size of the target population . Each entry in the vector is the number of mutations on the focal haplotype that are segregating in the target population with a specific count of derived alleles . Due to the accumulation of private mutations in the archaic population , we expect the IFS to capture the excess of alleles segregating at frequencies close to the admixture fraction in the introgressed population . This statistic is closely related to the conditional site frequency spectrum [25] . Next , we calculate the Euclidean distance between the focal haplotype and all other haplotypes , resulting in a vector of length n . Under a scenario of archaic admixture , the distribution of pairwise differences is expected to differ when we compare two haplotypes that are both modern human or archaic versus when we compare an archaic haplotype to a modern human haplotype . We also include the first four moments of this distribution , i . e . , the mean , variance , skew , and kurtosis . These summaries of haplotype distance are similar to the D1 statistic used in Hammer et . al . [14] . The next set of features rely on a present-day reference human population that has a different demographic history compared to the target population . The choice of the reference can alter the specific admixture events that our method is sensitive to: we expect the method to be sensitive to admixture events in the history of the target population since its divergence from the reference . While our method can also be applied in the setting where no such reference population exists , in the context of human populations where genomes from a diverse set of populations is available [1] , the use of the reference can improve the accuracy and the interpretability of our predictions . Given a reference population , we compute the minimum distance of the focal haplotype to all haplotypes in the reference population . A larger distance is suggestive of admixture from a population that diverged from the ancestor of the target and reference populations before the reference and target populations split . This feature shares some similarities with the D2 statistic from Hammer et . al . [14] . We also calculate the number of SNPs private to the focal haplotype , removing SNPs shared with the reference , as these SNPs are suggestive of an introgressed haplotype . Finally , we calculate S* [9] , a statistic designed for detecting archaic admixture by looking for long stretches of derived alleles in high LD . Using these features , we train a logistic regression classifier to distinguish between archaic and non archaic segments . In our training data , we define archaic haplotypes as those for which ≥ 70% of bases are truly archaic in ancestry and non-archaic as those for which ≤ 30% are archaic in ancestry . We discard haplotypes that fall in-between those values in the training data resulting in 988 , 372 training examples . We tested the accuracy of ArchIE by simulating data under a demography reflective of the history of Neanderthals and present-day humans [2] . We evaluated the ability of ArchIE to correctly predict the archaic ancestry at each SNP along an individual haplotype . Since ArchIE predicts archaic ancestry within a window , we simulated a 1 Mb segment , applied ArchIE in a 50 Kb window that slides 10 Kb at a time , and predicted archaic ancestry at a SNP by averaging predictions across all windows that overlap the SNP ( Methods ) . We compute Receiver Operator Characteristic ( ROC ) and Precision Recall ( PR ) curves by varying the threshold at which we call a SNP archaic and calculating the true positive rate ( TPR ) , false positive rate ( FPR ) , precision , and recall ( Fig 2 ) . We compared ArchIE to an implementation of the S*-statistic from Vernot and Akey using their hyper parameter choices [3] and to S’ , a new method for reference-free inference of archaic ancestry [22] ( Methods ) . At a 2% admixture fraction , ArchIE outperforms the S* and S’ statistics across all thresholds ( Fig 2A and 2B ) . At a precision of 0 . 80 , i . e . , false discovery rate of 20% , ArchIE obtains a recall of 0 . 21 , S* obtains a recall of 0 . 04 , and S’ obtains a recall of 0 . 09 . The area under the ROC curve ( AUROC ) is 0 . 94 ( ±0 . 008 ) for S* , 0 . 84 ( ±0 . 01 ) for S’ , and 0 . 97 ( ±0 . 005 ) for ArchIE and the area under the PR curve ( AUPR ) is 0 . 47 for S* ( ±0 . 031 ) , 0 . 28 ( ±0 . 032 ) for S’ , and 0 . 60 ( ±0 . 05 ) for ArchIE ( All standard error were estimated using a block jackknife [26] using 1 Mb blocks ) . We also note that while the ROC curves are similar , the PR curves show a large difference , indicative of the utility of PR curves in problems where there is an imbalance in the frequencies of the two classes . We also evaluated the ability of ArchIE to call archaic haplotypes . Since haplotypes can range from having none of their ancestry to being entirely from the archaic population , we called haplotypes archaic if they contain ≥ 70% archaic ancestry or not archaic if they contain ≤ 30% . We see that again , ArchIE has larger AUPR ( 0 . 53 for ArchIE , 0 . 38 for S* ) and AUROC ( 0 . 97 for ArchIE , 0 . 94 for S* ) compared to S* ( S4 Fig ) . We examined the absolute value of the standardized weights learned by ArchIE to understand the features that contribute substantially to its predictions . Examining single features , we find that the minimum distance between the focal haplotype and each of the reference haplotypes , as well as the skew of the distance vector have the largest weights ( Fig 3B ) . Intuitively , a larger distance to a reference population should indicate archaic ancestry . The next largest single statistic was the skew of the distance vector , which was negatively correlated with archaic ancestry . Under a simple scenario of admixture , we expect a bi-modal distribution of pairwise distances . However , when there is little archaic ancestry , the distribution will be unimodal resulting in a negative relationship between skew and archaic ancestry . The IFS contains mostly negative weights , suggesting that these features do not make a substantial contribution to the model predictions ( Fig 3A ) . As a further check , we wanted to determine how the performance of the model changes when trained on subsets of the features . First , since the “skew” feature has a large standardized absolute weight , we trained a model based only on this feature ( S5 Fig ) . We find that accuracy greatly decreases , indicating that the model does best when it combines multiple features that are informative of archaic introgression . However , when we train only on the number of private SNPs or only on the minimum distance to the reference population , we see improved accuracy indicating that these features are informative of archaic ancestry independent of other features . When we take a combination of three features ( skew , number of private SNPs , and minimum distance to the reference population ) , this model is still able to discern archaic from non-archaic haplotypes with slight decreased accuracy relative to the full model ( S5 Fig ) . Finally , we tested the contribution of the reference population to the accuracy of ArchIE . We trained the logistic regression without using any features that rely on the reference and found that model still retains reasonable accuracy ( AUPR = 0 . 36 ) to identify archaic ancestry ( S5 Fig ) . This suggests that ArchIE is useful even in scenarios where a reference population is not available . ArchIE relies on simulating data from a model with fixed demographic and population genetic parameters . In practice , these parameters are unknown and are inferred from data with some uncertainty . Thus , we wanted to determine the sensitivity of our method to demographic uncertainty . An exhaustive exploration of demographic uncertainty is challenging given the number of parameters associated with even the simplest models . As an alternative to an exhaustive exploration , we systematically perturbed each parameter at a time , simulated data using the perturbed model , and evaluated the performance of our classifier ( trained on the unperturbed parameters corresponding to the Neanderthal demographic history ) . ArchIE remains accurate when many aspects of the demography are misspecified , but has reduced precision or recall under some scenarios ( Fig 4 , S1 Fig ) . The most significant decrease in accuracy ( in terms of recall and precision at a fixed threshold ) arises when the reference population size is decreased or the split time of the reference and the target is increased . In this setting , the reference genomes are more drifted and hence , less representative of the ancestral population . We also compared the accuracies of ArchIE to S* across these perturbations and found that ArchIE remains relatively accurate across these settings ( S1 Table ) . We also tested the effect of variation in mutation rate ( μ ) and recombination rate ( r ) since we trained our model using fixed values of these parameters ( μ = 1 . 25 × 10−8 , r = 1 × 10−8 ) . To evaluate how ArchIE performs on real data , we simulated test data randomly drawing pairs of μ and r from a distribution chosen to match local recombination and mutation rates along the human genome ( see Methods ) . The overall AUPR is reduced ( 0 . 31 , S1i Fig ) , the log10 fold changes in precision and recall are −0 . 30 and +0 . 19 suggesting that ArchIE is relatively robust to variation in mutation and recombination rates . In addition , we tested the impact of the window size and found that reasonable choices of window size do not substantially impact the performance ( S2 Fig ) . We also assessed the impact of sample size by simulating 30 haplotypes ( 15 diploid individuals ) , representing a modestly sized genomic dataset , and found a reduction in power as expected ( AUPR = 0 . 45 ) ( S3 Fig ) . We tested the sensitivity of ArchIE to recent and ancestral structure in the demographic model . We simulated data under two scenarios of structure , one where 25% of the target population separates immediately after the target and reference population split , 2499 generations ago , and rejoins the generation prior to the archaic admixture , 2001 generations ago ( S6A Fig ) . We refer to this as the recent structure scenario . Additionally , we simulated data where 25% of the population in N0 separates 12 , 000 generations ago and rejoins the ancestral population right before the target and reference populations split ( 2600 generations ago , S6B Fig ) . We refer to this as the ancestral structure scenario . We observe that for both scenarios , the fraction of SNPs detected as archaic is 0 , suggesting that ArchIE is robust to introgression due to either recent or ancient structure at reasonable calling thresholds . We caution , however , that a more detailed exploration of structured demographic models is necessary . To identify segments of archaic ancestry in modern human populations , we applied ArchIE to genomes of European individuals in the 1000 Genomes Project [27] . We used all unrelated individuals from a European ( CEU ) population as our target population ( 99 diploid individuals ) and all unrelated individuals from an African ( YRI ) population as a reference ( 108 diploid individuals ) and calculated the summary statistics described above . We applied ArchIE in non-overlapping 50 Kb windows . We evaluated the average percent of windows inferred as archaic as a function of the calling threshold ( Fig 5A ) . Applying a threshold corresponding to a precision of 0 . 80 in simulations , we inferred 2 . 04% ( block jackknife SE = 0 . 6% using 1 Mb blocks ) of the genome as confidently archaic . This proportion is in line with proportion of Neanderthal ancestry from previous analyses [2 , 6 , 10] suggesting that the segments of archaic ancestry inferred by ArchIE likely correspond to segments of Neanderthal ancestry . To further investigate whether the haplotypes inferred as confidently archaic by our model are enriched for introgressed Neanderthal variants , we computed a Neanderthal match statistic ( NMS ) defined as the number of shared variants between an individual haplotype and the Altai Neanderthal reference genome sequence [10] divided by the total number of segregating sites in that window ( see Methods ) . We see that the archaic regions confidently inferred by ArchIE have a higher NMS suggesting that the archaic ancestry segments identified by our method are likely to represent introgressed Neanderthal sequence ( we reject the null hypothesis that the difference in NMS is zero for archaic vs non-archaic haplotypes with a P value = 1 . 7 × 10−3 via 100 Kb block jackknife ) . Further , as we make the calling threshold more strict , we see an increase in the mean NMS for the archaic haplotypes ( Fig 5B ) . We also compared the performance of ArchIE , S’ , and S* on real data from CEU Europeans . For each of these methods , we computed a matching rate with the Altai Neanderthal genome , defined as the fraction of SNPs called archaic that match the Altai Neanderthal sequence divided by the total number of SNPs called archaic . At a detection rate of ≈ 1% , S’ has a matching rate of 0 . 73 while ArchIE has a matching rate of 0 . 91 ( S9 Fig; see S1 Text for details ) . Comparing with the S* calls released from [28] , we found a match rate of ≈ 50% at a detection rate of ≈ 0 . 5% , consistent with results reported from the authors . We then focused on two genomic regions that have been shown to harbor introgressed Neanderthal haplotypes at elevated frequencies: the BNC2 gene ( Chromosome 9:16 , 409 , 501-16 , 870 , 786 ) [2] and the OAS gene cluster ( Chromosome 12:113 , 344 , 739-113 , 357 , 712 ) [7] . ArchIE detects substantially increased frequency of archaic ancestry in both these genes ( Fig 5C and 5D ) . Finally , we analyzed the correlation between a measure of selective constraint of a given genomic region ( B-value [23] ) and frequency of confidently inferred archaic segments in the CEU population in the same region . Sankararaman et al . 2014 [2] describe a relationship where more constrained regions ( lower B-value ) have a lower frequency of archaic ancestry . We observe the same trend where more neutral regions ( B-value ≥ 750 ) contain more archaic ancestry than constrained regions ( B-value ≤ 250 ) consistent with selection against the archaic ancestry ( P value = 7 . 86 × 10−9 via block jackknife; Fig 5E ) . These analyses suggest that ArchIE obtains results concordant with those from a previous reference-aware method [2] . We caution , however , that the observed concordance can be inflated due to any biases shared by the two methods .
A key challenge in detecting the contribution of deeply-diverged populations ( both deeply-diverged modern as well as archaic hominin populations ) to the ancestry of present-day human populations arises from the lack of accurate representative genomes for these populations . Here , we present a statistical model ( ArchIE ) for detecting regions of archaic local ancestry without the need for an archaic reference sequence . ArchIE combines weakly informative signals computed from present-day human genomes using a logistic regression model . The parameters of the model are estimated from data simulated under a specific demographic model . Using simulations , we show that ArchIE obtains improved accuracy over other approaches for reference-free local ancestry inference . While the accuracy of ArchIE will depend on how similar the demographic model used for training is to the true demographic model , our empirical results suggest that ArchIE is relatively robust even when the true demographic model differs from the assumed model . Applying ArchIE to genomes from the CEU population in the 1000 Genomes project data , we detect 2 . 03 ± 0 . 6% archaic ancestry ( at a threshold that corresponds to a false discovery rate of 0 . 2 ) . We find that segments confidently labeled as archaic by ArchIE are enriched for Neanderthal ancestry . One advantage of our approach is that the learning algorithm is general allowing it to be applied broadly to diverse inference problems as well as input summary statistics while its simplicity allows for a transparent interpretation of the features and the model . There are several limitations of our methodology , however . First , we require some knowledge of the demographic history of the target , reference and archaic populations . We have shown that ArchIE is robust to some demographic misspecification , but it is most powerful when the simulated demography is close to the true one . Second , we rely on the data being phased . Switch-errors in phasing will reduce the power of ArchIE , which can be a problem when applying the method to less-well studied populations . In principle it is possible to use ArchIE on unphased data , calculating features on the diploid individual level rather than the haplotype level , though we do not explore that here . Third , the use of a fixed-size window ignores long-range as well as variable-length dependence among the features . Models that account for this dependency can be expected to yield improved accuracy . An example of such an approach is a recently published method that uses a hidden Markov model ( HMM ) that models the distribution of private variants [12] . Combining such models with the framework outlined here has the potential to yield improved accuracies . Fourth , the use of a linear model is likely to underfit the true function between features and outputs . It is possible to train more expressive models like deep neural networks , which can learn and capture non-linear relationships between features and tend not to suffer from the curse of dimensionality [19] . These methods have been used to great success in tasks such as image classification [29] and we anticipate their use in population genetics could improve predictive power . Preliminary results applying deep learning to this problem with the features used here are promising , motivating future work ( S1 Text , S7 and S8 Figs ) . ArchIE relies on a careful choice of features as input . These hand crafted features are informed by population genetics theory , similar to other methods that have been proposed in population genetics [19 , 20 , 30 , 31 , 14] . Automatically learning features from genetic data is direction of high interest . Finally , while several methods [9 , 12 , 22] have been proposed to infer aspects of archaic ancestry without access to reference genomes , these methods are typically evaluated using simulations . Assessing the accuracy of these methods on real data remains challenging . Extrapolating simulation results to accuracy on real data depends on choices of the inference problem , population genetic models , parameters used for training and testing , genomic features used as input , and accuracy metrics of interest . A comprehensive comparison of these methods across a range of demographic histories and evolutionary forces is an important topic for future work . In conclusion , our method improves on previous methods for reference-free inference of archaic ancestry by combining informative summary statistics in a statistical learning framework . We anticipate that this method will be informative not only in human populations where questions about admixture with other hominins abound , but also in other species and systems where pervasive admixture has shaped the distribution of genetic variation .
We simulated training and test data sets using a modified version of ms [24] that tracks the ancestry of each site in each individual genome . Using a previously proposed demographic model relating modern humans and Neanderthals [2] , we sampled 100 haplotypes from the target , and 100 haplotypes from the reference over a region of length 50 Kb . We use a constant mutation rate μ = 1 . 25 × 10−8 and a recombination rate r = 1 × 10−8 . The general demography is as follows: an archaic population of size Na splits from a population of size N0 , T0 generations before present ( B . P . ) . Then , at TS , two populations split off from the ancestral population that then have effective population sizes N1 ( termed the reference ) and N2 ( termed the target ) respectively . Then , at time TA , the archaic population migrates into the target with an admixture fraction m . See Fig 1 for a graphical outline . Each simulation at a given locus generates 100 haplotypes in the target . For each haplotype , we calculate the following classes of summary statistics: individual frequency spectrum , distance vector to all haplotypes within the test population as well as the first four moments of this vector , minimum distance to haplotypes in the reference population , the number of private SNPs , and the S*-statistic . The individual frequency spectrum is created as follows: given a sample of n haplotypes , for each haplotype j , we construct a vector X of length n where entry Xi counts the number of derived alleles carried on the focal haplotype j whose derived allele frequency is i . For example , the first entry counts the number of singletons present in haplotype j , the second entry counts the number of doubletons and so on until n . The distance vector is a vector of length n where entry i is the Euclidean distance from haplotype j to haplotype i over all sites , where j is the focal haplotype and i is the haplotype being compared . The minimum distance to haplotypes in the reference population is computed as the minimum Euclidean distance from the focal haplotype to all haplotypes in the reference population . The number of private SNPs is calculated as the number of SNPs the focal haplotype contains that are not present in the reference population . This results in 208 features per example ( a 50 Kb window for a single haploid genome ) , with 100 examples per locus and 10 , 000 loci resulting in 1 , 000 , 000 examples for training before filtering haplotypes with intermediate levels of admixture . We used the “glm” function in R to construct a logistic regression model using the family = binomial ( “logit” ) option . We used the predict function to obtain a prediction and converted it to a probability using the “plogis” function . Due to the process of recombination , the ancestry of a haplotype may vary along its length . On the other hand , ArchIE predicts a single ancestry state for a haplotype across a specified window . We evaluate the ability of ArchIE to predict the ancestry at each SNP along a haplotype by simulating sequences of length 1 Mb and applying ArchIE in 50 Kb windows , sliding by 10 Kb at a time . We average the predictions that each SNP on a haplotype receives across all windows that overlap the SNP to obtain the predicted archaic ancestry . We compare the predicted and the true ancestry state at each SNP along a haplotype . We evaluated the performance using Precision-Recall ( PR ) curves as well as receiver operator characteristic ( ROC ) curves . We calculated precision ( equivalently 1− the false discovery rate ) , recall ( equivalently sensitivity ) and false positive rates as: R e c a l l ( t ) = T P ( t ) T P ( t ) + F N ( t ) S e n s i t i v i t y ( t ) = P r e c i s i o n ( t ) = T P ( t ) T P ( t ) + F P ( t ) F a l s e p o s i t i v e r a t e ( t ) = F P ( t ) F P ( t ) + T N ( t ) Here TP ( t ) is the number of true positives at threshold t , FN ( t ) is the number of false negatives at threshold t , FP ( t ) is the number of false positives at threshold t and TN ( t ) is the number of true negatives at threshold t . We summarize these results by reporting the recall at a fixed value of precision as well as by computing the area under the precision recall curve ( AUPR ) and the area under the ROC curve ( AUROC ) . We compute the AUPR using the method of Davis and Goadrich [32] . We compute standard errors of the AUPR and AUROC using a block jackknife [26] where we drop a single 1 Mb region and recompute the statistics . We compared ArchIE to the S* [9] and S’ [22] statistics . We calculate S* in a cohort of 100 haplotypes from the target population . Then , we convert the S* scores into a rank between [0-1] using the empirical cumulative distribution . We use a 50 Kb sliding window ( 10 Kb stride ) across the 1 Mb region , averaging the score for a SNP . We use a similar strategy for S’ . However , since S’ predicts archaic ancestry in a sample of individuals rather than on the haploid genome level , we use an algorithm to convert sample predictions to haploid genome predictions . We run S’ on the sample . Then , at some S’ score threshold , we find the longest stretch of SNPs at that score or higher and interpolate the scores across genotypes , building haplotypes when individuals have the archaic allele . Then , for each SNP , we evaluate whether the SNP is archaic or not and calculate the number of true positives , false positive , true negatives , and false negatives . We repeat this procedure across thresholds and calculate the precision , recall , and false positive rates . We examined the robustness of ArchIE to a specified demographic model by systematically perturbing one parameter at a time , simulating a dataset , and evaluating ArchIE’s performance . We doubled and halved the parameters , except when doing so would produce a demographic model that is not sensible . We evaluated the robustness of ArchIE to mutation and recombination rate variation by calculating local rates at 50 Kb windows and then randomly drawing combinations of the rates and simulating data . Mutation rates were calculated by estimating Watterson’s θ [33] from the number of segregating sites within 50 Kb windows across 50 randomly sampled west African Yoruba genomes from the 1000 Genomes Project Phase 3 release and calculating the mutation rate: μ = θw/4NeL where we set Ne = 10 , 000 . Recombination rates were estimated from the combined , sex-averaged HapMap recombination map [34] . We validated our method using the Neanderthal introgression scenario as a test case . We downloaded phased CEU genomes from the 1000 Genomes Phase 3 dataset [27] and calculated the features mentioned above in 50 Kb windows . For each individual haplotype , we inferred the probability that the window is archaic . We then intersected our calls with the 1000 Genomes strict mask using BEDtools v2 . 26 . 0 [35] , removing regions that are difficult to map to , measured as having less than 90% of sites in the callability mask . We calculated a Neanderthal match statistic ( NMS ) for focal haplotype i in a window as the fraction of alleles at which the the focal haplotype matches the Altai Neanderthal [10] genome: N M S i = S i N i + H i Here Si denotes the number of alleles that match between the focal haplotype and the Neanderthal genome within the window . Since the Neanderthal genome is not phased , we count sites as matching if it contained at least one single matching allele or more . Ni denotes the number of Neanderthal mutations , including both homozygous and heterozygous sites . Hi denotes the number of human mutations within the window . In order to test whether there is more Neanderthal matching in archaic haplotypes compared to non-archaic haplotypes , we computed the difference in NMS between the two classes of haplotypes at each window and test the hypothesis that the mean of this statistic averaged across the genome is zero . Specifically: Δ N M S , i = N M S ¯ a r c h , i - N M S ¯ n o n - a r c h , i N M S i ¯ For each window i , we compute ΔNMS , i , defined as the difference between the mean NMS for archaic ( N M S ¯ a r c h , i ) and non-archaic ( N M S ¯ n o n - a r c h , i ) haplotypes divided by the mean NMS of all haplotypes ( N M S i ¯ ) to control for mutation rate heterogeneity . We require a minimum of 90% callable sites within the window . We compute the mean of ΔNMS , i over all windows i as the genome-wide estimate and test if this estimate is significantly different from zero . To compute significance , we use a block jackknife and drop non-overlapping 100 Kb windows and recalculate the genome wide difference in means . In order to assess the relationship between background selection and inferred archaic ancestry , we use the B-values from McVicker et al . 2009 [23] and intersected them with our calls . For visualization , we binned the B-values into 4 bins , [0-250] , ( 250-500] , ( 500-750] , and ( 750-1000] . We tested for significant differences in allele frequency between the lowest and highest bins using a block jackknife using a 50 Kb block size .
|
Recent analyses of modern human genomes have shown that archaic hominins like Neanderthals and Denisovans contribute a few percentage of ancestry to many populations . These analyses rely on having accurate reference genomes from these archaic populations . Due to the difficulty in sequencing these genomes , we lack a complete collection of reference genomes with which to identify archaic ancestry . Here , we develop a method that identifies segments of archaic ancestry in modern human genomes without the need for archaic reference genomes . We systematically evaluate the accuracy and robustness of our method and apply it to modern European genomes to uncover signals of introgression which we confirm to be from a population related to Neanderthals .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"population",
"genetics",
"social",
"sciences",
"anthropology",
"human",
"genomics",
"genetic",
"mapping",
"simulation",
"and",
"modeling",
"neanderthals",
"archaic",
"humans",
"paleontology",
"population",
"biology",
"paleoanthropology",
"research",
"and",
"analysis",
"methods",
"hominids",
"introgression",
"hominins",
"haplotypes",
"physical",
"anthropology",
"heredity",
"earth",
"sciences",
"evolutionary",
"processes",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"evolutionary",
"biology",
"genomics",
"statistics",
"computational",
"biology"
] |
2019
|
A statistical model for reference-free inference of archaic local ancestry
|
MicroRNAs ( miRNAs ) play important roles in normal cellular differentiation and oncogenesis . microRNA93 ( mir-93 ) , a member of the mir106b-25 cluster , located in intron 13 of the MCM7 gene , although frequently overexpressed in human malignancies may also function as a tumor suppressor gene . Using a series of breast cancer cell lines representing different stages of differentiation and mouse xenograft models , we demonstrate that mir-93 modulates the fate of breast cancer stem cells ( BCSCs ) by regulating their proliferation and differentiation states . In “claudinlow” SUM159 cells , expression of mir-93 induces Mesenchymal-Epithelial Transition ( MET ) associated with downregulation of TGFβ signaling and downregulates multiple stem cell regulatory genes , including JAK1 , STAT3 , AKT3 , SOX4 , EZH1 , and HMGA2 , resulting in cancer stem cell ( CSC ) depletion . Enforced expression of mir-93 completely blocks tumor development in mammary fat pads and development of metastases following intracardiac injection in mouse xenografts . The effect of mir-93 on the CSC population is dependent on the cellular differentiation state , with mir-93 expression increasing the CSC population in MCF7 cells that display a more differentiated “luminal” phenotype . mir-93 also regulates the proliferation and differentiation of normal breast stem cells isolated from reduction mammoplasties . These studies demonstrate that miRNAs can regulate the states and fates of normal and malignant mammary stem cells , findings which have important biological and clinical implications .
miRNAs serve vital functions in many of normal developmental processes , as well as in carcinogenesis . A number of these miRNAs have been shown to function as oncogenes with increased expression in lung cancer , prostate cancer and colorectal cancer [1] , [2] , [3] , [4] , [5] , [6] , [7] , [8] . In contrast , other miRNAs such as Let7 are frequently downregulated in malignancies including breast cancer and lung cancer in these contexts functioning as a tumor suppressor gene [9] , [10] , [11] . The mir106b-25 cluster is composed of the highly conserved miRNA106b ( mir-106b ) , miRNA93 ( mir-93 ) and miRNA25 ( mir-25 ) that have been reported to be overexpressed in a number of cancers including gastric , prostate and pancreatic neural endocrine tumors , neuroblastoma and multiple myeloma [1] , [2] , [3] . These miRNAs are located in a 515-base region on chromosome band 7q22 in intron13 of the host MCM7 gene where they are co-transcribed in the context of MCM7 primary transcripts [1] . MCM7 is a DNA licensing factor obligate for cellular replication . Studies have suggested that the mir-106b-25 miRNA cluster functions as a proto oncogene . Several studies suggest that a primary mechanism of oncogenesis involves targeting of PTEN which cooperates with MCM7 to drive cellular proliferation [12] . Despite evidence for this miRNA cluster functioning as a proto oncogene , in some contexts it has been reported to function as a tumor suppressor inhibiting tumor growth [13] . The molecular mechanisms accounting for this discrepancy have not been determined . Studies associating miRNA expression with oncogenesis have largely been performed in bulk tumor populations . However , there is substantial evidence supporting the CSC hypothesis which suggests that tumors are hierarchically organized and that many tumors , including those of the breast , are maintained by a subpopulation of cells that displays stem cell properties [14] , [15] , [16] . These cells may mediate invasion and metastasis and contribute to treatment resistance [17] . miRNAs have also been found to play important roles in normal and malignant stem cell function . Silber et al , reported that mir-124 and mir-137 induced differentiation of neural and glioblastoma stem cells , a state associated with cell cycle arrest [18] . Furthermore , recent studies have shown that the miRNAs Let7 and mir-200c regulate self-renewal of BCSCs [10] , [19] . Stem cell regulatory genes such as BMI-1 and HMGA2 may mediate this process [10] , [19] . We have previously demonstrated that normal breast tissue , primary breast cancers and breast cancer cell lines contain subpopulations with stem cell properties that can be enriched by virtue of their expression of aldehyde dehydrogenase ( ALDH ) as assessed by the Aldefluor assay ( Stem Cell Technologies , Inc . , Vancouver , British Columbia ) or by tumor initiation in NOD/SCID mice [20] . Recently , Ibarra , et al , showed that Let7 , as well as mir-93 are highly depleted in mouse mammary stem/progenitor cells isolated with the stem cell marker ALDH [21] . We have utilized breast cancer cell lines representing different molecular subtypes of breast cancer as well as primary xenografts of breast cancer and normal mammary cells to examine the role of mir-93 in the regulation of normal and malignant breast stem cells . We demonstrate that this miRNA is able to regulate stem cell fate including cellular proliferation and differentiation . These studies suggest that miRNAs regulate the transition between CSC states findings which have important biological and clinical implications .
We have previously demonstrated that primary human breast cancers and established breast cancer cell lines contain subpopulations with stem cell properties that can be isolated by virtue of their expression of ALDH as assessed by the Aldefluor assay . These cells displayed increased tumor initiating capacity and metastatic potential compared to corresponding Aldefluor-negative cells [22] . mir-93 was shown as one of the most abundant miRNAs in ALDH− cells [21] . As assessed by qRT-PCR , mir-93 expression was significantly increased in the ALDH− compared to ALDH+ populations in SUM159 claudinlow and HCC1954 basal subtype of human breast cancer ( Figure 1A and Figure S1A ) . As shown in Figure S3 , mir-93 expression was lower in CSCs which were characterized by their expression of the CSC markers: ALDH+ or CD24−CD44+ . To determine the relationship between mir-93 expression and tumor initiating capacity , we constructed a mir-93 sensor tagged with GFP ( mir-93-sensor-GFP ) containing a mir-93 target UTR coupled to GFP . In cells transfected with this vector , mir-93 expression results in degradation of GFP mRNA ( sensor-negative ) , whereas mir-93-negative cells express GFP ( sensor-positive ) ( Figure 1B ) . mir-93 expression was significantly higher in GFP-negative cells than GFP-positive cells ( Figure S4 ) and the ALDH1A1 was much lower in GFP-negative cells than GFP-positive cells as accessed by western blot or immunohistochemical staining ( Figure S5 ) . Furthermore , GFP was significantly reduced by overexpression of mir-93 ( Figure S6 ) , demonstrating that the sensor reports mir-93 function . The relationship between mir-93 expression and tumor initiation was determined by introducing serial dilutions of sensor-positive ( mir-93-negative ) and sensor-negative ( mir-93-positive ) SUM159 cells into the mammary fatpads of NOD/SCID mice . As shown in Figure 1C , sensor-positive ( mir-93-negative ) cells had significantly higher tumor initiating capacity and CSC frequency than sensor-negative ( mir-93-positive ) cells . Moreover , mir-93-negative cells gave rise to tumors containing both mir-93-negative and mir-93-positive populations , whereas mir-93-positive cells gave rise only to small , slow growing tumors containing exclusively mir-93-positive populations ( Figure 1D ) . Similar findings were seen using HCC1954 cells ( data not shown ) . These studies demonstrated that in these breast cancer cell lines low mir-93 expression is associated with the CSC phenotype characterized by increased aldehyde dehydrogenase expression , tumor initiating capacity and the ability to generate heterogeneous tumors containing both stem cell and non-stem cell populations . We utilized a tetracycline ( TET ) inducible mir-93 construct tagged with RFP ( pTRIPZ-mir-93-RFP ) to determine the functional role of mir-93 in CSCs . mir-93 levels were significantly increased by ten hours following tetracycline induction in these cells ( Figure 2A ) . Induction of mir-93 was associated with a significant decrease in the CSC population as assessed by the Aldefluor assay ( Figure 2B ) , which were also seen in two basal breast cancer cell lines HCC1954 and SUM149 ( Figure S1B and Figure S7 ) . Furthermore , this decrease did not result from induction of apoptosis in these cells as assessed by Annexin V staining ( Figure 2B ) . Our group and others have previously shown that CSCs were relatively resistant to cytotoxic chemotherapy . Consistent with this , addition of the cytotoxic agent docetaxel resulted in a relative increase in the percentage of Aldefluor-positive cells ( Figure 2B ) , an increase associated with induction of apoptosis in the bulk cell population ( 42 . 8% versus 1 . 1% control ) ( Figure 2B ) . The relative increase in the Aldefluor-positive population seen with docetaxel treatment was abrogated by simultaneous mir-93 expression ( Figure 2B ) . These experiments suggested that unlike cytotoxic agents which primarily target the bulk cell population , mir-93 overexpression was able to reduce the CSC population . Moreover , this did not appear to result from increased CSC apoptosis suggesting a potential role for mir-93 in promoting differentiation of CSCs . Furthermore , since the TET-inducible mir-93 system allows for the controlled regulation of CSC populations , it provides a valuable tool for assessing the role of CSCs in tumor growth in mouse xenograft models . Furthermore , the ability to regulate the CSC population during different phases of tumor growth allows for the assessment of the role of these cells in tumor initiation and maintenance . We first determined the effect of mir-93 induction on the growth of established tumors and compared these effects to those of cytotoxic chemotherapy . When tumors reached 0 . 2–0 . 3 cm in diameter , we induced mir-93 ( with doxycycline treatment , hereafter DOX ) or initiated cytotoxic chemotherapy with docetaxel or the combination . Induction of mir-93 significantly inhibited the growth of SUM159 and HCC1954 xenografts ( Figure 2C and Figure S1C ) . Furthermore , induction of mir-93 further reduced tumor growth when added to the docetaxel chemotherapy ( Figure 2C and Figure S1 ) . Following five weeks of treatment , animals were sacrificed and CSC populations were assessed by the Aldefluor assay and ALDH1 immunohistochemistry . Induction of mir-93 alone or in combination with docetaxel reduced the Aldefluor-positive population by more than 60% compared to control or docetaxel alone ( Figure 2D and Figure S1D ) . These observations were confirmed by immunohistochemistry of ALDH1 expression ( Figure 2D and Figure S1D ) . mir-93 expression was significantly higher in DOX group compared to the control group at the end of treatment ( Figure S2 ) . To provide a more definitive assessment of CSCs , we determined the ability of serial dilutions of cells obtained from primary tumors to form tumors in secondary NOD/SCID mice . Tumor cells isolated from docetaxel treated mice , initiated tumors at lower concentrations with accelerated growth compared to control animals ( Figure 2E , Figure S1E ) . This was consistent with previous studies demonstrating a relative increase in CSCs following chemotherapy [17] . In contrast , cells isolated from tumors with mir-93 induction with or without docetaxel chemotherapy had markedly reduced tumor initiating capacity in secondary mice with no tumors observed from introduction of fifty cells from the mir-93 docetaxel treated group ( Figure 2E , Figure S1E ) . The CSC frequency was lower in the groups of DOX alone and DOX+docetaxel , and was significantly increased in the docetaxel group ( Figure 2E and Figure S1E ) . These studies demonstrated that mir-93 induction reduced the CSC population reducing growth of established tumor xenografts . In order to determine whether down-regulation of mir-93 promoted tumorigenesis , we utilized a mirZip anti-sense miRNA in SUM159 cells . qRT-PCR was utilized to confirm the efficient knock-down of mir-93 ( Figure S8A ) . ALDH+ cells were significantly increased after mir-93 was knocked down ( mirZip93-DsRed ) ( Figure S8B ) . As shown in Figure S8C , knockdown of mir-93 significantly promoted the growth of SUM159 cells in tumor xenografts and increased the CSC frequency . Furthermore , the proportion of ALDH+ cells were significantly increased after mir-93 was knocked down ( mirZip93-DsRed ) ( Figure S8D ) . Preclinical models have suggested that CSCs play a role in tumor recurrence and metastasis following adjuvant therapy [23] . This suggests that targeting of CSCs may have more dramatic effects in the adjuvant than in the advanced tumor settings . To simulate the adjuvant setting we induced mir-93 and/or administered docetaxel immediately after tumor cell implantation . Although tumors grew after four to five weeks in control animals , there was no observed tumor growth following mir-93 induction and/or docetaxel treatments for eight weeks ( Figure 2F , Figure S1F ) . After eight weeks , treatments were stopped and animals observed for an additional ten weeks . In SUM159 xenografts , tumors developed in all mice who received eight weeks of docetaxel alone . In contrast , no tumors developed in mice following mir-93 induction with or without docetaxel ( Figure 2F , Figure S1F ) . In order to extend these observations to primary breast tumors , we examined the effect of mir-93 induction on three primary breast xenografts , MC1 ( Figure S10A ) , UM2 ( Figure S10B ) and UM1 ( Figure S10C ) which were directly established from patient tumors and not passaged in vitro . MC1 and UM1 were derived from claudinlow and UM2 from a basal breast carcinoma . Induction of mir-93 upon cell implantation completely prevented tumor growth in this model . Together these studies suggested that mir-93 regulated the CSC population and that this population mediates tumor growth following adjuvant therapy . Previous studies have demonstrated that CSCs mediate tumor invasion and metastasis . To determine the effect of mir-93 expression on tumor invasion , we examined the effect of mir-93 induction and/or downregulation on invasion of SUM159 cells using a matrigel invasion assay . Overexpression of mir-93 significantly inhibited the ability of SUM159 cells to invade in this assay ( Figure S8E ) . In contrast , knockdown of mir-93 utilizing the mirZip93-DsRed promoted tumor invasion ( Figure S8F ) . To determine whether the expression of mir-93 affect the growth of tumor metastasis in vivo , SUM159 ( Figure 2G ) and HCC1954 ( 1G ) cells co-transfected with the inducible mir-93 vector and luciferase were introduced into NOD/SCID mice by intracardiac injection and metastasis formation monitored by bioluminescence imaging . DOX and/or docetaxel treatments were initiated following intracardiac injection . As shown in Figure 2G , mir-93 induction completely suppressed whereas docetaxel partially suppressed metastasis formation . Metastasis was confirmed by histologic examination with pan-cytokeratin staining ( Figures S9 , S1H ) . Treatments were stopped at eight weeks and animals were observed for an additional ten weeks for development of metastasis . In animals receiving docetaxel alone , metastasis rapidly developed following cessation of therapy . In contrast , no metastases developed in mice following mir-93 induction with or without docetaxel chemotherapy ( Figure 2G ) . In animals injected with HCC1954 cells , animals from all groups developed metastasis following cessation of therapy . However , development of metastasis were delayed and reduced in mice following mir-93 induction with or without docetaxel chemotherapy ( Figure S1G ) . Human breast cancer represents a heterogeneous set of diseases with distinct molecular profiles and clinical behaviors [24] . These subtypes may represent different cells of origin and/or differentiation state . It has been proposed that the most undifferentiated “claudinlow” tumors originate from and resemble normal mammary stem cells , whereas the triple-negative basal tumors arise from a more differentiated luminal progenitor cell and the most differentiated luminal tumors which express estrogen and progesterone receptors originate from and are composed of the most differentiated cells [24] . To determine the relationship between mir-93 expression and level of cellular differentiation , we compared the expression of mir-93 in claudinlow ( SUM159 ) , basal ( HCC1954 ) and luminal ( MCF7 ) cells . As shown in Figure S11 , mir-93 levels correlate with postulated differentiation state of these cell lines . Furthermore , in the claudinlow SUM159 cells and basal HCC1954 cells , mir-93 expression is significantly lower in Aldefluor-positive as compared to Aldefluor-negative populations ( Figure S11 ) . In contrast , the CSC population in MCF7 cells characterized by the phenotype CD24−CD44+ [25] expressed the same high level of mir-93 as did the other ( non-stem ) cells constituting the bulk population ( Figure 3A , Figure S11 ) . This suggests that mir-93 may play a different role in more differentiated luminal breast cancer than in the more undifferentiated claudinlow and basal subtype . Consistent with this , induction of mir-93 in MCF7 cells increased the CD24−CD44+ population ( Figure 3B ) . Docetaxel also increased this population , as did the combination of mir-93 plus docetaxel ( Figure 3B ) . In xenografts , induction of mir-93 accelerated the growth of MCF7 xenografts compared to control ( Figure 3C ) , findings which were confirmed using two additional luminal cell lines MDA-MB-453 and T47D ( Figures S12A and S13A ) . In contrast , docetaxel reduced tumor growth ( Figure 3C ) . Analysis of treated MCF7 tumors confirmed that mir-93 induction increased the proportion of CD24−CD44+ cells and ALDH+ cells in tumors as did docetaxel or DOX plus docetaxel ( Figure 3D ) . mir-93 expression level was significantly higher in DOX group compared to the control group at the end of treatment ( Figure S14 ) . mir-93 induction increased the proportion of ALDH+ cells from 1 . 01% to 9 . 5% in MDA-MB-453 tumors ( Figure S12B ) and from 1 . 26% to 3 . 84% in T47D tumors ( Figure S13B ) . Furthermore , the calculated tumor initiating frequency was significantly increased after mir-93 induction ( Figure 3E , Figures S12C and S13C ) . These results were confirmed and extended by demonstrating that mir-93 induction in primary tumors increased their tumor-initiating capacity when implanted into secondary recipients ( Figure 3E , Figures S12C , S13C ) . Together , these experiments suggested that the effects of mir-93 on the CSC population differed in different molecular subtypes of breast cancer , an observation consistent with the hypothesis that miRNA effects might be differentiation state dependent . In order to determine the cellular targets of mir-93 in BCSCs , ALDH+ and ALDH− populations of SUM159 cells were separated and cultured in suspension in the presence or absence of DOX for ten hours . Gene expression profiles in the four populations were determined utilizing Affymetrix oligonucleotide microarrays ( Figure 4A ) . Of the 2 , 000 genes downregulated at least two-fold upon DOX treatment in the ALDH+ population ( Table S1 ) , 127 overlapped with the predicted target sequences of mir-93 including twenty-four genes known to be involved in stem cell regulation ( Figure 4A and Table S2 ) including JAK1 , SOX4 , STAT3 , AKT , E2H1 and HMGAZ . The downregulation of these genes in pTRIPZ-SUM159-mir-93 , pTRIPZ-HCC1954-mir-93 cell lines and pTRIPZ-MC1-mir-93 were confirmed with customized PCR array plates ( Figures S15 , S16 , S17 ) . In contrast , only 352 genes were significantly downregulated by DOX in the ALDH− population ( Table S3 ) with twelve of these genes ( no stem cell genes ) overlapping with the predicted mir-93 targets . These studies suggest that mir-93 regulates the CSC population by simultaneously targeting a number of stem cell regulatory genes . To confirm this , we utilized a luceriferase reporter assay to determine the effect of mir-93 on the expression of the stem cell regulatory genes AKT3 , SOX4 and STAT3 selected from the expression profiling data . Expression of mir-93 reduced the level of these stem cell regulatory genes in SUM159 ( Figure 4B ) and HCC1954 cells ( Figure S18 ) but not in luminal MCF7 and MDA-MB-453 cells ( Figure S19 ) . Furthermore , knockdown of STAT3 or SOX4 but not AKT3 decreases the proportion of ALDH+ SUM159 cells suggesting these genes play a role in the regulation of CSC self-renewal ( Figure S20 ) . The 127 genes in pTRIPZ-MCF7-mir-93 were also tested with customized PCR array plates , and interestingly , most of the stem cell genes were not knocked-down by mir-93 induction in the ALDH+ proportion of MCF7 ( Figure S21 ) . To determine the relationship between mir-93 expression and cell cycle kinetics , we assessed mir-93 expression in quiescent and cycling ALDH+ and ALDH− populations . Cycling ( S/G2/M ) cells expressed significantly higher levels of mir-93 compared to quiescent ( G0/G1 ) cells in both the ALDH− and ALDH+ compartments ( Figure 5A ) . To determine whether mir-93 induces or is a consequence of cellular proliferation , we utilized the DOX inducible mir-93 construct to determine the effect of mir-93 induction on cell cycle distribution . Induction of mir-93 reduced the quiescent cell population from 64% to 42% suggesting that this miRNA has the capacity to directly regulate the cell cycle ( Figure 5B ) . Furthermore , induction of mir-93 increased the proliferation of SUM159 by 29% ( Figure S22 ) . Although mir-93 induction had similar effects on the basal HCC1954 cell lines it had no significant effect on the cell cycle of the luminal MCF7 cells ( Figure S22 , Figure S23 ) . To determine the relationship between the stem cell phenotype and cell cycle kinetics , we determined the cell cycle distribution of ALDH+ and ALDH− populations . The ALDH+ population in SUM159 cells had a higher fraction of non-cycling cells compared to ALDH− cells ( Figure 5A ) . This finding was confirmed by Ki67 and MCM7 staining ( Figure S24 ) . SUM159 cells are derived from a “claudinlow” subtype of breast cancer which is characterized as having a high proportion of cells displaying “epithelial-mesenchymal transition ( EMT ) ” . This state is characterized by loss of epithelial characteristics such as apical basal polarity and E-Cadherin expression and acquisition of mesenchymal characteristics , including loss of cell polarity and expression of Vimentin . We determined the effects of mir-93 expression on MET of SUM159 cells by assessing markers of these states at the protein and mRNA levels . SUM159 cells have a mesenchymal morphology and express Vimentin , but not the epithelial marker E-Cadherin , an effect not dependent on cell density ( Figure 6A ) . Expression of mir-93 in these cells caused them to assume a more epithelial appearance associated with a decrease in Vimentin and an increase in E-Cadherin expression ( Figure 6A ) . Similar effects were seen in the basal HCC1954 cell line ( Figure S25 ) although these were less pronounced . To confirm and extend these results we determined the effect of mir-93 expression on mRNA expression of a wider panel of epithelial and mesenchymal markers . We also determined the time course of expression of epithelial and mesenchymal marker mRNAs expressed in ALDH+ stem cells and ALDH− non-stem cell populations . Expression of mir-93 in SUM159 cells resulted in a time dependent decrease in expression of mesenchymal markers , Vimentin , N-cadherin and Twist , and an increase in the epithelial markers E-Cadherin and Claudin ( Figure 6B ) . Furthermore , although these effects were seen in ALDH− populations , they were even more pronounced in the ALDH+ stem cell compartment . Since TGFβ is a major inducer of the EMT [26] , [27] , we examined the effect of mir-93 expression on components of this pathway . Interestingly , expression of mir-93 significantly reduced expression of the mRNA for TGFβR2 in both ALDH+ and ALDH− SUM159 cells . This effect was seen as early as twelve hours , suggesting a potential role for down regulation of TGFβ signaling in inducing the MET . In addition to breast cancer cells , we also determined the effects of mir-93 expression on normal breast cell differentiation . We utilized flow cytometry to access expression of EpCAM and CD49f in breast epithelial cells obtained from reduction mammoplasties . It has previously been shown that mammary stem cells are contained within the EpCAM−CD49f+ population while double positive ( EpCAM+CD49f+ ) cells are luminal progenitors , EpCAM+CD49f− more differentiated Luminal cells , while EpCAM−CD49f− constitute stromal cells [28] . We compared mir-93 expression levels in these four populations . Interestingly , we found that the highest level of mir-93 is expressed in the EpCAM+CD49f+ population ( Figure 7A ) , which suggested mir-93 was required to maintain the cells as EpCAM+CD49f+ . Furthermore , overexpression of mir-93 in freshly isolated normal breast cells or in immortalized non-transformed MCF-10A cells increased the proportion of cells expressing EpCAM ( Figure 7B , 7C ) . These studies suggested that mir-93 played a role in maintaining normal breast cells in an epithelial ( EpCAM+ ) state .
In these studies , we demonstrate that mir-93 is capable of modulating breast CSC populations by regulating their proliferation and differentiation states . To examine this , we utilized breast cancer cell lines representing different states of differentiation . The levels of endogenous mir-93 expression paralleled cellular differentiation states with mir-93 levels lowest in the most primitive “claudinlow” SUM159 cells , highest in the “luminal” MCF7 cells and intermediate in the “basal” HCC1915 cells . We utilized a DOX inducible system to determine the effects of enforced mir-93 expression on the CSC populations assessed by expression of the stem cell markers ALDH and CD24−CD44+ as well as by mouse xenograft assays [14] , [22] . Enforced mir-93 expression in claudinlow and basal breast cancer cell lines significantly reduced the CSC populations as assessed by the Aldefluor assay . To assess the functional relevance of this , we determined the effect of mir-93 induction in SUM159 and HCC1954 cells on tumor growth in NOD/SCID mouse xenografts . The effects of mir-93 expression on tumor initiating capacity was confirmed using two primary breast xenografts generated without in vitro culture . mir-93 expression decreased the CSC in these claudinlow primary xenografts . In contrast , overexpresson of mir-93 in the luminal MCF7 cells line resulted in an increase in CD24−CD44+ CSC resulting in increased tumor growth . This demonstrates that the effect of mir-93 on CSC populations is dependent on the cellular differentiation state . This model allowed us to simulate potential clinical scenarios involving CSC targeting agents . To simulate the effects of CSC targeting agents in advanced disease , tumors were inoculated into mammary fatpads and when the tumors were palpable mir-93 was induced by addition of doxycycline to the mouse drinking water . In this setting , mir-93 induction had only a modest effect in reducing tumor growth . Addition of the chemotherapeutic agent docetaxel resulted in a more significant reduction in tumor size , an effect that was accentuated by mir-93 induction . CSC models predict that the efficacy of CSC targeting agents should be most pronounced in the adjuvant setting where tumor growth from micrometastasis is dependent on stem cell self-renewal [29] . Consistent with this model , induction of mir-93 immediately after fatpad implantation or after development of micrometastasis by intracardiac injection completely blocked tumor recurrence . Furthermore , when treatment was discontinued at eight weeks , animals that received chemotherapy alone developed local tumor growth and metastasis while those with mir-93 induction with or without chemotherapy showed no recurrence when animals were sacrificed after four months . These studies provide strong support for the CSC hypothesis and provide a valuable animal model for clinical trial design using CSC targeting agents . To determine the molecular mechanisms of mir-93 CSC regulation , we employed an unbiased approach assessing the effect of mir-93 expression on early changes in global gene expression profile coupled with prediction of miRNA target sequences . Interestingly , this analysis revealed that twenty-four genes known to be involved in stem cell self-renewal including JAK1 , SOX4 , STAT3 , AKT , EZH1 , HMGA2 are targeted by mir-93 . In addition , this miRNA targets two important regulators of TGFβ signaling , TGFβR2 and SMAD5 . mir-93 expression was also associated with and in turn regulates cellular proliferation . Quiescent G0/G1 cells expressed lower levels of mir-93 than proliferating cells in S/G2/M phase . Furthermore , enforced expression of mir-93 increased the fraction of cycling cells . We demonstrate that induction of mir-93 in mesenchymal-like SUM159 cells induces an MET in the ALDH+ CSC population characterized by increased expression of E-Cadherin and Claudin , with concomitant downreguation of mesenchymal genes , such as Vimentin , N-Cadherin and Twist . mir-93 also inhibits TGFβ signaling by targeting TGFβR2 , an effect seen within twelve hours of mir-93 induction . This was followed by an MET in the Aldefluor-positive CSC population . Since TGFβ is a major regulator of EMT , abrogation of this signaling pathway may facilitate MET . Of interest , it has been recently reported that the mir-106b-25 cluster including mir-93 is induced in the early stages of nuclear reprogramming of fibroblasts into IPS cells [30] . This is accompanied by a mesenchymal to epithelial conversion in these cells which is obligatory for reprogramming to recur . This suggests that this miRNA cluster may regulate MET in multiple biological contexts . In summary , our experiments suggest that CSCs can exist in two alternative epithelial and mesenchymal states , the balance of which is regulated by miRNAs including mir-93 ( Figure 8 ) . The mesenchymal state associated with an invasive phenotype characterized by quiescence and low mir-93 expression is maintained by growth factors such as TGFβ . Upon activation of cellular proliferation , MYC and E2F are induced leading to expression of MCM7 , a licensing factor required for DNA synthesis . Concomitantly , mir-93 and its related miRNA cluster is co-synthesized which promotes further proliferation while simultaneously downregulating TGFβ signaling . This facilitates a mesenchymal to epithelial transition in the CSC population characterized by decreased invasiveness and increased proliferation . Continued expression of mir-93 simultaneously downregulates a number of stem cell self-renewal pathways including JAK/STAT , AKT , EZH1 and HMGH2 , promoting cellular differentiation and depleting the CSC population . The model depicted in Figure 8 is consistent with our observation that mir-93 level is highest in the EpCAM+CD49f+ normal mammary cells and decreased with terminal differentiation . In contrast , the effects of mir-93 depend on the cellular differentiation state accounting for differences we observed in claudinlow , basal and luminal breast cancers , with mir-93 level highest in the luminal MCF7 cell line compared to basal HCC1954 and claudinlow SUM159 cell lines . MCF7 cells are highly proliferative although unlike normal mammary cells incapable of terminal differentiation ( Figure 8 ) . The existence of alternative CSC states , associated with expression of different protein markers has important implications for understanding the plasticity of CSCs . For example , it has been claimed that CSCs may be generated from non-CSC tumor populations through induction of EMT [31] . However , the existence of alternative CSC state suggests that the acquisition of stem cell markers may reflect transition of CSC states rather than generation of CSCs from non-CSC populations . In addition , the existence of multiple stem cell states suggests the necessity of developing of therapeutic strategies capable of effectively targeting CSCs in all of these states .
Breast cancer cell line SUM159 and SUM149 have been extensively characterized ( http://www . asterand . com ) [32] . HCC1954 , MCF-7 , MDA-MB-453 and MCF10A were purchased from ATCC . The cell lines were grown using the recommended culture conditions . Briefly , the culture medium for SUM159 and SUM149 is Ham's F-12 ( Invitrogen ) supplemented with 5% FBS , 5 ug/mL insulin , and 1 ug/mL hydrocortisone ( both from Sigma , St . Louis , MO ) . MCF7 , MDA-MB-453 and HCC1954 cells were maintained in RPMI1640 medium ( Invitrogen , Carlsbad , CA ) supplemented with 10% fetal bovine serum ( ThermoFisher Scientific , Pittsburgh , PA ) , 1% antibiotic-antimycotic ( Invitrogen , Carlsbad , CA ) , and 5 µg/ml insulin ( Sigma-Aldrich , St Louis , MO ) . 100–200 g of normal breast tissue from reduction mammoplasties was minced with scalpels , dissociated enzymatically , and single cells were isolated as described previously [33] , [34] . The single cells were utilized for FACS sorting or were cultured in suspension as described previously [33] , [34] . Mammospheres were dissociated into single cells enzymatically and mechanically , and then cultured in regular cell culture plates [33] , [34] . For construction of the mir-93 sensor , miRNA-complementary oligonucleotides were annealed and cloned into a Marx vector that directs GFP expression . The mir-93 miRNA target sequence was engineered into the 3′ untranslated region ( UTR ) of the cDNA encoding for the GFP fluorescent protein . Expression of this construct in cells that express the mir-93 miRNA results in a RNAi pathway-dependent degradation of GFP mRNA and thus no green fluorescence . In contrast , in cells with repressed mir-93 miRNA , the GFP mRNA is not degraded and resulting in expression of the fluorescent GFP . shRNA oligos for STAT3 , AKT3 or SOX4 were inserted to PlentiLox3 . 7-DsRed lentiviral vector . A highly efficient lentiviral expression system ( TRIPZ lentivral vector;www . openbiosystems . com/RNAi ) was used to generate mir-93-expressing lentiviruses; and mirZIP-lentivector ( SBI , Mountain View , CA ) was used to generate mir-93-knockdown lentiviruses in UM Vector Core Facility . The cell lines were infected with the lentiviruses as described previously [34] . The Aldefluor kit ( StemCell Technologies , Inc , Vancouver , BC , Canada ) was used to isolate cells with high ALDH enzymatic activity as illustrated in the manufacturer's instructions . Briefly , single cells were suspended in buffer containing ALDH substrate – BAAA ( 1 µmol/l per 1×106 cells ) and incubated at 37°C for 40 minutes . In each experiment , the specific ALDH inhibitor diethylaminobenzaldehyde ( DEAB ) was used as negative control at 50 mmol/L . A FACStarPLUS ( Becton Dickinson ) was used for FACS . Aldefluor fluorescence was excited at 488 nm and fluorescence emission was detected using a standard fluorescein isothiocyanate ( FITC ) 530/30 band pass filter . The sorting gates were established based on negative controls . CD44/CD24 staining was performed as previously described [14] . Briefly , cells were stained with primary antibodies anti-CD44 labeled APC ( dilution 1∶10 , BD Pharmingen ) , and anti-CD24 labeled FITC ( dilution 1∶10 , BD Pharmingen ) . In all in vivo experiments , mouse cells were eliminated by excluding H2Kd+ ( mouse histocompatibility class I , BD Pharmagen ) cells during flow cytometry . 0 . 5 µg/ml 4′ , 6-diamidino-2-phenylindole ( DAPI ) ( Sigma ) was used to access cell viability . Cells ( 1×106 ) were harvested and washed in cold PBS followed by fixation in 70% alcohol for thirty minutes on ice . After washing in cold PBS three times , cells were resuspended in 0 . 8 mL of PBS solution with 40 µg of propidium iodide and 0 . 1 µg of RNase A for thirty minutes at 37°C . Samples were analyzed for DNA content using aFACSCalibur cytometer ( Becton Dickinson , San Jose , CA ) . The effect of mir-93 on cell proliferation was measured using an MTT assay . Briefly , 200–500 cells from Control and DOX-treated groups were seeded in 96-well culture plates and were cultured in the absence ( CTRL ) and presence ( DOX ) of DOX for 7 days . Subsequently , 0 . 025 ml of MTT solution ( 5 mg/ml ) was added to each well , and the cells were incubated for 2 h . After centrifugation , the supernatant was removed from each well . The colored formazan crystal produced from MTT was dissolved in 0 . 15 ml of isopropanol with 4 mM HCl and 0 . 1% NP40 , and the optical density ( OD ) value was measured at 590 nm . Total RNA was isolated using RNeasy Micro Kit ( Qiagen , Valencia , CA ) , and total RNA with enriched miRNA was isolated using miRNeasy mini Kit , according to the manufacturer's instructions . Gene expression analyses used Affymetrix U133 Plus 2 . 0 human oligonucleotide microarrays containing over 47 , 000 transcripts and variants including 38 , 500 well-characterized human genes . Preparation of cRNA , hybridizations , washes and detection were done as recommended by the supplier ( http://www . affymetrix . com/index . affx ) . Expression data were analyzed by the RobustMultichip Average method in R using Bioconductor and associated packages [35] . MiRNA expression level was measured utilizing TaqMan qRT-PCR ( Applied Biosystems , Carlsbad , CA ) . Single-stranded cDNA was synthesized from 10 ng of miRNA enriched total RNA using specific miRNA primers ( TaqMan MiRNA Assay , PN 4427975 , Applied Biosystems ) and the TaqMan MiRNA Reverse Transcription Kit ( PN 4366596 , Applied Biosystems ) . Two ul of cDNA was used as a template in a 20 ul PCR reaction . PCR products were amplified using specific primers ( TaqMan MiRNA Assay ) and the Taq-Man Universal PCR Master Mix ( PN 4324018 , Applied Biosystems ) , and PCR was performed in a ABI PRISM 7900HT sequence detection system with 384-Well block module and automation accessory ( Applied Biosystems ) by incubation at 50°C for two min and then 95°C for ten min followed by forty amplification cycles ( fifteen seconds of denaturation at 95°C and one min of hybridization and elongation at 60°C ) . PCR reactions for each sample were run in triplicate . The number of cycles required for amplification to reach the threshold limit , the Ct-value was used for quantification . RNU24 was used as an endogenous control for miRNA data normalization , and TBP was used as an endogenous control for other gene normalization . All TaqMan miRNA assays used in this study were obtained from Applied Biosystems . All mice were housed in the AAALAC-accredited specific pathogen-free rodent facilities at the University of Michigan . Mice were housed on sterilized , ventilated racks and supplied with commercial chow and sterile water both previously autoclaved . All experimentation involving live mice were conducted in accordance with standard operating procedures approved by the University Committee on the Use and Care of Animals at the University of Michigan . Six-week old female NOD/SCID mice were purchased from Jackson Laboratories ( Bar Harbor , ME ) and housed in SPF microisolator cages in the animal facility of University of Michigan . Tumorigenicity of 10 , 000 ( Adjuvant setting ) cells or 100 , 000 ( Advanced setting ) cells in the mamary fatpads of NOD/SCID mice was accessed . Six mice were included in each cohort . The animals were euthanized when the tumors were 1 . 0–1 . 5 cm in diameter , in compliance with regulations for use of vertebrate animal in research . A portion of each fat pad was fixed in formalin and embedded in paraffin for histological analysis . Another portion was analyzed by the ALDH or CD24/CD44 cytometric staining . Human breast tumors were obtained as biopsy cores or pieces of tumors after surgery and implanted in humanized cleared fat pads of NOD/SCID mice for establishing xenotransplants . The success of xenotransplantation was approximately 20% , similar to previous reports in the literature . Three xenotransplants were used: an ER−PR−ERBB2− tumor at the 20th passage in animals ( MC1 ) , an ER−PR−ERBB2− tumor at the 5th passage ( UM1 ) , and an ER+PR+ERBB2− tumor at the 8th passage ( UM2 ) . All procedures were approved by the University Committee for the Use and Care of Animals ( UCUCA ) of the University of Michigan . The intracardiac injection was carried out according to previously published methods [36] . Briefly , six-week-old NOD/SCID mice were anesthetized with isoflurane gas ( a 2% isofluorane/air mixture ) and injected in the left ventricle of the heart with 100 , 000 cells in 100 µl of sterile Dulbecco's PBS lacking Ca2+ and Mg2+ . For each of the cell lines ( SUM159-luc , HCC1954-luc ) , five animals were injected . Baseline bioluminescence was assessed before inoculation and each week thereafter . Mice were anesthetized with isoflurane gas and given a single i . p . dose of 150 mg/kg D-luciferin ( Promega ) in PBS . For photon flux counting , we used a charge-coupled device camera system ( Xenogen ) with a nose-cone isofluorane delivery system and heated stage for maintaining body temperature . Results were analyzed after six min of exposure using Living Image software provided with the Xenogen IVIS imaging system . For ALDH1 staining , paraffin-embedded sections of breast tumors from xenografts were deparaffinized in xylene and rehydrated in graded alcohol . Antigen enhancement was done by incubating the sections in citrate buffer pH 6 . 0 ( Dakocytomation , Copenhagen , Denmark ) as recommended . Slides were stained using Peroxidase histostain-Plus Kit ( Zymed ) according to the manufacturer's protocol . ALDH1 antibody ( BD biosciences ) was used at a 1∶50 dilution . AEC ( Zymed ) was used as substrate for peroxidase . Slides were counter-stained with hematoxylin and coverslipped using glycerin . For E-Cadehrin , Vimentin , MCM7 , Ki67 and DAPI fluorescent staining , cells were fixed in ice-cold methanol and permeablized with 0 . 15% triton X-100 . E-Cadherin antibody ( Santa Cruz , 1∶100 dilution ) , Vimentin antibody ( Santa Cruz , 1∶200 dilution ) , MCM7 antibody ( Cell signaling , 1∶100 dilution ) , p21 antibody ( Cell Signaling , 1∶400 dilution ) and Ki67 antibody ( Dako , 1∶150 dilution ) were used and incubated for 1 hour at room temperature . PE and FITC labeled secondary antibodies ( Jackson Labs ) were used at the dilution 1∶200 and incubated for twenty min . Nuclei were counterstained with DAPI/antifade ( INVITROGEN ) and cover slipped . Sections were examined with a Leica fluorescent microscope . The pMIR-REPORT luciferase reporter plasimds with the 3′ UTR sequence of AKT3 , SOX4 , STAT3 or the control ACTB were transfected into the cell lines using Fugene HD tansfection reagent ( Roche Applied Science ) according to the manufacturer's instruction . After transfection , cells were dissociated and cultured with or without DOX . Luciferase activity was assayed by luciferase assay kit ( Promega ) . Luciferase activities were measured after forty-eight hrs utilizing a luminometer . The results were presented as the luciferase activity of cells transfected with 3′ UTR sequence of AKT3 , SOX4 , or STAT3 normalized to cells transfected with the luciferase activity of cells transfected with 3′ UTR sequence of ACTB . Results are presented as the mean ± standard deviation ( STDEV ) for at least three repeated individual experiments for each group using Microsoft Excel . Statistical differences were determined by using ANOVA and student's t-test for independent samples . A p-value of less than 0 . 05 was considered statistically significant .
|
Recent evidence suggests that many cancers , including those of the breast , are maintained by a population of cancer cells that display stem cell properties . These “cancer stem cells” may also contribute to tumor metastasis , treatment resistance , and relapse . Recently , miRNAs ( small non-coding RNAs ) have been reported to be capable of functioning as oncogenes or tumor suppressors . miRNA93 ( mir-93 ) is frequently overexpressed in human cancer but , paradoxically , has been found to function as a tumor suppressor in some contexts . Using a series of breast cancer cell lines representing different stages of differentiation and mouse xenograft models , we demonstrate that mir-93 modulates the fate of breast cancer stem cells by regulating their proliferation and differentiation states . In less differentiated tumors , enforced expression of mir-93 completely blocks tumor development in mammary fat pads and development of metastases following intracardiac injection in mouse xenografts by reducing breast cancer stem cells . However , the effect of mir-93 on the cancer stem cell population is dependent on the cellular differentiation state , with mir-93 expression increasing the cancer stem cell population in more differentiated breast tumors . These studies demonstrate that miRNAs can regulate breast stem cell proliferation and differentiation , an observation with important biological and clinical implications .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology"
] |
2012
|
MicroRNA93 Regulates Proliferation and Differentiation of Normal and Malignant Breast Stem Cells
|
Neuroblastoma is a cancer of the developing sympathetic nervous system that most commonly presents in young children and accounts for approximately 12% of pediatric oncology deaths . Here , we report on a genome-wide association study ( GWAS ) in a discovery cohort or 2 , 101 cases and 4 , 202 controls of European ancestry . We identify two new association signals at 3q25 and 4p16 that replicated robustly in multiple independent cohorts comprising 1 , 163 cases and 4 , 396 controls ( 3q25: rs6441201 combined P = 1 . 2x10-11 , Odds Ratio 1 . 23 , 95% CI:1 . 16–1 . 31; 4p16: rs3796727 combined P = 1 . 26x10-12 , Odds Ratio 1 . 30 , 95% CI: 1 . 21–1 . 40 ) . The 4p16 signal maps within the carboxypeptidase Z ( CPZ ) gene . The 3q25 signal resides within the arginine/serine-rich coiled-coil 1 ( RSRC1 ) gene and upstream of the myeloid leukemia factor 1 ( MLF1 ) gene . Increased expression of MLF1 was observed in neuroblastoma cells homozygous for the rs6441201 risk allele ( P = 0 . 02 ) , and significant growth inhibition was observed upon depletion of MLF1 ( P < 0 . 0001 ) in neuroblastoma cells . Taken together , we show that common DNA variants within CPZ at 4p16 and upstream of MLF1 at 3q25 influence neuroblastoma susceptibility and MLF1 likely plays an important role in neuroblastoma tumorigenesis .
Neuroblastoma is a cancer of the developing sympathetic nervous system that most commonly affects children under 5 years of age , with a median age at diagnosis of 17 months [1] . Approximately 50% of cases present with disseminated disease at the time of diagnosis , and despite intense multi-modal therapy , the survival rate for this high-risk subset remains less than 50% [1] . Somatically acquired segmental DNA copy number alterations , such as MYCN amplification and deletions of 1p and 11q , are associated with aggressive disease and poor survival [2] . However , recent whole genome and exome sequencing studies have revealed a relative paucity of somatic point mutations in neuroblastoma tumors [3–6] . In terms of the etiology of neuroblastoma , only 1–2% of patients present with a family history of disease; the vast majority of cases appear to arise sporadically . Familial neuroblastoma is largely explained by germline mutations in ALK [7 , 8] or PHOX2B [9 , 10] . To understand the genetic basis of sporadic neuroblastoma , we are performing a genome-wide association study ( GWAS ) . To date , this effort has identified single nucleotide polymorphisms ( SNPs ) within or upstream of CASC15 [11 , 12] and CASC14 [11] , BARD1 [13 , 14] , LMO1 [15] , DUSP12 [16] , HSD17B12 [16] , DDX4/IL31RA [16] , HACE1 [17] , LIN28B [17] , and TP53 [18] , along with a common copy number variation ( CNV ) within NBPF23 [19] at chromosome 1q21 . 1 , each being highly associated with neuroblastoma . Importantly , several of the neuroblastoma susceptibility genes identified by GWAS have been shown to not only influence disease initiation , but also drive tumor aggressiveness and/or maintenance of the malignant phenotype [15 , 17 , 20–22] . Here , to identify additional germline variants and genes influencing neuroblastoma tumorigenesis , we imputed genotypes across the genome ( see Methods ) and performed a discovery GWAS of genotyped and imputed variants in a cohort of 2 , 101 neuroblastoma patients and 4 , 202 control subjects of European ancestry [17] . This effort refined previously reported susceptibility loci and identified two new association signals at 3q25 and 4p16 which were replicated in three independent cohorts comprising 1 , 163 cases and 4 , 396 controls . In addition , based on expression quantitative trait loci ( eQTL ) analysis and in vitro studies following manipulation of candidate genes in neuroblastoma cell lines , we demonstrate that the 3q25 signal likely targets the myeloid leukemia factor 1 ( MLF1 ) gene in neuroblastoma , resulting in increased MLF1 expression and promoting cell growth .
To discover germline variants associated with neuroblastoma , we performed a GWAS following genome-wide genotype imputation in 2 , 101 neuroblastoma patients accrued through the North American-based Children’s Oncology Group ( S1 Table ) and 4 , 202 control subjects of European ancestry ( see Methods; S1 Fig ) [17] . Individuals were genotyped using the Illumina HumanHap550 or Quad610 Beadchip . Multi-dimensional scaling was used to infer ancestry , and the first twenty components were recorded for subsequent use as co-variates in association testing to control for potential population substructure . To generate imputed genotypes , we first selected SNPs present on both platforms that passed our quality control metrics and applied SHAPEIT to infer haplotypes[23] . We then utilized IMPUTE2 [24] with default parameters and Ne = 20000 , along with a multi-population reference panel from the world-wide 1000 Genomes Project Phase 1 Release 3 to impute genotypes across the entire genome . For quality control purposes , variants with minor allele frequency ( MAF ) <1% and/or IMPUTE2-info quality score <0 . 7 were removed following imputation . The remaining variants were tested for association with neuroblastoma using the frequentist association test under the additive model using the “score” method implemented in SNPTEST [25] ( Fig 1 and S2 Fig ) . The genomic inflation factor was 1 . 04 ( S3 Fig ) . We first confirmed previous reports of neuroblastoma-associated loci , and identified variants of greater statistical significance through imputation at each locus ( Fig 2; S2–S8 Tables ) . Specifically , we observed association at 2q35 implicating BARD1 [13] ( Fig 2A , rs58430496: p = 3 . 05 x 10−11; OR: 1 . 36 , 95% CI: 1 . 25–1 . 48 ) , 6p22 implicating CASC15 [11] , ( Fig 2B , rs4712656: p = 8 . 07 x 10−16; OR: 1 . 37 , 95% CI: 1 . 27–1 . 47 ) , and 6q16-q21 implicating HACE1 [17] ( Fig 2C , rs72990858: p = 1 . 37 x 10−13; OR: 0 . 59 , 95% CI: 0 . 51–0 . 69 ) . After conditioning on rs72990858 at 6q16 , we identified a second independent association signal at 6q16-q21 implicating LIN28B [17] ( Fig 2D , rs17065417: p = 4 . 72 x 10−9; OR: 0 . 70 , 95% CI: 0 . 62–0 . 80 ) . We also confirmed association at 11p15 implicating LMO1 [15] ( Fig 2E , rs2168101: p = 3 . 18 x 10−16; OR: 0 . 70 , 95% CI: 0 . 70–0 . 65 ) , 11p11 implicating HSD17B12 [16] ( Fig 2F , rs10742682: p = 1 . 31 x 10−7; OR: 1 . 24 , 95% CI: 1 . 15–1 . 34 ) 17p13 implicating TP53 [18] ( Fig 2G , rs35850753: p = 1 . 39 x 10−8; OR: 1 . 95 , 95% CI: 1 . 57–2 . 43 ) . We observed two new genome-wide significant associations , the first at 3q25 ( rs6441201: p = 3 . 01 x 10−7; Odds Ratio: 1 . 21 , 95% C . I . : 1 . 12–1 . 30; Fig 3A; Table 1; S9 Table ) and the other at 4p16 ( rs3796727: p = 5 . 25 x 10−9; Odds Ratio: 1 . 26 , 95% C . I . : 1 . 16–1 . 36; Fig 3B; Table 1; S10 Table ) . The novel association signal at 3q25 spans a large 470-Kb linkage disequilibrium ( LD ) block in the HapMap CEU population , encompasses the arginine/serine-rich coiled-coil 1 ( RSRC1 ) gene , and maps just upstream of the myeloid leukemia factor 1 gene ( MLF1 ) ( Fig 3A ) . The signal at 4p16 marks an approximate 27 . 5-Kb LD block in the CEU population and maps within the carboxypeptidase Z ( CPZ ) gene ( Fig 3B ) . To identify potential causal variants at each susceptibility locus , we developed an annotation tool incorporating data from ENCODE [26] , the Roadmap Epigenomics Project [27] , evolutionary conservation , and transcription factor binding motifs ( see Methods ) . We applied this tool to all variants with a discovery p-value < 10−4 , MAF > 0 . 005 , and info score > 0 . 5 . This approach confirmed the recently identified causal variant ( rs2168101 ) at the LMO1 locus shown to disrupt a canonical GATA binding site in neuroblastoma[20] , and identified several other variants that warrant further study . ( S2–S10 Tables ) . To investigate whether more than one association signal may exist at 3q25 or 4p16 , we conditioned our analysis of 3q25 on rs6441201 and our analysis of 4p16 on rs3796727 . No evidence for a separate association signal was observed at either locus ( S4 Fig ) . In addition , no association was observed between rs6441201 or rs3796727 genotypes and clinical/biological covariates , including markers of tumor aggressiveness ( S11 and S12 Tables ) . An interaction analysis between rs6441201 or rs3796727 and the most statistically significant SNPs at each of the previously reported susceptibility loci revealed only weak evidence for epistasis ( S13 Table ) , suggesting that these loci may contribute independently to neuroblastoma risk . We next sought to replicate the new 3q25 and 4p16 association signals in three independent cohorts ( S2 Fig ) . First , we analyzed an African American cohort of 365 neuroblastoma cases and 2 , 491 genetically matched controls [28] . These individuals were genotyped on the Illumina HumanHap550 or Quad-610 bead chips , and SHAPEIT and IMPUTE2 [24] were applied to infer genotypes at the 3q25 and 4p16 loci using data from the 1000 Genomes Phase I Release 3 in a manner similar to the European American cohort . Utilizing the proportion of African admixture as a covariate to correct for varying degrees of admixture among our samples , we confirmed the association of rs6441201 at 3q25 ( p = 5 . 70 x 10−3; Odds Ratio: 1 . 23 , 95% CI: 1 . 04–1 . 45; Table 1; S5 Fig ) . Genotype imputation at the 4p16 locus was of low confidence in this cohort and therefore was not included . Next , we performed PCR-based genotyping in two additional independent cohorts for the top genotyped SNP at 3q25 ( rs6441201 ) , and two SNPs at the 4p16 locus since they were imputed ( rs3796727 and rs3796725 ) . First , we genotyped an Italian cohort of 427 neuroblastoma cases and 783 controls and observed a trend toward association in the same direction seen in the European and African American samples at 3q25 ( rs6441201: P = 0 . 11 , OR: 1 . 15 , 95% CI: 0 . 97–1 . 36 ) and a robust replication at 4p16 ( rs3796727: P = 0 . 010 , OR: 1 . 28 , 95% CI: 1 . 07–1 . 54; rs3796725: P = 4 . 36 x 10−3 , OR: 1 . 33 , 95% CI: 1 . 01–1 . 61 Table 1 ) . Second , we genotyped both SNPs in a cohort of 371 cases and 1 , 122 controls from the United Kingdom , and confirmed all associations ( rs6441201: P = 8 . 45 x 10−4 , OR: 1 . 32 , 95% CI: 1 . 12–1 . 56; rs3796727: P = 1 . 71 x 10−3 , OR: 1 . 33 , 95% CI: 1 . 11–1 . 59; rs3796725: P = 0 . 028 , OR: 1 . 23 95% CI: 1 . 02–1 . 48; Table 1 ) . Meta-analysis using the inverse-variance method within METAL[29] resulted in a highly significant associations with neuroblastoma ( Table 1; rs6441201: P = 1 . 21x10-11 , Odds Ratio 1 . 23 , 95% CI:1 . 16–1 . 31 and rs3796727: P = 126x10-12 , Odds Ratio 1 . 30 , 95% CI:1 . 21–1 . 40; rs3796725: P = 2 . 08 x 10−11 , Odds Ratio 1 . 29 , 95% CI:1 . 19–1 . 38 ) . To investigate whether the neuroblastoma susceptibility variants may function as methylation quantitative trail loci ( meQTL ) , we performed a methylation genome-wide association study based on additive risk genotype of rs6441201 ( 3q25 ) or rs3796727 ( 4p16 ) in a cohort of 769 individuals without cancer for whom we have both SNP and methylation array data , as described previously [30] . Briefly , M-values ( log2 ratio between the methylated and unmethylated probe intensities [31] ) were compared using an additive model based on SNP genotype . Principal component analysis ( PCA ) was first applied to infer ancestry ( S6 Fig ) , and we focused initially on 395 individuals of European ancestry . No evidence was observed for rs6442101 functioning as a meQTL in this cohort . However , in our analysis of rs3796727 genotypes , we observed a single genome-wide significant meQTL signal mapping to the same neuroblastoma-associated locus at 4p16 ( cg14339343 , p = 1 . 33 x 10−16; S7 Fig; S14 Table ) ; this signal replicated in the independent cohort comprised of 332 individuals of African ancestry ( cg14339343 , p = 1 . 36 x 10−6 S8 Fig; S15 Table ) . Analyzing all 769 individuals together in a multi-ethnic meGWAS yielded a highly significant association between rs3796727 genotype and methylation status of cg14339343 ( cg14339343 , p = 5 . 98 x 10−21 Fig 4A; S16 Table ) . Closer examination revealed that this meQTL resides directly within the 5′ UTR of the CPZ gene ( Fig 4B ) , and the rs3796727 risk allele is associated with decreased methylation ( Fig 4C , S9 Fig ) . These data suggest that rs3796727 genotype may influence CPZ expression . While RNA was not available to assess CPZ expression in these individuals , interrogation of the Genotype-Tissue Expression ( GTEx ) Portal revealed that CPZ expression was primarily limited to ovary , cervix and fallopian tube ( S10 Fig ) . Cervix and fallopian tube did not include matched genotype data and thus eQTL analysis was not possible , but ovary tissue showed increased CPZ expression in cells homozygous for the rs3796727 risk allele ( p = 0 . 17; S11 Fig ) . While not reaching statistical significance , this trend is consistent with the observed genotype-methylation correlation . Taken together , these data suggest that rs3796727 genotype may be associated with decreased methylation and increased CPZ expression; further study is necessary to confirm this role for rs3796727 in neuroblastoma directly . To determine if the neuroblastoma-associated SNPs at 3q25 are eQTLs , we utilized the GTEx Portal . The rs6441201 variant at 3q25 was identified as a multi-tissue cis-eQTL for both RSRC1 ( p = 1 . 05 x 10−78; S12 Fig ) and LOC100996447 , a recently discovered long non-coding RNA located at 3q25 ( p = 1 . 14 x 10−145; S13 Fig ) . In addition , rs6441201 was identified as a cis-eQTL for MLF1 in esophagus ( p = 6 . 33 x 10−11; S14 Fig ) . We next analyzed a set of 19 neuroblastoma cell lines with matched genome-wide SNP genotyping and mRNA expression data . The rs6441201 variant was not observed to be an eQTL for RSRC1 in neuroblastoma cells . However , MLF1 expression was significantly higher in neuroblastoma cells harboring the rs6441201 risk allele compared those homozygous for the protective allele ( P = 0 . 02; Fig 5A ) . We further interrogated seven additional genes in the region , but did not observe association of rs6441201 genotype with mRNA levels . Consistent with these findings , silencing of MLF1 , but not RSRC1 , using pooled siRNA resulted in significant cell growth inhibition in neuroblastoma cells ( Fig 5A–5D ) .
Neuroblastoma is an embryonal tumor of the autonomic nervous system thought to arise from developing and incompletely committed precursor cells derived from neural crest tissues; it is the most common cancer diagnosed in the first year of life [1] . Here , in order to identify germline genetic risk factors and genes influencing neuroblastoma tumorigenesis , we performed a genome-wide association studying ( GWAS ) comprising a total of 3 , 264 neuroblastoma patients and 8 , 598 healthy control subjects from four independent cohorts . Two new neuroblastoma susceptibility loci were identified , one at chromosome 3q25 and the other at 4p16 . The 4p16 variants map to the CPZ gene locus , and the 3q25 variants map within RSRC1 and upstream of MLF1 . The CPZ gene encodes a member of the carboxypeptidase E subfamily of metallocarboxypeptidases which represent Zn-dependent enzymes implicated in intra- and extracellular processing of proteins . Through an unbiased meGWAS , we observed strong evidence for rs3796727 functioning as a meQTL for sites within the 5′ UTR of CPZ . Specifically , the rs3796727 risk allele was associated with decreased methylation , suggesting the risk allele may be associated with increased expression of CPZ . CPZ is a Zn-dependent enzyme with an N-terminal cysteine-rich domain ( CRD ) and a C-terminal catalytic domain . CPZ is enriched in the extracellular matrix and expressed during early embryogenesis . In addition to containing a metallocarboxypeptidase domain , CPZ also contains a Cys-rich domain with homology to Wnt-binding proteins [32] . Indeed , studies in chick embryos suggest that CPZ is involved in WNT signaling[33] . In addition , CPZ has been shown to modulate Wnt/beta-catenin signaling and terminal differentiation of growth plate chondrocytes[34] . Among the tissues interrogated in GTEx , CPZ expression was primarily observed in ovary , where there was a trend toward increased expression in cells homozygous for the risk allele ( S10 and S11 Figs ) . Our methylation GWAS based on additive risk allele at the 4p16 susceptibility locus revealed significantly decreased methylation in the 5' UTR of CPZ of cells harboring the risk allele , consistent with increased CPZ expression . Matched RNA was not available to assess mRNA expression in the methylation GWAS cohort , and a genotype-expression correlation was not observed in neuroblastoma cell lines . However , CPZ may influence tumor initiation and thus require assessment of precursor cells from the developing sympathetic nervous system . The 3q25 variants map within RSRC1 which encodes a member of the serine and arginine rich-related protein family . The gene product has been shown to play a role in constitutive and alternative splicing , and is involved in the recognition of the 3′ splice site during the second step of splicing [35] . Variants in RSRC1 are associated with the neurological disease schizophrenia , and RSRC1 is involved in prenatal brain development and cell migration to forebrain structures [36] . RSRC2 , a member of the same gene family , has been proposed as a tumor suppressor gene in esophageal carcinogenesis[37] . Increased expression of RSRC2 has been observed in neuroblastomas harboring somatic gain of chromosome 12q [38] , and a MIER2-RSRC1 fusion has been observed in prostate cancer [39] . Taken together , existing studies suggest that RSRC1 may play an important role in both neural stem cell proliferation and cancer development . The MLF1 gene , also mapped to 3q25 , encodes an oncoprotein that is thought to play a role in the phenotypic determination of hematopoetic cells . It was first identified as the C-terminal partner of the leukemic fusion protein nucleophosmin ( NPM ) -MLF1 that resulted from a t ( 3;5 ) ( q25 . 1;q34 ) chromosomal translocation [40] . MLF1 is overexpressed in more than 25% of MDS-associated cases of AML , in the malignant transformation phase of MDS , and in lung squamous cell carcinoma [41 , 42] . MLF1 overexpression is thought to suppress a rise in the CDK inhibitor CDKN1B , preventing the activation of Epo-activated terminal differentiation pathway and promoting proliferation [43] . MLF1 is expressed in a wide variety of tissues , shuttles between the cytoplasm and the nucleus , and has also been shown to reduce proliferation by stabilizing the activity of TP53 by suppressing its E3 ubiquitin ligase , COP1 [44] . These data suggest that MLF1 may play both a tumor suppressing and an oncogenic role depending on the biological context . Since both RSRC1 and MLF1 have been previously implicated in cancer , we investigated the 3q25 locus in more detail . Based on GTEx data , rs6441201 is a multi-tissue eQTL for both RSRC1 and a recently discovered long non-coding RNA LOC100996447 at 3q25 . While we did not observe a genotype-expression correlation for RSRC1 or LOC100996447 in neuroblastoma cells , we cannot rule out the possibility that variants at 3q25 influence expression of RSRC1 and/or LOC100996447 genes early in tumorigenesis within developing neural crest cells . However , MLF1 expression was observed in nineteen distinct neuroblastoma cell lines interrogated in this study , with the highest expression in cells homozygous for the risk allele at rs6441201 . Silencing of MLF1 resulted in significant growth inhibition in four distinct neuroblastoma cell lines . Taken together , these data are consistent with the hypothesis that MLF1 promotes neuroblastoma tumorigenesis , and that the 3q25 risk alleles are associated with growth advantage through increased MLF1 expression . Given that the observed cell growth phenotype was independent of rs6441201 genotype , alternative mechanisms driving MLF1 expression to promote neuroblastoma cell growth likely exist . In conclusion , here we refine previously reported susceptibility loci , identify common variation at chromosome 3q25 and 4p16 associated with neuroblastoma , and provide insight into potential causal variants at the newly identified susceptibility loci . The newly associated variants at 4p16 are located within CPZ , and the top associated SNP is a meQTL for sites located directly within the 5′ UTR of CPZ . The associated variants at 3q25 appear to function in cis to alter MLF1 expression in neuroblastoma . Based on initial functional studies , it is likely that germline susceptibility alleles at 3q25 play and important role in both initiation and disease progression . Ongoing studies will further elucidate the role of both CPZ and MLF1 in neuroblastoma tumorigenesis .
A primary European-American cohort of 2 , 101 cases and 4 , 202 matched controls were assayed with Illumina HumanHap550 v1 , Illumina HumanHap550 v3 , and Illumina Human610 SNP arrays as previously described [17] . Genotypes were phased using SHAPEIT [23] v2 . r790 and data from 1000 Genomes Phase 1 Release 3 . Subsequently , imputation was performed genome-wide using IMPUTE2 [24] v2 . 3 . 1 for all SNPs and indel variants annotated in 1000 Genomes Phase I Release 3 . To minimize potential errors in phasing and imputation performed genome-wide , we employed a genome-tiling approach . Each position in the genome was covered by a minimum of three tiles ( sliding windows ) . Variants with MAF <1% and/or IMPUTE2-info quality score <0 . 7 were removed . Testing for association with neuroblastoma was performed under an additive genetic effect model using the frequentist likelihood score method implemented in SNPTEST [25] v2 . 4 . 1 . After genome-wide assessment , regions with p < 5 . 0 x 10−7 were re-imputed without tiling and tested for association in a similar manner . Genotypes for a previously described African-American replication cohort of 365 cases 2491 controls [28] were imputed and tested for neuroblastoma association using the same analytic pipeline . Statistical adjustment for gender was performed in both cohorts . For population stratification adjustment , the first 20 multidimensional scaling ( MDS ) components were included as covariates in the European-American cohort , while a measure of African admixture as estimated by the ADMIXTURE software program was used in the African-American cohort . Genotyping of the top associated SNPs at MLF1 ( rs6441201 ) and RSRC1 ( rs3796725 and rs3796727 ) was performed using TaqMan SNP genotyping assays ( Life Technology ) . The Italian cohort was comprised of a total of 432 neuroblastoma cases and 780 controls . The replication cohort from the United Kingdom included 371 cases and 1 , 122 controls in total . Association with neuroblastoma was assessed using an additive genetic effect model of the frequentist likelihood score method implemented in SNPTEST [25] v2 . 4 . 1 in the same manner as the discovery cohort . DNA from 769 children without cancer was extracted from blood and genotyped using Illumina HumanHap550 v1 , Illumina HumanHap550 v3 , and Illumina Human610 SNP arrays . DNA from the same individuals was also profiled for genome-wide methylation using Illumina 450K methylation arrays . Genotypes were phased using SHAPEIT [23] v2 . r790 and data from 1000 Genomes Phase 1 Release 3 . Subsequently , imputation was performed genome-wide using IMPUTE2 [24] v2 . 3 . 1 for all SNPs and indel variants annotated in 1000 Genomes Phase I Release 3 . Principal component analysis ( PCA ) was performed based on genotype data and ancestry was inferred . A threshold of 0 . 9 was applied to rs3796727 imputed genotype probabilities for the purpose of methylation association testing; genotypes from individuals not reaching this threshold were excluded . Association testing was subsequently performed using linear regression with the R software . Meta-analysis was performed using the inverse-variance method within the METAL [29] software package , and a fixed-effects model was assumed . Genome-wide methylation profiles were generated from gDNA isolated from peripheral blood mononuclear cells from a total of 854 subjects recruited by the Center for Applied Genomics ( CAG ) at the Children’s Hospital of Philadelphia ( CHOP ) on the Infinium HumanMethylation450 BeadChip Kit according to the manufacturers' protocols . and analyzed as Methylation data were exported from GenomeStudio and subjected to quantile color balance adjustment , background level correction , and simple scaling normalization as described previously [30] . Principle component analysis identified 425 subjects of European ancestry , 374 African Americans , 20 East Asians , and 24 Hispanics among these subjects . Methylation probes known to overlap with common SNPs , were identified and removed using the IMA R package . M-values ( the log2 ratio between the methylated and unmethylated probe intensities ) were extracted and stored as a matrix . Additive genotypes at rs3796727 for subjects of European ancestry were extracted from existing genotyping data using PLINK . There are a total of 402 subjects of European ancestry without missing genotype at rs3796727 and extreme outlier values of methylation M-values ( ≥median M-value of the genotype group±3 s . d . ) . Methylation data in gene CPZ were analyzed as the response variable in a linear regression , with genotype at as the predictor variable among these 402 subjects . Sex , age , and 10 genotype-derived principle components were included as covariates . Linear regression and generation of boxplots was performed using base packages in R . Genome-wide mRNA expression profiling in neuroblastoma cell lines was performed using the Illumina WG-6 expression array according to the manufacturer’s specifications . Data were normalized using the average normalization method provided in Illumina GenomeStudio software . ANOVA test was performed at the gene level to assess differential expression in cell lines . P < 0 . 05 was considered significant . Data is available from the Gene Expression Omnibus ( GEO ) database ( Accession: GSE78061 ) . TaqMan Gene expression assays for MLF1 ( Hs00963682_m1 ) , RSRC1 ( HS00963694_m1 ) and HPRT ( Hs02800695_m1 ) were purchased through Life Technologies . Reactions were set up in triplicate . Starting with 200 ng RNA , reverse transcription was performed followed by 1:4 dilution and 2 ul of cDNA was subsequently used in a 10-μl reaction with 1× TaqMan Universal PCR Master Mix ( Life Technologies ) . Standard curves were generated using serial dilutions of cDNA from the neuroblastoma cell line Kelly , produced in the same RT reaction as the experimental samples . Samples were amplified on an Applied Biosystems 7900HT Sequence Detection System using standard cycling conditions , and data were collected and analyzed with SDS 2 . 3 software . MLF1 and RSRC1 expression levels were normalized to HPRT expression . Neuroblastoma cell lines were grown in T75 flasks under standard cell culture conditions . Cells were plated into 6 well plates for transfection with siRNA , 2 wells per target for protein analysis . Replicate samples were pooled on collection . Whole-cell lysates were extracted with 100 μl of protein lysis buffer containing Tris Base ( 25mM ) , NaCl ( 150 mM ) , EGTA and EDTA ( 1 mM each ) , NaF ( 10 mM ) DTT ( 1 mM ) , Triton X-100 ( 1% ) , and protease/phosphatase inhibitors ( Cell Signaling , #5872 ) on ice for at least 30 minutes before brief sonication . After 15 min of centrifugation at 4°C , the supernatant was removed , and protein quantification was performed using the Pierce BCA Protein Assay Kit ( Life Technologies , 23225 ) . Lysates ( 12 μg ) were separated on 10% Criterion TGX gels ( BioRad ) and were transferred to PVDF membranes . Membranes were washed and incubated with antibodies directed against MLF1 ( Abcam , ab70211 ) , RSRC1 ( Abcam , ab106650 ) and Ku80 ( Cell Signaling , 2753 ) . All blocking and antibody dilution was performed in 5% milk in TBST . For routine maintenance , cells were grown in RPMI 1640 complete medium ( Gibco , 22400 ) containing 10% FBS ( Hyclone , SH 30073–03 ) , 1× antibiotic antimycotic ( Gibco , 15240–062 ) and 2 mM l-glutamine ( Gibco , 25030 ) . On day 0 , cells were seeded in triplicate into antibiotic-free medium in 96-well RT-CES plates ( ACEA ) . On day 1 , using DharmaFECT 1 ( Dharmacon , T-2001-03 , 0 . 1% ) , cells were transiently transfected with 25 nM of either a non-targeting negative control siRNA ( Dharmacon , D-00810-10-20 ) or pooled siRNA directed against MLF1 ( L-019478-00-0005 ) or RSRC1 ( L-028584-01-0005 ) . Real-time cell growth was monitored every hour for at least 96 h using the RT-CES system , as previously described . Data presented are representative of at least three independent experiments . To monitor efficiency of MLF1 and RSRC1 knockdown , transfection was performed as described , and RNA was isolated 48 hours later using the Qiagen mini extraction kit . Total RNA ( 200 ng ) was primed with oligo ( dT ) and reverse transcribed using SuperScript First Strand Synthesis System for RT-PCR ( Life Technologies ) . Quantitative RT-PCR using TaqMan gene expression assays ( ABI ) was performed as described above . Similarly , protein was isolated 72 hours after transfection to monitor MLF1 and RSRC1 protein knockdown using Western blot analysis as described . Variants directly genotyped , or imputed from the 1000 Genomics phase 1 release 3 data with discovery p-value < 10−4 , MAF > 0 . 005 , and info score > 0 . 5 were annotated and ranked based on a DNase I hypersensitivity data , evolutionary conservation , transcription factor binding site scores , and Roadmap Epigenomics data . Conservation scores were computed as the average of the phastCons46way Placental UCSC conservation track score for all bases from the −10 position to the +10 position surrounding each candidate variant . A DNase I hypersensitivity score was calculated by counting the number of sequencing tags from the −100 position to the +100 position around each candidate variant in ENCODE data for the neuroblastoma cell line , SK-N-SH . Scanning for transcription factor binding motifs was performed using a custom implementation of the MATCH algorithm[45] using JASPAR 2014[46] position weight matrices ( PWMs ) as input . Briefly , to quantify the conservation of position i in a PWM described by a frequency matrix , fi , B , the information vector was computed as follows: I ( i ) =∑B∈ ( A , C , G , T ) fi , Blog2 ( 4fi , B ) For a given input sequence , bi , an absolute information-weighted match score was computed as Score=∑i=1LI ( i ) fi , bi and a normalized matrix similarity score ( mSS ) was computed as previously described . This scan was completed both for the entire human reference genome ( hg19 ) and a modified version of the reference genome ( hg19_alt ) , where each reference base was replaced by its alternative base at each SNP position . A match was called for a PWM if the mSS was greater than 0 . 8 for either hg19 or hg19_alt at a given position overlapping a SNP . At these positions , an mSS difference ( delta-nrm ) and an absolute score difference ( delta-abs ) were computed between hg19_alt and hg19 as two separate metrics to quantify the predicted effect of each SNP on transcription factor binding .
The URLs for data presented herein are as follows: 1000 Genomes Project , http://www . 1000genomes . org LiftOver , http://genome . ucsc . edu/cgi-bin/hgLiftOver SHAPEIT , https://mathgen . stats . ox . ac . uk/genetics_software/shapeit/shapeit IMPUTE2 , http://mathgen . stats . ox . ac . uk/impute/impute_v2 SNPTEST , https://mathgen . stats . ox . ac . uk/genetics_software/snptest/snptest LocusZoom , http://csg . sph . umich . edu/locuszoom METAL , http://www . sph . umich . edu/csg/abecasis/metal
|
Neuroblastoma is an embryonal tumor of the developing sympathetic nervous system that accounts for 12% of childhood cancer deaths . Approximately 1–2% of cases are inherited in an autosomal dominant fashion . These familial cases often harbor germline mutations in ALK or PHOX2B . However , the vast majority of neuroblastomas appear to arise sporadically . We are studying sporadic neuroblastoma through an ongoing genome-wide association study ( GWAS ) . To date , this effort has identified single nucleotide polymorphisms ( SNPs ) within or upstream of CASC15 and CASC14 , BARD1 , LMO1 , DUSP12 , HSD17B12 , DDX4/IL31RA , HACE1 , LIN28B , and TP53 , along with a common copy number variation ( CNV ) within NBPF23 at chromosome 1q21 . 1 , each being highly associated with neuroblastoma . Here , we report on genome-wide association study ( GWAS ) comprising 3 , 264 neuroblastoma patients and 8 , 598 control subjects . We identify two new association signals at 3q25 and 4p16 ( 3q25: rs6441201 combined P = 1 . 2x10-11 , Odds Ratio 1 . 23 , 95% CI:1 . 16–1 . 31; 4p16: rs3796727 combined P = 1 . 26x10-12 , Odds Ratio 1 . 30 , 95% CI: 1 . 21–1 . 40 ) . The 3q25 signal resides upstream of the MLF1 gene and the 4p16 signal maps to the CPZ gene . We further demonstrate that neuroblastoma cells homozygous for the risk allele at 3q25 express higher levels of MLF1 and that silencing of MLF1 in neuroblastoma cells results in significant growth inhibition .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods",
"Web",
"resources"
] |
[
"genome-wide",
"association",
"studies",
"medicine",
"and",
"health",
"sciences",
"biological",
"cultures",
"blastomas",
"cancers",
"and",
"neoplasms",
"variant",
"genotypes",
"neuroblastoma",
"cells",
"african",
"americans",
"genetic",
"mapping",
"ethnicities",
"oncology",
"methylation",
"genome",
"analysis",
"cell",
"cultures",
"research",
"and",
"analysis",
"methods",
"genomic",
"signal",
"processing",
"neuroblastoma",
"chemistry",
"genetic",
"loci",
"people",
"and",
"places",
"signal",
"transduction",
"cell",
"biology",
"heredity",
"genetics",
"biology",
"and",
"life",
"sciences",
"chemical",
"reactions",
"cultured",
"tumor",
"cells",
"genomics",
"physical",
"sciences",
"population",
"groupings",
"cell",
"signaling",
"computational",
"biology",
"human",
"genetics"
] |
2017
|
Common variants upstream of MLF1 at 3q25 and within CPZ at 4p16 associated with neuroblastoma
|
New architectures of multilayer artificial neural networks and new methods for training them are rapidly revolutionizing the application of machine learning in diverse fields , including business , social science , physical sciences , and biology . Interpreting deep neural networks , however , currently remains elusive , and a critical challenge lies in understanding which meaningful features a network is actually learning . We present a general method for interpreting deep neural networks and extracting network-learned features from input data . We describe our algorithm in the context of biological sequence analysis . Our approach , based on ideas from statistical physics , samples from the maximum entropy distribution over possible sequences , anchored at an input sequence and subject to constraints implied by the empirical function learned by a network . Using our framework , we demonstrate that local transcription factor binding motifs can be identified from a network trained on ChIP-seq data and that nucleosome positioning signals are indeed learned by a network trained on chemical cleavage nucleosome maps . Imposing a further constraint on the maximum entropy distribution also allows us to probe whether a network is learning global sequence features , such as the high GC content in nucleosome-rich regions . This work thus provides valuable mathematical tools for interpreting and extracting learned features from feed-forward neural networks .
Multilayer artificial neural networks ( ANNs ) are becoming increasingly important tools for predicting outcomes from complex patterns in images and diverse scientific data including biological sequences . Recent works have applied multilayer networks , also called deep learning models , to predicting transcription factor ( TF ) binding [1 , 2] and chromatin states [3–5] from DNA sequence , greatly advancing the state-of-the-art prediction rate in these fields . The success of these multilayer networks stems from their ability to learn complex , non-linear prediction functions over the set of input sequences . The main challenge in using deep learning currently resides in the fact that the complexity of the learned function coupled with the typically large dimension of the input and parameter spaces makes it difficult to decipher which input features a network is actually learning . In spite of this challenge , the rise of deep learning has spurred efforts to interpret network prediction . While interpretation methods have been proposed in the fields of computer vision and natural language processing , we focus on genomics and specifically on methods for identifying features in individual input sequences used by an ANN for classification . To the best of our knowledge , two classes of interpretation methods have been proposed to address this problem: the first class of interpretation methods measures network feature importance by expressing the change in the network output for two distinct inputs as a sum of importance values assigned to the input units that encode biological sequence . One such decomposition assigns to each input unit importance given by its term in the 1st order Taylor approximation of the change in the network output when the input units encoding a specific sequence are set to zeros . This method of assigning network importance to DNA sequence is called a Saliency Map by Lanchantin et al . [6] who adapted the method from the computer vision Saliency Map [7] . The DeepLIFT interpretation method uses an alternative approach for assigning input sequence importance based on comparing network activations elicited by the input sequences to those of a “reference” network input [8] . This approach to assigning importance has been motived as an approximation to Shapley values , which describe distribution of credit in game theory [9] . The second class of interpretation method , called in silico mutagenesis ( ISM ) , measures changes in network output produced by simulated point mutations . Flexibility in the types of mutations performed means that ISM can , in principle , reveal the network’s dependence on sequence in detail . However , computational cost limits the number and type of progressive mutations that can be performed . As a result , to investigate learned network features using ISM , one must employ prior notions of important features to design a manageable number of sequential mutations for testing . In this paper , we use the rigorous formalism of statistical physics to develop a novel method for extracting and interpreting network-learned sequence features . The method makes direct reference to the nonlinear function learned by the network by sampling a maximum entropy distribution over all possible sequences , anchored at an input sequence and subject to constraints implied by the learned function and by the background nucleotide content of the genome from which the network’s input sequences are derived . To extract learned features from inputs , we study two complementary quantities derived from sequences sampled from the constrained maximum entropy ( MaxEnt ) distribution via Markov Chain Monte Carlo ( MCMC ) : ( 1 ) a local profile of nucleotide contents for revealing sequence motifs , and ( 2 ) a feature importance score based on the sample variance of a summary statistic for focusing on a particular sequence characteristic of interest . The latter score directly measures the effect of a global sequence feature , such as GC content , on the non-linear function learned by the network , and it can be used to rank global features , thereby answering questions about the relative importance of such features in the context of network prediction . Our approach can be viewed as a compromise between the two classes of interpretation method described above , extracting features by examining several sequences , instead of just two as in the first class , and eliminating ISM’s need for a priori specification of sequential mutations . Importantly , our method is distinguished from other previous approaches in that , rather than assigning an importance score to each base of a given input sequence , our method reveals features by sampling unseen sequences that are assessed by the trained network to be similar to the original input . We apply our method , termed MaxEnt interpretation , to three separate deep neural networks and compare with the first class of interpretation methods: DeepLIFT and Saliency Map . The first network is a simple yet instructive example , and it demonstrates how the DeepLIFT and Saliency Map methods can encounter difficulties , while the MaxEnt method successfully captures the logic learned by the network . The remaining two networks are trained to predict transcription factor ( TF ) binding activity from CTCF ChIP-seq [2] and nucleosome positioning from chemical cleavage nucleosome mapping data [10] , thus putatively learning the CTCF binding motifs and nucleosome positioning signals encoded in DNA , respectively . In the motif discovery task , while all methods give good results , correctly localizing the learned CTCF motifs in most cases , our MaxEnt approach achieves better agreement with motif positions called by the conventional motif discovery programs MEME and FIMO . In the task of extracting nucleosome positioning signals , MaxEnt surpasses DeepLIFT and Saliency Mapping in detecting a learned network preference for G/C and A/T nucleotides at positions separated by 10 base pairs ( bps ) . Furthermore , we demonstrate the use of a global sequence feature score , unique to our method , to estimate the fraction of nucleosomal sequences for which the GC content is an important feature for network classification . We do not compare with ISM because , as discussed above , it is less general than other methods and requires the specification of the types of mutations to test .
Consider a trained , multilayer feed-forward neural network that takes a length L genomic sequence as input and performs a classification task , assigning the sequence to one of K classes indexed by {0 , 1 , … , K-1} . The network consists of a stack of layers , each of which contains a real-valued vector whose entries are called the activations of the layer’s units . The stack starts with an input layer whose activations represent the genomic sequence; activations of each subsequent layer are determined by applying a trained non-linear transformation to the proceeding layer . Activations of units in the output layer encode the predicted probabilities of the K classes , thereby assigning the input sequence to the class whose output unit activation is the largest . ( S1 Text describes the types of layers used in this work ) The standard motivation behind multilayer networks is that intermediate layers may learn to recognize a hierarchy of features present in the set of inputs with features becoming more abstract with the depth of the intermediate layer . Since changes in the features detected by a layer are encoded in changes in the intermediate layer’s vector of activations , we propose that it is possible to identify learned features by looking for commonalities among the set of all input sequences that approximately preserve that layer’s vector of activations . While this approach could be applied to identifying learned features of any intermediate layer , this work focuses on features learned by the penultimate layer , the layer making direct connections to the output layer . Penultimate layer features are interesting for two reasons . First , we are interested only in input sequence patterns that elicit an intermediate layer representation relevant to the classification task and discard sequence variations irrelevant to classification . Since the non-linear functions computed by intermediate layers are , in general , many-to-one mappings , one important role of a layer is to identify which differences in the preceding layer’s activations are irrelevant to the learned classification . Because intermediate layer activations are calculated from the input sequence by composing such non-linear functions , the number of identified classification-irrelevant differences in inputs should increase with the intermediate layer depth , making the penultimate layer the intermediate layer least affected by classification-irrelevant features . Second , the network output layer is either a logistic or a softmax classifier applied to penultimate activations; by uncovering features learned by the penultimate layer , we are thus finding the sequence patterns directly used by these output layer classifiers to make predictions . To formalize the search for new input sequences that approximately preserve a given penultimate activation , let the vector x0 represent an input genomic sequence . The entries in x0 encode the presence of one of the nucleotides A , C , G , T at a certain position in the sequence . Let Φ ( x0 ) denote the vector of penultimate layer activations elicited by x0 . For notational convenience , assume the trained network has assigned x0 to the 0th class . We measure the extent to which an arbitrary input sequence x reproduces the penultimate activation Φ ( x0 ) by weighing the set of all length L genomic sequences with the probability mass function ( PMF ) px0 that is most similar to a pattern-agnostic PMF q , subject to a constraint on the average distance to Φ ( x0 ) . More precisely , we define px0=argminpDKL ( p∥q ) subjecttoEx[d ( Φ ( x0 ) , Φ ( x ) ) ]=D , ( 1 ) where DKL ( p ‖ q ) is the Kullback-Leibler ( KL ) divergence between p and q , Ex denotes expectation calculated when x is distributed according to p , d is a distance metric on the set of penultimate activation vectors , and D is a positive constant , with smaller values corresponding to a more exact reproduction of the activation vector Φ ( x0 ) . The PMF q describes prior beliefs about the background nucleotide content of the genomic regions from which network inputs are derived , and the fact the px0 minimizes DKL ( p ‖ q ) subject to the constraint ensures that differences between px0 and these prior beliefs arise from the need to reproduce the sequence features encoded in Φ ( x0 ) . We take q be a product of L identical single nucleotide distributions with probabilities of G/C and A/T chosen to reflect the genome-wide GC content . In this case , the solution to ( 1 ) is px0 ( x ) =1Ze−βd ( Φ ( x0 ) , Φ ( x ) ) +μN ( x ) ( 2 ) where Z is a normalization constant , β is a Lagrange multiplier whose value is chosen to yield the desired value of D , N ( x ) is the number of G and C nucleotides in sequence x , and μ=log ( c1−c ) where c is the GC content of the genome ( S2 Text ) . When μ = 0 , px0 ( x ) is determined solely by the learned function Φ , and Eq ( 2 ) is the maximum entropy distribution over all length L sequences subject to the constraint in ( 1 ) . For simplicity , we use the term MaxEnt samples to refer to the samples from px0 for any value of μ . We use a weighted Euclidean metric for d , d ( Φ ( x0 ) , Φ ( x ) ) = ( ∑i ( Φi ( x0 ) −Φi ( x ) ) Wi2 ( Φi ( x0 ) −Φi ( x ) ) ) 12 ( 3 ) where our choice of Wi depends on the type of classification task . For binary classification , Wi = w0 , i , the weight connecting the ith penultimate unit to the output unit encoding the assigned class label of x0 ( when there is only one output unit , Wi is the weight of connection to this unit ) . For multiclass classification , Wi = w0 , i − wk , i , where k ∈ {1 , 2 , …K−1} is a user-specified class . This choice of Wi corresponds to weighting each penultimate activation by its effect on the log ratio of predicted class 0 and class k probabilities ( S1 Text ) . The mean distance D is an increasing function of 1/β , whose scale is set by nearest-neighbor distances among penultimate activations , as measured by the metric ( 3 ) . When β is large , D approaches 0 , the PMF over the set of penultimate activations is a single spike at Φ ( x0 ) , and px0 consists of a ( relatively ) small number of non-zero probability masses on sequences x that exactly reproduce Φ ( x0 ) . Conversely , decreasing β smooths the PMF over penultimate activations and causes px0 to shift probability mass onto sequences that yield penultimate activations similar to Φ ( x0 ) . When β = 0 , px0 and q are identical , D is the expected distance under the distribution q , and px0 contains no information on the features encoded in Φ ( x0 ) . This intuition informs one method for choosing β ( and implicitly D ) : select β so that the PMF over the set of penultimate activations has small width relative to an empirical distribution of penultimate activations , while still assigning appreciable probability to sequences with penultimate activations near Φ ( x0 ) . Alternatively , because a sufficiently large value of β effectively fixes the nucleotide content at certain indices in sequences sampled from px0 , one can examine the samples from distributions at different values of β to uncover a hierarchy of important features in xo . We give examples of both methods in the following sections , where we sample the distribution ( 2 ) using MCMC ( Methods ) . Fig 1 , illustrates the sampling of sequences x according to their similarity to data set element x0 in the space of penultimate layer activations . We extract features of network input x0 , captured by penultimate activation Φ ( x0 ) , by examining several statistics of the MCMC samples from px0 . In our first example , the dimension of the input space is low enough to directly visualize the empirical distribution of samples . For higher dimensional input spaces , we summarize the distribution by plotting sample single nucleotide frequencies at each genomic index and also by examining the variance of linear combinations of nucleotide indicator variables that serve to define sequence features of interest . Plots of single nucleotide frequencies reveal important features by illustrating the extent to which preserving penultimate activation forces the single nucleotide distribution of px0 to be away from that of q . Large divergence of sample nucleotide frequencies from q signals importance and determines which nucleotide substitutions dramatically affect the penultimate layer representation . By contrast , if the sampled nucleotide distribution of px0 is similar to that of q at a given base position , and if we assume that the content of the position is independent of other positions under px0 , then this position is irrelevant for determining Φ ( x0 ) and the classification of x0 . To quantify the importance of sequence features , in a way that accounts for interactions among base positions , we define an input-wide sequence feature V as a function that maps an input sequence to a weighed sum of indicator variables for specific nucleotide content: V ( x ) =∑i=1LciIi ( xi ) ( 4 ) where xi denotes the nucleotide at index i in x , Ii ( ∙ ) is the indicator variable for one of a set of nucleotides at index i and ci is a real valued weight . We define the concordance of sequence x with the input-wide feature V to be V ( x ) . For example , if we are interested in the importance of GC content , V would be the sum of indicator variables for S ( G or C nucleotide ) at each input index , and x would have large concordance with this input-wide sequence feature when it contains many G and C nucleotides . To relate changes in feature concordance to changes in penultimate layer activation , let Xv denote the set of length L sequences x , such that V ( x ) = v . The quantity f ( v ) = ( ∑x∈Xvq ( x ) e−βd ( Φ ( x0 ) , Φ ( x ) ) ∑x∈Xvq ( x ) ) ( 5 ) is the mean value of e−βd ( Φ ( x0 ) , Φ ( x ) ) under PMF q , conditioned on V ( x ) = v . The rate of decay of f ( v ) from its maximum measures the dependence of e−βd ( Φ ( x0 ) , Φ ( x ) ) on feature concordance V ( x ) . To see this , observe that when V sums indicator variables for nucleotides important in eliciting the penultimate activation Φ ( x0 ) , the exponential factor e−βd ( Φ ( x0 ) , Φ ( x ) ) will , on average , decay as V ( x ) deviates from V ( x0 ) . More rapid decay corresponds to greater importance , since in this case smaller changes in V ( x ) suffice to produce large changes in Φ ( x ) . By contrast , when V sums indicator variables for only those nucleotides unimportant in eliciting Φ ( x0 ) , the factor e−βd ( Φ ( x0 ) , Φ ( x ) ) does not decay with changes in V ( x ) . In S2 Text , we approximate the decay of f from its maximum as f ( v ) ∝e−12δ ( v−v* ) 2 ( 6 ) where v* maximizes f ( v ) and δ≡1s2−1σ2 ( 7 ) where s2 is the variance of V ( x ) under the PMF px0 and σ2 is the variance of V ( x ) under PMF q . The variance s2 is estimated from MCMC samples while the variance σ2 can be calculated explicitly from q . Larger values of δ correspond to more rapid decay of f ( v ) , signaling greater input-wide feature importance . Our derivation of the proportionality ( 6 ) requires that the marginal distributions of V ( x ) when x is distributed according to q or px0 are approximately normal . Approximate normality of V ( x ) when x is distributed according to q is guaranteed by the Lindeberg version of the Central Limit Theorem , provided that V sums indicator variables at a large number of base positions with weights roughly equal in magnitude . Approximate normality of V ( x ) when x is distributed according to px0 can be checked directly by estimating V ( x ) from MCMC samples . We have checked that these normal approximations are valid when the number of base positions considered by the function V is large , explaining our choice of the name input-wide sequence features . In Application 3 below , we consider two uses of the importance measure δ . First , we choose the weights ci and indicators Ii ( ∙ ) , so that the resulting input-wide feature V measures the importance of GC content for a network predicting nucleosome positioning . Second , we use MCMC samples from ( 2 ) to find sets of weights ci that approximately maximize δ . When V captures an important input-wide feature , the exponential factor e−βd ( Φ ( x0 ) , Φ ( x ) ) in ( 2 ) should make the marginal distribution of V under px0 much narrower than the marginal distribution of V under q; that is , s2 ≪ σ2 . In this case , we can thus approximate , δ≈1s2 . ( 8 ) Under this approximation , the most important features are given by the lowest variance principal components ( PCs ) of our MCMC samples . Examining the elements of these low variance PC vectors reveals important input-wide features . ANN interpretation methods , such as Saliency Map and DeepLIFT , that assign a real-valued importance score to each input unit provide intuitive pictures that are easy to understand , but must confront the challenge of summarizing learned nonlinear interactions between base positions using a base-wise score . To illustrate the practical consequences of these issues , we trained ANNs on an artificial data set of DNA dinucleotides and applied the MaxEnt , Saliency Map and DeepLIFT interpretation methods . Class labels were assigned to each of the 16 possible dinucleotides according to an XOR logic , where sequences ( with positions indexed by 0 and 1 ) were assigned to class 0 unless one , but not both , of the following conditions was satisfied , in which case class 1 was assigned: We represented sequences with a one-hot encoding scheme and chose a simple convolutional architecture with 2 convolutional filters of stride 1 and taking a single base of the dinucleotide as input , followed by a layer of two fully connected units and then a single output unit indicating the predicted class label . Rectified-linear units ( ReLU ) were used throughout the network , except for the output which was modeled using a sigmoid function ( S1 Text ) . We obtained 30 of these convolutional architectures trained to achieve 100% classification accuracy on the set of all possible dinucleotide inputs ( Methods ) . Fig 2 shows the results of interpretation analysis on the AA and GG inputs in class 1 for each of the 30 models achieving 100% validation accuracy . Although each network represents the same classification rule , the interpretation results of Saliency Map and DeepLIFT show model dependence in the sign of their interpretation score , indicating in some cases that a nucleotide is evidence for the correct class label and in other cases that the same nucleotide is evidence against this class label ( Fig 2A–2D ) ( see S3 Text for description of our application of Saliency Map and DeepLIFT ) . In contrast , our MaxEnt interpretation illuminates , in almost every case , how the trained networks implement the XOR logic defined above , indicating that the GG input is similar to CG and that the AA input is similar to any input with A or T in the 0th position and not G in the 1st position ( Fig 2E and 2F ) . For this analysis , we sampled the distribution ( 2 ) with μ = 0 and β chosen based on the distribution of penultimate layer activation associated with the 16 dinucleotide inputs ( Methods ) . S1 Fig shows similar results for the other dinucleotide inputs . Fig 2 highlights a key difference that distinguishes the MaxEnt interpretation approach from Salience Map and DeepLIFT . By replacing base-position scores with samples from a distribution , MaxEnt interpretation is able to capture nonlinear classification rules that escape the other methods . The cost of this extra flexibility is some additional effort in assigning meaning to the MaxEnt samples . We applied MaxEnt interpretation to a network trained on a benchmark motif discovery data set constructed by [2] from ENCODE CTCF ChIP-seq data [11] . CTCF is a well-studied DNA binding protein with important transcription factor and insulator functions [12 , 13] . In this motif discovery task , the network distinguished elements of the positive class , consisting of 101 base-pair ( bp ) sequences centered on ChIP-seq peaks , from elements of the negative class consisting of positive class sequences shuffled to maintain dinucleotide frequency . We represented network inputs with a one-hot encoding scheme and trained an architecture consisting of a convolutional layer of 64 convolutional filters each with a stride of 1 and taking 24 bps as input , followed by a layer of 100 units fully connected to the preceding layer and a two unit softmax output layer . ReLUs were used in all layers preceding the output ( S1 Text ) . The trained network performed well , achieving a mean area under the receiver operating characteristic ( AUROC ) of 0 . 978 with standard deviation 0 . 001 in 5-fold cross-validation ( Methods ) . We picked a neural network trained on a random fold of the cross-validation , selected 2500 random sequences from all correctly classified CTCF-containing sequences in the test set , and applied MaxEnt , DeepLIFT and Saliency Map interpretation methods . Our application of MaxEnt to this network used β = 400 , chosen by examining samples collected at a range of β′s for a few network inputs and selecting the smallest β sufficient to fix the nucleotide content at positions where MaxEnt marginal distributions signaled greatest importance . Because single nucleotide frequencies for the data set were nearly uniform ( P ( A ) = P ( T ) = 0 . 27 and P ( C ) = P ( G ) = 0 . 23 ) , we set μ = 0 when sampling from the distribution ( 2 ) . Fig 3A and 3B show nucleotide frequencies as a function of base index for MCMC MaxEnt samples associated with two input sequences . In both cases , the location of the motif identified by the network was indicated by an interval of single nucleotide frequencies that diverged dramatically from the uniform distribution over nucleotides implied by the distribution ( 2 ) for sequence locations with little effect on penultimate layer activations . Sequence logos were generated from the nucleotide frequencies on these intervals using WebLogo [14] . We confirmed that the discovered motifs in Fig 3A and 3B correspond to the canonical CTCF motif and its reverse complement by using the motif database querying tool Tomtom [15] ( Methods , S2 Fig ) . DeepLIFT and Saliency Map interpretation of these inputs yielded visually similar results ( S3 Fig ) . However , MaxEnt single nucleotide frequencies provide direct interpretation as motif position-specific scoring matrices utilized by other bioinformatics tools and thus provide advantages over Saliency Map and DeepLIFT base-wise scores . To make a more global comparison of interpretation methods , we calculated the distribution of relative distances from motif positions called using each interpretation method to motif positions identified by the conventional motif discovery programs MEME and FIMO [16 , 17] . Combined application of MEME and FIMO to the 2500 interpreted inputs found a single 19 bp consensus CTCF motif in 1652 of these input sequences ( Methods ) . Using this motif length as a guide , we called MaxEnt motifs by calculating , for each sequence input , the KL divergence of sample nucleotide frequencies at each base from a uniform distribution and finding the 19 bp running window that has the largest average KL divergence . DeepLIFT and Saliency Map motifs were similarly called for each sequence input at the 19 bp widow with largest interpretation score ( see S3 Text for definition ) . Fig 3C shows the empirical distribution of the signed center-to-center distances between network interpretation motifs and MEME/FIMO motifs in the 1652 sequences . Fig 3D shows the cumulative distribution of the same unsigned distances . MaxEnt interpretation gives significantly more motif calls within 0 , 1 , 2 , 3 and 4 bp of MEME/FIMO motifs than Saliency Map and DeepLIFT interpretation . Finally , we tested ANN interpretation methods on a genomic data set where we expected learned sequence features to be more diffuse . We constructed a data set based on the chemical cleavage map of nucleosome dyads in S . cerevisiae [10] . Each input was a 201 bp sequence with positive class elements centered on nucleosome dyads and negative class elements centered on locations uniformly sampled from genomic regions at least 3 bps from a dyad . We chose to allow sampling of negative sequences within the 73 bps of nucleosomal DNA flanking the dyad to encourage the network to learn features that direct precise nucleosome positioning as well as those determining nucleosome occupancy . Our trained network consisted of a convolutional layer with 30 filters , each with a stride of 1 and taking 6 bp windows as input , followed by a 400-unit layer with full connections and a 2-unit output softmax layer . Sigmoid activation functions were used in all layers preceding the output ( S1 Text ) . The trained network performed well , achieving an AUROC of 0 . 956 on the test set ( Methods ) . We applied interpretation methods to 2500 input sequences randomly selected from validation set elements corresponding to nucleosomal sequences correctly classified by the network . MaxEnt interpretation sampled the distribution ( 2 ) using μ = −0 . 49 and β = 40 . 0 . The value of μ was determined using the 38% GC content of the S . cerevisiae genome . The value of β was determined by examining the plots of nucleotide frequencies for a range of β values and selecting the largest value that permitted fluctuation in the nucleotide content at all of 201 bp . S3 Text describes our application of DeepLIFT and Saliency Mapping to this network . Fig 4A shows single nucleotide frequencies of MaxEnt samples for one of the 2500 nucleosomal sequences analyzed . Fig 4B and 4C show the results of DeepLIFT and Saliency Map interpretation for the same network input , respectively . This example shows that MaxEnt interpretation surpasses DeepLIFT and Saliency Mapping in capturing the importance of G/C and A/T nucleotides preferentially positioned at anti-phased 10 bp intervals . To confirm this trend across all interpreted nucleosomal sequences , we calculated for each input the Discrete Fourier Transform ( DFT ) of single nucleotide frequencies of MaxEnt samples , DeepLIFT interpretation scores , and Saliency Map interpretation scores ( S3 Text ) . DFT represents each of these signals as a sum of sinusoidal functions , with each sinusoid described by a period , phase and amplitude of oscillation . The contribution of a sinusoid to the signal is measured by its amplitude relative to the amplitudes at all periods; we normalized the components of each DFT to account for this relative comparison ( Methods ) . Fig 4D shows the average of these normalized amplitudes over the set of all interpreted inputs , confirming that MaxEnt single nucleotide frequencies provide the strongest evidence for the learned 10bp-periodicity preference in nucleotide positioning . Consistent with Fig 4D , 10 bp-periodic signals were found in many individual sets of MaxEnt samples . However , S4 Fig shows counterexamples to this trend , highlighting that the network does not need to detect the pattern to classify a sequence as nucleosomal . S4 Fig also shows a plot of nucleotide frequencies averaged over the set of 2500 input nucleosomal sequences . It is important to recognize that even though these plots of sample nucleotide frequencies illuminate learned features , they do not imply that the periodic features are present in individual MaxEnt samples . Indeed , it has recently been shown that most nucleosomal sequences in S . cerevisiae do not contain significant 10 bp periodicity [18] . To confirm that the MaxEnt samples produced by our method were also not individually enriched for 10 bp periodicity , we calculated the normalized Fourier spectrum of each sample separately and then averaged the amplitudes over all samples associated with the 2500 nucleosomal sequences ( Methods ) . Fig 4D shows that this Fourier amplitude at 10 bp averaged over the pooled MCMC samples is greatly suppressed relative to the Fourier amplitude of nucleotides frequencies averaged over nucleosomal sequences . In this way , MaxEnt samples capture the true nature of the 10 bp periodic feature learned by the network . That is , to be considered similar to an input nucleosomal sequence , it is enough for MaxEnt samples to possess G/C and A/T nucleotides at only some of the “hot-spots” separated by 10 bps; at the same time , averaging over these samples gives a coarse and conceptually useful representation of the learned feature . It is also widely believed that nucleosomal DNA often possesses high GC content relative to the genomic background [19]; we thus explored the importance of GC content to our network’s classification . Fig 5A shows that , while mean GC content of MaxEnt samples generally agreed with the background 38% tuned by our choice of μ , there was also a significant positive correlation between sample mean GC content and GC content of the associated input . The correlation indicated that changes in GC content affected the penultimate layer activations to the extent that samples tended to preserve the GC enrichment or depletion of their associated input . To rigorously measure the importance of the GC content feature , we defined an input-wide sequence feature V that sums the indicators for G or C nucleotides at each of the central 147 bases of the network input . For comparison , we defined “dummy” input-wide features which also sum indicator variables at each of the central 147 bases of network input , but where , at each position , the set of two nucleotides for which the indicator is 1 is uniformly sampled from the list {G , C} , {G , A} , {G , T} , {C , A} , {C , T} , {A , T} . For 1000 inputs chosen at random from the 2500 analyzed , we calculated the feature importance score δ , defined in ( 7 ) , for the GC content feature V and for 300 random variables measuring dummy features . We then computed the percentile of the importance score of the GC content variable in the distribution of importance scores of the dummy feature variables for each input . Fig 5B shows the distribution of these percentiles , with enrichment of nucleosomal sequences near the 100th percentile; setting a threshold at the 90th percentile in the distribution of dummy feature importance scores , we estimate that GC content is a learned network feature of about 26% of the 1000 nucleosomal sequences analyzed . While assigning relative importance to a chosen input-wide feature is useful , we were also interested in automatically discovering the most important input-wide features from MaxEnt samples , without prior specification of the weights ci in ( 4 ) . For this purpose , we chose V to sum over indicator variables for G/C at each position , with the ci’s to be determined . The variance of this V ( x ) , with x distributed according to ( 2 ) , can be written for an arbitrary vector c ≡ ( c1 , c1 , … , cL ) T of weights as Var ( V ( x ) ) =cTSc where S is the covariance matrix of the indicator variables estimated from MCMC samples . Since the approximation ( 8 ) implies that feature importance decreases with increasing variance , we seek weight vectors minimizing Var ( V ( x ) ) under the constraint that c has unit Euclidean norm . Thus , the problem of identifying low-variance input-wide features amounts to selecting low variance principal components ( PCs ) in Principal Component Analysis ( PCA ) using S . The elements of the low variance PCs then give the weights of an input-wide feature V . Moreover , because the PCs are uncorrelated with respect to S , we expect several of the low variance PCs to be interpretable . Fig 5C shows a sharp decrease in the variance of the input-wide features determined by PCA on MaxEnt samples for a single nucleosomal sequence . We empirically observed that this sharp drop , signaling the prominent importance of features corresponding to the lowest variance PC vectors , is typical for our samples . S5 Fig plots the weight vectors obtained from the two lowest variance PC vectors associated with the MaxEnt distribution depicted in Fig 4A . The lowest variance feature concentrates on a spike at the +3 position relative to the dyad . The strong network dependence on this position is also seen in Fig 4A–4C . The second lowest variance feature shows 10 bp periodicity with the weights changing sign roughly every 5 bp . While the pattern is much like that of Fig 4A , it accounts for correlations in nucleotide content , demonstrating that it is the collective alignment of G/C and A/T content with this weight template that changes the network’s penultimate representation . Finally , we demonstrated the utility of these features by constructing a simple 10-nearest neighbor classifier , where we used the lowest variance PCs to compute the inter-sequence distance . Briefly , we randomly selected 1200 correctly classified nucleosomal sequences to which we applied our interpretation method with the values of β and μ given above . For a given nucleosomal sequence , we represented each of its MCMC samples as a 201 dimensional binary vector by evaluating the G/C indicator variable at each base and used the element-wise mean over these vectors to represent the nucleosomal sequence itself as a positive exemplar in the nearest neighbor classification . Likewise , we repeated this task for 1200 correctly classified non-nucleosomal sequences to obtain negative exemplars . We then selected a balanced test set of 10500 sequences that were previously classified correctly by the network and represented each test sequence as a 201 dimensional binary vector indicating the G/C nucleotide composition of its bases . To compute the distance of a test sequence to an exemplar , we projected the vector joining the exemplar and the test sequence onto the space spanned by the exemplar’s 5 lowest variance PC vectors , scaling the projected coordinates by the inverse standard deviation of the associated PC vectors and then computing the Euclidean distance . Test set elements were assigned to the majority class of the 10 nearest exemplars . This simple method yielded a classification accuracy of 76% . For comparison , we repeated this classification replacing the 5 lowest variance PC vectors of each exemplar with 5 mutually orthogonal vectors randomly sampled from the 201 dimensional space ( Methods ) . Using this control , nearest neighbor classification accuracy dropped to 51% . This result thus demonstrates the ability of our interpretation method to extract de novo features used in the neural network’s classification .
Deep neural networks provide researchers with powerful tools for making predictions based on complex patterns in biological sequence . Methods for extracting learned input features from these networks can provide valuable scientific insights , and several efforts in this direction [6 , 8] have made deep learning an even more appealing approach for tackling complex problems in genomics and other scientific disciplines . We have contributed to these efforts by introducing a novel feature extraction method based on sampling a maximum entropy distribution with a constraint imposed by the empirical non-linear function learned by the network . From a theoretical standpoint , this constraint allows the derivation of relationships between the statistics of the sampled distribution and the dependence of network classification on specific sequence features . In particular , we have developed a scheme for assessing input-wide feature importance that has been difficult to measure otherwise with currently available approaches to network interpretation . From a practical standpoint , the MaxEnt approach to feature extraction is distinct from other interpretation methods that assign base-wise importance to sequences . Admittedly , different interpretation schemes may thus have distinct advantages and disadvantages . For example , in Application 1 , the MaxEnt method is able to capture the XOR logic that is learned by a simple ANN , but the same logic is difficult to infer using the methods based on base-wise importance . In Application 2 , all schemes give similar results , but the MaxEnt interpretation method also provides probabilistic position-specific scoring matrices that are commonly used in bioinformatics . However , DeepLIFT and Saliency Map may be preferred in some cases for their computational efficiency . Interpreting an input in Application 2 via DeepLIFT , Saliency Map and MaxEnt takes 0 . 64 ms , 0 . 11 ms , and 181 s , respectively , on a quad-core 3 . 2 GHz Intel CPU , where the clear computational cost of MaxEnt interpretation stems from our MCMC sampling approach . This cost could be mitigated via a parallel implementation of multiple MCMC chains . Finally , Application 3 illustrates a setting in which MaxEnt interpretation surpasses other methods in elucidating the learned features that are consistent with the current understanding of nucleosome positioning [18] . The success of our MaxEnt approach signals that statistical physics may have much to contribute to the task of interpreting deep learning models . Indeed , a central goal of statistical mechanics is to understand constrained MaxEnt distributions of many degrees of freedom that interact according to known microscopic rules . While our formulation addresses an inverse problem of inferring unknown characteristics of network inputs from observed statistics of a constrained MaxEnt distribution , statistical mechanics provides a wide range of tools that could be further explored in this new context . This theoretical paper provides an example of the growing synergy between machine learning and physics towards assessing the role of diffuse and subtle sequence features that direct important biological outcomes , such as the positioning of nucleosomes .
Given a representation x0 of a sequence classified to class 0 by the trained network , we sampled sequences x from the PMF ( 2 ) using a Markov chain Monte Carlo ( MCMC ) method . We initialized the Markov random sequence x at x0 , then repeatedly selected an index i of x with uniform probability and proposed a mutation of nucleotide xi to nucleotide xi* sampled from the set {G , C , A , T}–{xi} with uniform probability . The proposed mutation was accepted with probability given by the Metropolis criterion: Paccept=min ( 1 , e−β[d ( Φ ( xo ) , Φ ( x* ) ) −d ( Φ ( xo ) , Φ ( x ) ) ]+μ[N ( x* ) −N ( x ) ] ) where x* denotes the random variable x with the proposed mutation at index i [20] . To generate the results of Application 1 , we used 50 chains in parallel for each x0 , with each chain constructed from 100 proposed mutations . Each chain was sampled after every proposal . To generate the results of Application 2 , we used 100 Markov chains in parallel for each x0 , with each chain constructed from 3 × 104 proposed mutations . Each chain was sampled every 100 proposals . To generate the results of Application 3 , we used for each x0 a single Markov chain constructed by proposing 1 × 106 mutations . We sampled the chain every 100 proposals . We trained the network architecture using stochastic gradient descent with batch size of 8 and the categorical cross entropy loss function . Our training set was augmented with reverse complement sequences , and gradient descent used a learning rate of 0 . 1 , momentum parameter of 0 . 5 , and L2 weight penalty of 0 . 001 . Training was done with the python package Keras [21] . Normalized Fourier amplitudes were calculated by performing discrete Fourier transform with the python package numpy [23] , setting the zero frequency component to 0 , then normalizing by the Euclidean norm of the Fourier components and calculating the amplitude at each frequency . These normalized amplitudes were averaged to produce the plots in Fig 4 ( D ) . We generated sets of 5 orthogonal basis vectors over the unit sphere embedded in 201 dimensions by sampling the 201 components of each vector from standard normal distributions and then performing QR decomposition on the 201 × 5 matrix of column vectors . The source code , simulation data for Application 1 , and a Python example workbook are available at https://github . com/jssong-lab/maxEnt .
|
Deep learning is a state-of-the-art reformulation of artificial neural networks that have a long history of development . It can perform superbly well in diverse automated classification and prediction problems , including handwriting recognition , image identification , and biological pattern recognition . Its modern success can be attributed to improved training algorithms , clever network architecture , rapid explosion of available data , and advanced computing power–all of which have allowed the great expansion in the number of unknown parameters to be estimated by the model . These parameters , however , are so intricately connected through highly nonlinear functions that interpreting which essential features of given data are actually used by a deep neural network for its excellent performance has been difficult . We address this problem by using ideas from statistical physics to sample new unseen data that are likely to behave similarly to original data points when passed through the trained network . This synthetic data cloud around each original data point retains informative features while averaging out nonessential ones , ultimately allowing us to extract important network-learned features from the original data set and thus improving the human interpretability of deep learning methods . We demonstrate how our method can be applied to biological sequence analysis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"sequencing",
"techniques",
"neural",
"networks",
"neuroscience",
"artificial",
"neural",
"networks",
"nucleosome",
"mapping",
"artificial",
"intelligence",
"network",
"analysis",
"computational",
"neuroscience",
"sequence",
"motif",
"analysis",
"molecular",
"biology",
"techniques",
"nucleotide",
"mapping",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"computer",
"and",
"information",
"sciences",
"network",
"motifs",
"gene",
"mapping",
"bioinformatics",
"molecular",
"biology",
"nucleotide",
"sequencing",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"computational",
"biology"
] |
2017
|
Maximum entropy methods for extracting the learned features of deep neural networks
|
Plasmacytoid dendritic cells ( pDC ) provide an important link between innate and acquired immunity , mediating their action mainly through IFN-α production . pDC suppress HIV-1 replication , but there is increasing evidence suggesting they may also contribute to the increased levels of cell apoptosis and pan-immune activation associated with disease progression . Although having the same clinical spectrum , HIV-2 infection is characterized by a strikingly lower viremia and a much slower rate of CD4 decline and AIDS progression than HIV-1 , irrespective of disease stage . We report here a similar marked reduction in circulating pDC levels in untreated HIV-1 and HIV-2 infections in association with CD4 depletion and T cell activation , in spite of the undetectable viremia found in the majority of HIV-2 patients . Moreover , the same overexpression of CD86 and PD-L1 on circulating pDC was found in both infections irrespective of disease stage or viremia status . Our observation that pDC depletion occurs in HIV-2 infected patients with undetectable viremia indicates that mechanisms other than direct viral infection determine the pDC depletion during persistent infections . However , viremia was associated with an impairment of IFN-α production on a per pDC basis upon TLR9 stimulation . These data support the possibility that diminished function in vitro may relate to prior activation by HIV virions in vivo , in agreement with our finding of higher expression levels of the IFN-α inducible gene , MxA , in HIV-1 than in HIV-2 individuals . Importantly , serum IFN-α levels were not elevated in HIV-2 infected individuals . In conclusion , our data in this unique natural model of “attenuated” HIV immunodeficiency contribute to the understanding of pDC biology in HIV/AIDS pathogenesis , showing that in the absence of detectable viremia a major depletion of circulating pDC in association with a relatively preserved IFN-α production does occur .
Plasmacytoid dendritic cells ( pDC ) are one of the two main subtypes of human dendritic cells . pDC , like the classical myeloid dendritic cells ( mDC ) , are able to present antigens to T cells [1] , but have a distinctive feature of producing type I interferons ( IFN ) [2] . pDC are able to secrete IFN-α at levels up to 1000 fold higher than any other blood cell following viral infection [2] . They recognize pathogens mainly via two pattern recognition receptors: Toll-like receptor 7 ( TLR7 ) , which recognizes single-strand RNA , and TLR9 , which recognizes unmethylated DNA . The triggering of these receptors induces pDC activation and IFN-α production [3] . IFN-α is a potent stimulator of other immune cells , like mDC and NK cells , playing a central role in the development of immune responses , in addition to its well-documented antiviral effects [2] . pDC are thought to be particularly important in immune responses against viral infections , including HIV . Accordingly , IFN-α is one of the most important cytokines able to suppress HIV replication [4] , [5] . However , increasing evidence suggests that IFN-α contributes to the generalized pan-immune activation and increased levels of cell apoptosis associated with AIDS progression , and thus the exact role of pDC in HIV/AIDS pathogenesis remains debatable [6]–[10] . HIV-2 infection is associated with low levels of circulating virus at all disease stages [11]–[15] . This is thought to be the main reason for the reduced HIV-2 transmission and its geographical confinement to West Africa and a few related European countries , in particular Portugal [16] , [17] . Despite being associated with a clinical spectrum similar to HIV-1 [18] , the rate of disease progression and CD4 decline is much slower irrespective of the disease stage [19] , [20] , leading to a limited impact on the survival of the majority of infected adults [21] . The reasons for the relatively benign course of HIV-2 infection remain poorly understood , and its potential to generate valuable insights into HIV immunopathogenesis has been little explored [16] , [17] , [22] , [23] . Importantly , we have previously shown that in HIV-2 infected patients , as in HIV-1 infection , CD4 depletion is directly linked to immune activation [22] , [24] . HIV-2 is closely related to HIV-1 , sharing ∼60% homology at the amino acid level in the group antigens ( GAG ) and polymerase ( POL ) and 30–40% in the regions encoding the envelope protein ( ENV ) [23] , and has been shown to be equally cytopathic in vitro [25] . Moreover , despite plasma viremia remaining low or undetectable throughout HIV-2 infection , the levels of proviral DNA do not significantly differ from those found in HIV-1 infected individuals [26]–[29] . These data suggest that HIV-2 , like HIV-1 , is able to disseminate and establishes a similar pool of infected cells . The reduced productive viral replication and the slow rate of the progressive immune activation and CD4 decline through the natural history of the disease are in agreement with distinct viral-host equilibrium during HIV-2 infection . Evidence exists to support preserved polyfunctional cellular specific responses [30]–[32] , and broad neutralizing antibodies are found in HIV-2 infected patients [33] , [34] . However , the debate continues as to whether these are the cause or the consequence of the control of viral replication and/or of a better preserved immune system [23] . Given the importance of the innate immunity in defining host-pathogen interactions , it is plausible that DC and other components of the innate response play a role . Accordingly , NK numbers and cytolytic activity have been shown to be better maintained in HIV-2 than in HIV-1 infection [35] . Importantly , a recent study showed that pDC are less susceptible to HIV-2 than to HIV-1 infection in vitro [36] . pDC express CD4 and both the HIV co-receptors CXCR4 and CCR5 , and may be infected by HIV-1 in vitro and in vivo [37] , [38] . Moreover , pDC apoptosis may be triggered by the binding of HIV-1 envelope in the absence of direct infection [39] . In this study we characterized for the first time circulating pDC in HIV-2 infected patients in order to generate insights into their role in HIV/AIDS pathogenesis . A similar marked reduction in the frequency of circulating pDC was found in untreated HIV-1 and HIV-2 infections that correlated with the degree of CD4 depletion and T cell activation , in spite of the absence of detectable viremia documented in the majority of HIV-2 patients . However , in contrast with HIV-1 , IFN-α production upon TLR9 stimulation was relatively preserved in HIV-2 infection , except in the few HIV-2 patients with detectable viremia in whom major impairments were found .
HIV-2 infection is characterized by reduced to undetectable viremia [11]–[15] and a much slower rate of CD4 decline as compared to HIV-1 [19] , [20] . We first asked whether this “favourable” outcome is associated with the preservation of circulating pDC . For this purpose , pDC were defined within freshly isolated peripheral blood mononuclear cells ( PBMC ) as HLA-DR+CD123+CD11c− after the exclusion of cell-lineage markers , as illustrated in Fig . 1A . Cohorts of untreated HIV-2 and HIV-1 individuals with similar degrees of CD4 T cell depletion but distinct viremia were evaluated ( Table 1 ) . HIV-1 and HIV-2 infected patients exhibited a similar marked reduction in blood pDC as compared to seronegative controls , assessed both as percentage of total PBMC and as absolute numbers ( Fig . 1B ) . This was not ascribed to sex , ethnicity or age distribution since no significant differences were found between males and females , Caucasians and non-Caucasians , and individuals with more or less than 45 years within each cohort or all cohorts combined ( data not shown ) . In order to evaluate whether the two infections also have similar levels of pDC depletion in early and advanced HIV disease , we stratified the HIV-1 and HIV-2 cohorts according to CD4 T cell counts ( >350 and <350 CD4 T cells/µl ) . As previously reported [40]–[43] , in HIV-1 infection pDC depletion was more marked in the advanced disease stage ( Fig . 1C ) . Of note , we found comparable pDC levels in advanced HIV-2 infected patients ( Fig . 1C ) . Moreover , a similar significant depletion was also documented in early disease in both infections as compared to seronegative subjects ( Fig . 1C ) . The association of pDC levels with disease progression in HIV-2 infection was further demonstrated by the statistically significant positive correlation found between the frequencies of pDC and circulating CD4 T cells ( Fig . 2A ) . Although pDC levels were found to negatively correlate with the frequency of CD4 T cells in some HIV-1 studies [42]–[46] , we found no significant correlation in our untreated HIV-1 cohort , possibly due to the reduced representation of patients with very low CD4 counts . Hence , HIV-1 and HIV-2 diseases are associated with a similar extent of pDC depletion in spite of the slower rate of CD4 decline and the better prognosis that characterize HIV-2 infection . Plasma viremia is thought to be a major determinant of pDC depletion in HIV-1 infection [7] , [40] , [42]–[44] , [47] . HIV-1 and HIV-2 infections are associated with markedly distinct plasma viral loads [11]–[15] . As shown in Table 1 , 20 out of 28 HIV-2 infected patients had undetectable viremia and those with detectable viremia showed levels significantly lower than the ones found in HIV-1 patients . Of note , the highest viremia documented in the HIV-2 cohort was 26 , 263 RNA copies/ml . In order to address the impact of viremia on the levels of circulating pDC , we divided the patients into two groups , viremic and “aviremic” ( levels below the test cut-off ) . As shown in Fig . 2B , the HIV-2 group with undetectable viremia exhibited significantly lower pDC levels than the seronegative cohort . In addition , HIV-2 infected patients with detectable viremia had significantly lower pDC levels than the “aviremic” HIV-2 patients ( Fig . 2B ) . However , it is important to stress that HIV-2 viremic individuals had significantly lower CD4 T cell counts than HIV-2 “aviremic” ( 356±60 cells/µl , n = 8 , and 790±97 cells/µl , n = 20 , respectively , p = 0 . 0112 ) . Nevertheless , a significant inverse correlation between the frequency of pDC within PBMC and viremia was observed in both HIV-2 ( r = −0 . 4485; p = 0 . 0089; n = 28 ) and HIV-1 ( r = −0 . 7684; p<0 . 0001; n = 22 ) cohorts . As shown in Fig . 2B , HIV-1 individuals with undetectable viremia do not exhibit pDC depletion . The HIV-1 patients able to control viral replication in the absence of antiretroviral drugs are considered to represent less than 1% of HIV-1 infected individuals [48] . Our small group of 4 HIV-1 “aviremic” individuals had follow-ups with undetectable viremia ranging from 2 to 10 years ( 6 . 45±3 . 28 years of follow-up as compared to 7 . 03±1 . 21 in “aviremic” HIV-2 ) and showed relatively well preserved CD4 T cell counts ( 814±242 cells/µl , range 344–1425; as compared to 790±97 cells/µl , range 52–1511 , in “aviremic” HIV-2 ) . Similar findings were obtained when circulating pDC numbers were analyzed instead of pDC frequency ( data not shown ) . In summary , in agreement with previous reports [40] , [42] , [43] , [47] , we found a significant negative correlation between viremia and pDC levels in HIV-1 infection . However , a major reduction of circulating pDC levels was found in HIV-2 infected patients with undetectable viremia , showing that HIV-2 infected patients exhibited a major reduction in circulating pDC irrespective of the presence of detectable plasma viral load . Both HIV-1 and HIV-2 infections are associated with a persistent generalized immune-activation , which is considered a main determinant of the immunodeficiency and that inversely correlates with CD4 T cell counts [22] , [24] , [49] . We assessed the relationship between pDC levels and expression of activation markers on T cells . HIV-2 infected cohort exhibited a significant inverse correlation between the frequency of pDC and the proportion of CD4 T cells expressing HLA-DR as well as of CD8 T cells that simultaneously expressed HLA-DR and CD38 ( Fig . 2C and 2D ) . In the case of HIV-1 infection , a significant inverse correlation was only found with CD8 T cell activation , as shown in Fig . 2C and 2D . This is relevant since CD8 T cell activation is considered a better marker of the hyper-activation state associated with HIV infection with prognostic value [50] . On the other hand , CD4 T cell activation may be in part related to the homeostatic response to CD4 depletion , and , as described above , in our HIV-1 cohort no inverse correlation was documented between pDC and CD4 circulating levels . Similar findings were obtained in relation to the absolute number of circulating pDC as well as in relation to other parameters of CD8 T cell activation , namely the percentage and mean fluorescence intensity ( FI ) of CD38 expression ( data not shown ) . Overall , pDC depletion directly correlates with T cell activation in both infections . We next asked whether the phenotype of circulating pDC differ in the two infections . The co-stimulatory molecule CD86 was similarly overexpressed on pDC in the HIV-1 and HIV-2 infected cohorts and this increase was statistically significant as compared to healthy controls ( Fig . 3A ) . No significant correlation was found between the CD86 expression , as assessed by percentage or geomean FI , and percentage of CD4 T cell ( r = −0 . 0699 for HIV-1; r = −0 . 08539 for HIV-2 , in the case of CD86 geomean FI ) or viremia ( r = −0 . 01922 for HIV-1; r = 0 . 1975 for HIV-2 , in the case of CD86 geomean FI ) in both infections . Moreover , we also found no correlation with the different parameters of CD4 and CD8 T cell activation evaluated . The ex vivo expression of the co-stimulatory molecules CD40 and CD80 within pDC was minimal in all individuals ( data not shown ) . Programmed death-1 ( PD-1 ) signaling mediates an inhibitory pathway of T cell response and its overexpression is considered to contribute significantly to the impairment of specific T cell responses in HIV-1 infected individuals [51] . We compared the expression of PD-1 ligands on pDC and found a statistically significant increase in the percentage of PD-L1+ pDC in both infections as compared to healthy controls ( Fig . 3A ) . In addition , the increase in the PD-L1 geomean FI within total pDC also reached statistical significance in HIV-1 infected individuals in comparison with healthy controls ( Fig . 3A ) . Again , no significant correlation was found between PD-L1 expression and percentage of CD4 T cell ( r = 0 . 1200 for HIV-1; r = −0 . 02762 for HIV-2 , in the case of PD-L1 geomean FI ) , or viremia ( r = −0 . 02476 for HIV-1; r = −0 . 2501 for HIV-2 , in the case of PD-L1 geomean FI ) , or the T cell activation markers assessed in both infections . pDC expression of PD-L2 was minimal in all the three cohorts ( data not shown ) . In summary , both CD86 and PD-L1 were similarly up-regulated on pDC of both HIV-2 and HIV-1 cohorts , irrespective of disease stage . pDC are known to express TLR9 , which binds to unmethylated CpG motifs , and to mature upon TLR9 signaling [3] , [52] . Studies on the modulation of pDC phenotype in vitro have been scarce and mainly conducted in HIV-1 patients under antiretroviral therapy [53] . We assessed the modulation of pDC phenotype upon TLR9 stimulation by stimulating freshly isolated PBMC with a TLR9 ligand ( CpG type A ) or a non-CpG oligodeoxynucleotide ( ODN ) as a negative control . After 22 h , cells were harvested and analyzed within a pDC gate as described above . CD86 and PD-L1 were found to be up-regulated by the control non-CpG ODN ( data not shown ) , precluding their use to evaluate pDC maturation induced by CpG . Therefore , we focused our analysis on the CD40 and CD80 molecules that , although exhibiting reduced ex vivo expression , were specifically up-regulated upon CpG stimulation ( Fig . 3B ) . Results are shown as stimulation index ( SI ) calculated as the ratio of the geomean FI measured in the presence of CpG and medium alone . The capacity of pDC to up-regulate CD40 after CpG stimulation was significantly decreased both in HIV-1 and in HIV-2 individuals relative to healthy controls ( Fig . 3C ) . An impairment of CD80 up-regulation was also documented in both infections , though without reaching statistical significance in comparison with controls ( CD80 SI: 2 . 4±0 . 3 for seronegatives , 1 . 9±0 . 1 for HIV-1 , 1 . 9±0 . 1 for HIV-2 ) . The stimulation index for CD40 geomean FI was found to have a significant positive correlation with the percentage of CD4 T cells ( r = 0 . 8451; p<0 . 0001 ) and a negative correlation with viremia ( r = −0 . 7312; p = 0 . 002 ) in the HIV-1 cohort , but no significant correlations were found in the HIV-2 cohort ( r = −0 . 065 with percentage of CD4 T cells; and r = 0 . 0846 with viremia ) . Moreover , in contrast to the HIV-1 cohort , similar levels were found when HIV-2 patients with more and less than 350 CD4 T cells/µl ( data not shown ) or with detectable and undetectable viremia were compared ( Fig . 3C ) . These data suggest that the impairment in CD40 up-regulation upon CpG stimulation was present throughout HIV-2 disease and was not further aggravated in late stages . Overall , the circulating pDC of HIV-infected individuals showed a reduced ability to further differentiate upon CpG-A stimulation as compared to seronegative controls . IFN-α production is mainly triggered through TLR7 and TLR9 [3] . CpG-A has been shown to preferentially act on pDC [52] and was used here to assess pDC ability to secrete IFN-α upon TLR9 stimulation . Using single-cell assessment by flow cytometry , we further confirmed that in our experimental conditions the CpG-A used selectively induced IFN-α production by pDC ( Fig . S1 ) . Both HIV-1 and HIV-2 infected cohorts exhibited a significantly lower IFN-α production upon CpG stimulation as compared to healthy controls ( Fig . 4A ) . Of note , similar levels of IFN-α production were found in the infected cohorts in Caucasians and non-Caucasians , individuals with more or less than 45 years , as well as males and females . Thus , despite recent data showing increased IFN-α production in women upon TLR7 stimulation in vitro [54] , we were unable to detect any difference upon CpG stimulation . Patients with less than 350 CD4 T cells/µl tended to produce lower IFN-α levels than the patients with higher CD4 counts ( Fig . 4B ) . In agreement , IFN-α production was found to positively correlate with the frequency of circulating CD4 T cells and inversely with the up-regulation of activation markers in CD4 and CD8 T cells in both infected cohorts ( Fig . 5A ) . Noteworthy , the ability to produce IFN-α was also significantly lower in viremic than “aviremic” cohorts ( Fig . 4C ) and a significant correlation was found with viremia in both infections ( Fig . 5A ) . We estimated the IFN-α production on a per cell basis , by dividing the concentration of IFN-α produced upon CpG stimulation by the absolute number of pDC in the culture . Although there was a decrease in the ability of pDC to produce IFN-α in both infections , significantly higher estimated IFN-α levels per pDC were found in HIV-2 than in HIV-1 infected patients ( Fig . 4D ) . Patients with less than 350 CD4 T cells/µl had lower levels of IFN-α production per pDC ( Fig . 4E ) , but no significant correlation was found between IFN-α production per pDC and the degree of CD4 depletion in either infection ( Fig . 5B ) . In addition , no correlation was found between the estimated levels of IFN-α production per pDC and CD4 or CD8 T cell activation in the HIV-1 cohort , in contrast with the significant correlations found in HIV-2 infection ( Fig . 5B ) . Importantly , when the cohorts were divided according to the viremia status , the HIV-2 patients with undetectable viremia exhibited preserved levels of IFN-α production per pDC ( Fig . 4F ) . Moreover , detectable HIV-2 viremia was associated with a significant decrease in the levels of IFN-α per pDC , reaching levels similar to the ones documented in HIV-1 viremic patients ( Fig . 4F ) . In agreement , a significant correlation was documented between IFN-α production per pDC and viremia in the HIV-2 cohort that was not observed in the HIV-1 cohort ( Fig . 5B ) . These data suggest a major role of plasma viral load , even at low levels such as those found in HIV-2 patients , in the impairment of IFN-α production . Therefore , we next assessed the relationship between IFN-α production and proviral DNA . The levels of proviral DNA were similar in the HIV-1 and HIV-2 cohorts , including in “aviremic” groups ( Fig . S2 ) . No significant correlations were found between proviral DNA and pDC levels in both infections ( data not shown ) . Of note , a significant correlation between viremia and proviral DNA was only found in the HIV-2 cohort ( r = 0 . 4480 , p = 0 . 0168 for HIV-2; and r = 0 . 0145 , p = 0 . 9518 for HIV-1 ) . Remarkably , a statistically significant positive correlation between proviral DNA and Net IFN-α production was found in the HIV-2 cohort that was not documented in HIV-1+ patients , showing that the higher the number of infected cells the higher the IFN-α production in HIV-2 infection ( Fig . 5A ) . In contrast , the estimated levels of IFN-α per pDC correlated negatively with proviral DNA in the HIV-1 but not in the HIV-2 cohort ( Fig . 5B ) . These data suggest that the number of infected cells contributed more to the impairment of IFN-α production on a per pDC basis upon TLR9 stimulation in HIV-1 than in HIV-2 infection , possibly related to the higher levels of effective viral replication in HIV-1 infection . CpG has been suggested to modulate the production of other cytokines both directly and indirectly through effects mediated by CpG-induced IFN-α [2] . We investigated the effect of CpG stimulation on the production of IL-10 , IL-12p40 , TNF-α and the β-chemokine MIP-1β by measuring their levels in the culture supernatants using a Luminex-based multiplex assay . Of note , the analysis of the combined cohorts revealed a direct correlation between the levels of IFN-α production upon CpG stimulation and the levels of TNF-α ( r = 0 . 4033 , p = 0 . 0027 for Net IFN-α and r = 0 . 4207 , p = 0 . 0017 for Net IFN-α/1000pDC; n = 53 ) and MIP-1β ( r = 0 . 3978 , p = 0 . 0035 for Net IFN-α and r = 0 . 4272 , p = 0 . 0016 for Net IFN-α/1000pDC; n = 52 ) . No such correlations were found in the case of IL-10 or IL-12p40 . As illustrated in Fig . 6A , the most striking finding was a reduced ability of HIV-1 infected individuals to produce MIP-1β and TNF-α , as compared to healthy and HIV-2 infected cohorts upon CpG stimulation . These decreases were clearer when the stimulation index was analyzed as illustrated in Fig . 6B , showing that a statistically significant reduction was only found in the case of TNF-α production in the HIV-1 cohort as compared to healthy controls . Of note , when the HIV-2 cohort was split accordingly to viremia status , the individuals with undetectable viremia exhibited a preserved ability to produce TNF-α ( Fig . 6C ) and MIP-1β ( data not shown ) in response to CpG stimulation , whereas the patients with detectable viremia showed a decrease in stimulation indexes similar to HIV-1 infected patients . However , despite the clear trends ( p = 0 . 07 in the case of the viremic cohorts as compared to healthy controls ) , none of these differences reached statistical significance ( Fig . 6C ) . These data suggest that viremia impacts on the ability to produce pro-inflammatory cytokines upon TLR9 stimulation ex vivo in HIV-2 infected individuals , as documented for IFN-α . Despite the consensual data regarding the decreased ability of pDC to produce IFN-α upon in vitro stimulation in HIV-1 infected patients [40] , [41] , [47] , [55] , there are conflicting results regarding circulating IFN-α levels [43] , [56]–[58] . To our knowledge , there are no reports of serum IFN-α levels in chronic HIV-2 infection . We assessed serum IFN-α by ELISA and found that , similarly to seronegative donors , all HIV-2 and the majority of HIV-1 infected patients had undetectable levels ( <12 , 5 pg/ml ) . The only exceptions were two advanced HIV-1 infected patients with 13 , 18 pg/ml and 22 , 16 pg/ml of IFN-α ( patients with 6 . 5 and 5 . 7 log10 RNA copies/ml , and with 18 and 290 CD4 T cells/µl , respectively ) . In spite of the reduced circulating IFN-α levels and impaired production upon TLR9 and/or TLR7 stimulation in vitro , there is increasing evidence of increased IFN-α production in vivo during HIV-1 infection [7] , [39] , [59] . In order to evaluate this possibility , we quantified in freshly isolated PBMC the relative mRNA expression levels of MxA , a gene that is mainly induced by IFN-α and , thus , has been used as an indicator of IFN-α activity [46] , [60]–[63] . As depicted in Table 2 , significantly higher levels were found in HIV-1 than in HIV-2 infected patients . Importantly , in the HIV-1 cohort , the MxA levels were directly correlated with both viremia and levels of CD8 T cell activation , and inversely correlated with the frequency of circulating pDC and with the Net IFN-α production upon CpG stimulation in vitro ( Table 2 ) . No significant correlations were found in the HIV-2 cohort ( Table 2 ) . The “aviremic” patients tended to have lower levels of MxA expression than the viremic individuals in both infections , though not reaching statistical significance ( relative MxA expression: 50±39 and 380±155 , for the HIV-1 cohort; 81±23 and 122±50 , for the HIV-2 cohort; respectively ) . These data are in agreement with lower activity of IFN-α in vivo during HIV-2 as compared to HIV-1 infection , that possibly explain the reduced refractoriness to IFN-α production upon further in vitro pDC stimulation observed in HIV-2 infected patients . In summary , despite the similar decrease of pDC in both infections , pDC from HIV-2-infected patients appear to better preserve their ability to produce IFN-α upon CpG stimulation . Of note , no increase in the ex vivo levels of serum IFN-α was documented in HIV-2 infection in parallel with absence of evidence of significant increase in the IFN-α activity in vivo , as assessed by MxA relative expression levels . However , detectable viremia was associated with a similar impairment of IFN-α production in both infections , suggesting a major role of circulating virus in the pDC functional impairment ex vivo .
This study characterized for the first time circulating pDC in individuals with HIV-2 infection . A similar decrease in pDC levels was found in untreated HIV-2 and HIV-1 infections in spite of the much lower viremia and slower rate of disease progression that distinguishes HIV-2 disease . Importantly , a significant depletion of circulating pDC was documented even in HIV-2 infected individuals with undetectable viremia . The pDC levels were directly correlated with the degree of CD4 T cell depletion and T cell activation in both infections . Conversely , viremia appears to have a major impact on the ability of the remaining pDC to produce IFN-α upon TLR9 stimulation in vitro in HIV-2 infected patients . HIV-1 infection has been consistently shown to be associated with reduced frequency and impaired function of circulating pDC , both during primary and chronic infection [40]–[44] , [47] , [55] , [56] . These defects have been found to be more pronounced in individuals with higher viremia [40] , [42]–[44] , [47] and/or lower CD4+ T cell counts [42]–[44] and to be associated with the development of opportunistic infections and tumors [40] . Viral infection and induction of pDC apoptosis are thought to significantly contribute to the pDC depletion during both HIV-1 and SIV disease [39] . However , HIV-2 was shown to be less able than HIV-1 to infect pDC in vitro [36] and reduced levels of viral replication are documented in HIV-2 infected patients [11]–[15] . Therefore , the finding of a similar reduction of pDC levels was unexpected and suggests that it may be related to other mechanisms than direct viral effects . This possibility is further supported by the delayed and frequently incomplete recovery of pDC numbers and function following long-term successful antiretroviral treatment in HIV-1 seropositive patients [44] , [47] , [56] . Interestingly , the 4 HIV-1 infected patients with undetectable viremia in the absence of antiretroviral therapy evaluated in this study showed better preserved pDC levels than aviremic HIV-2 individuals . This was not apparently related to distinct length of patients' follow-up or proviral load , and could not be ascribed to the higher HIV-2 viremia cut-off , since viremia was quantified in the last 14 HIV-2 patients evaluated using an up-dated assay with a limit of detection of 40 RNA copies/ml and found to be undetectable ( data not shown ) . The distinct pDC levels in the “aviremic” HIV-1 and HIV-2 groups suggest that the study of larger cohorts of these particular HIV-1 infected individuals , usually called “Elite controllers” , will be instrumental to better understand pDC biology in HIV/AIDS . Traffic alterations have also been suggested to contribute to the reduced levels of circulating pDC . Several in vitro studies have documented a viral-associated up-regulation of molecules , such as CCR7 on pDC , that may contribute for their preferential homing to lymphoid tissues [64] . However , the changes in cell redistribution have not been consistently confirmed either in HIV-1 infected patients [7] , [65]–[67] or in non-human primates infected with SIV [68]–[71] . Although there are studies that demonstrated an increase in pDC counts in lymph nodes and spleen during HIV-1 and SIV infections [65]–[67] , [69] , [70] , other studies reported a parallel pDC decrease in the peripheral blood and lymphoid tissues [68] and an increase in pDC primed to apoptosis in the lymph nodes [71] . There are no data available on lymphoid tissues during HIV-2 disease . Of note , the establishment of HIV-2 infection is associated with levels of proviral DNA similar to those found in HIV-1 , suggesting an equivalent viral dissemination despite the reduced HIV-2 viremia [26]–[29] , [72] . Therefore , it is plausible that HIV-2 RNA and/or HIV-2 proteins may induce pDC maturation and migration . In agreement , we found the same alterations in the phenotype of circulating pDC in the two infections , and these changes were not associated with disease stage or viremia status . Of note , PD-L1 has been shown to be up-regulated in pDC upon TLR stimulation by HIV-1 products [73] . Strong correlations between pDC decline and up-regulation of markers of CD8 T cell activation both in HIV-2 and HIV-1 infections represent another important finding of our study . We have previously shown that generalized immune activation is likely to be a main determinant of HIV-2 disease progression [22] , [24] , as has been demonstrated in HIV-1 infection [22] , [24] , [50] , [58] . Persistent HIV-2 infection is thought to induce a chronic stimulation of the immune system leading to a progressive T cell impairment and CD4 depletion , though at much slower rates than in HIV-1 infection [22] , [24] , [49] . The generalized pro-inflammatory state is likely to contribute to pDC depletion both by altered cell traffic and apoptosis susceptibility , as well as through the impairment of the ability of DC precursors to differentiate . We showed a similar up-regulation of the levels of expression of co-stimulatory and co-inhibitory molecules in the two infections , which may represent a state of incomplete differentiation that may preclude adequate antigenic presentation or be associated with tolerogenic properties , as previously reported [66] , [74] . On the other hand , besides their antiviral properties , pDC are increasingly viewed as conductors of the immune activation associated with HIV/AIDS pathogenesis [6] . In this respect , a dual role has been suggested . A deleterious contribution of the pDC-mediated activation of other cells of the immune system [9] was further supported by the recently documented impairment of pDC activation and IFN-α production in sooty mangabeys , natural hosts of SIV infection that are known to exhibit reduced levels of immune activation and do not progress to AIDS [8] . Additionally , pDC were shown to be able to induce regulatory T cells upon activation by HIV-1 through an indoleamine 2 , 3-dioxygenase ( IDO ) -mediated mechanism [75] and , in this way , modulate immune activation and limit HIV specific responses [74] , [75] . Despite some discrepant results , there is usually no significant increase in circulating IFN-α levels until advanced HIV-1 disease stages [43] , [56]–[58] . In agreement , we were unable to detect increased serum IFN-α in the large majority of HIV-1 infected patients . We also found no detectable serum levels of IFN-α during HIV-2 infection . Importantly , although a decrease in the pDC ability to produce IFN-α in vitro has been consistently observed during HIV-1 infection [40] , [41] , [47] , [55] , several lines of evidence suggest that there is increased production of IFN-α in vivo . HIV-1 infected patients have been shown to exhibit increased transcriptional levels of IFN-α and of several genes that are known to be induced by IFN-α [7] , [39] , [59] . There are no data on HIV-2 infected patients . Here , we showed that HIV-2 as compared to HIV-1 infection is associated with lower levels of MxA expression , a gene induced by IFN-α [60] , [61] . Our data support a lower IFN-α activity in vivo throughout the course of HIV-2 infection , possibly contributing to the slower progression of immune activation and consequently lower rate of CD4 decline that distinguishes the HIV-2 disease [19] , [20] , [22] , [24] . Worth noting , we found that in HIV-1 infection the levels of MxA expression directly correlated with viremia and were inversely related to the circulating pDC levels and the amount of IFN-α production upon in vitro TLR9 stimulation . These data further support the possibility that continuous pDC stimulation by TLR ligands in vivo leads to a refractory state of the pDC that explains the apparent paradox between reduced in vitro production of IFN-α and indirect evidence of increased IFN-α production in vivo [7] . HIV-1 itself has been shown to modulate pDC function , particularly through the binding of viral RNA to TLR7 [76] , as well as through the impairment of TLR9 signaling by the envelope protein gp120 [77] . The corresponding HIV-2 envelope protein , gp105 , was shown to exhibit distinct impacts in several immunological systems [78]–[81] , but there are no data on its effects on TLR9 signaling . We tested here the IFN-α production upon TLR9 stimulation in vitro and found it to be significantly impaired in HIV-2 infected individuals . Given the limited volume of patient samples , we selected CpG type A as a standard TLR9 ligand thought to target mainly pDC [52] , as confirmed here by single-cell analysis of IFN-α production . A similar decrease in IFN-α production upon CpG stimulation of whole blood cultures was also recently reported in HIV-2 and HIV-1 infected cohorts in Guinea-Bissau , West Africa , as compared to non-infected individuals [82] . As discussed above , HIV-2 infected individuals are thought to have reduced levels of viral replication and , therefore , it is expected that their pDC would be exposed to much less HIV-related molecules able to signal through TLR . Importantly , when we split the HIV-2 cohort according to viremia status , we found that the individuals with undetectable viremia exhibited a preserved IFN-α production on a per pDC basis . These data suggested that despite their reduced number , pDC function was preserved in HIV-2 infected patients without detectable circulating virus . In contrast , a similar impairment in IFN-α production was found in viremic HIV-2 and HIV-1 infected patients despite the average 2 log difference in the number of plasma viral RNA copy numbers . These results suggest that even low levels of circulating virus are sufficient to intrinsically impair IFN-α production by pDC or to induce a pDC refractory state that prevents their subsequent response to further TLR9 stimulation . In conclusion , we reported here for the first time a major depletion of circulating pDC during HIV-2 infection , a unique natural model of “attenuated” HIV immunodeficiency . This decrease was observed early in disease and also in HIV-2 infected patients with undetectable viremia , suggesting that mechanisms other than pDC direct viral infection play major roles in their depletion during persistent infections . On the other hand , viremia was associated with an impairment of IFN-α production on a per pDC basis upon TLR9 stimulation , in agreement with the possibility that diminished function in vitro is likely a consequence of prior activation by HIV virions in vivo .
The study was approved by the Ethical Board of the Faculty of Medicine , University of Lisbon . Subjects gave written informed consent to blood sampling and processing . In exceptional cases , related to cultural factors , oral informed consent was chosen by the patient and the assistant physician provided a written declaration of the permission obtained . A cross-sectional study was performed involving 28 HIV-2 and 22 HIV-1 infected patients without ongoing opportunistic infections or tumours , followed at Hospital de Santa Maria in Lisbon , Portugal . 18 HIV-seronegative age-matched control subjects were studied . Cohort characterization is summarized in Table 1 . PBMC were isolated from heparinized blood immediately after venopuncture by Ficoll-Hypaque density gradient centrifugation . PBMC were cultured at 2×106 cells/ml in 24-well plates in 1 . 5 ml of RPMI1640 supplemented with 100 U/ml penicillin/100 µg/ml streptomycin , 2 mM glutamine ( all from Gibco-Invitrogen , Paisley , U . K . ) and 10% human AB serum ( Sigma-Aldrich , St Louis , MO ) at 37°C with 5% CO2 , in the absence or presence of 10 µg/ml class A CpG-ODN 2336 ( 5′-gggGACGACGTCGTGgggggg-3′ ) or the non-CpG-ODN 2243 control ( 5′-gggGGAGCATGCTGgggggg-3′ ) provided by Coley Pharmaceutical Group ( Wellesley , MA ) . After 22 h , cells were harvested for phenotypic analysis and culture supernatants stored at −80°C for subsequent cytokine evaluation . PBMC surface staining was performed as previously described [83] , and analyzed for pDC frequency ex vivo with the following anti-human conjugated antibodies: FITC conjugated lineage ( Lin ) markers ( CD3 and CD14 from Sanquin , Amsterdam , Netherlands; CD16 from BD Biosciences , San Jose , CA; and CD20 from eBioscience , San Diego , CA ) ; HLA-DR PerCP ( L243 , BD Biosciences ) ; CD123 PE-Cy7 ( 6H6 , eBioscience ) , and CD11c APC ( B-ly6 , BD Biosciences ) . Analysis was done within a large gate including lymphocytes and monocytes , defined according to their forward/side scatter characteristics . pDC were defined as Lin−HLA-DR+CD123+CD11c− . Gated pDC were further analyzed using PE-conjugated mAbs against CD40 ( 5C3 ) and PD-L2 ( MIH18 ) from eBioscience , CD80 ( L307 . 4 , BD Biosciences ) and APC-conjugated mAbs against CD86 ( FUN-1 , BD Biosciences ) and PD-L1 ( MIH1 , eBioscience ) . The same strategy was applied to evaluate the phenotype of pDC after in vitro culture of PBMC with CpG-ODN . At least 400 , 000 events were acquired within a lymphocyte+monocyte gate , using a CANTO flow cytometer ( BD Biosciences ) , and analyzed using FlowJo ( Tree Star , Inc , Ashland , OR ) . The absolute numbers of pDC/µl of blood were calculated by multiplying their representation by the sum of the absolute lymphocyte and monocyte counts obtained at the clinical laboratory . The expression of pDC surface markers was evaluated both in terms of percentage and of geomean FI . IFN-α was quantified in serum samples and in culture supernatants using the VeriKine™ Human IFN-Alpha Serum Sample ELISA and Human IFN-Alpha ELISA Kit , respectively ( PBL InterferonSource , Piscataway , NJ ) , according to manufacturer's instructions . IFN-α production at the single cell level was assessed by flow cytometry in PBMC cultured for 18 h in the presence of Brefeldin A ( last 16 h culture; 10 µg/ml; Sigma ) by intracellular staining with the anti-human IFN-α ( clone 225 . C; eBioscience ) after surface staining , using a previously described protocol [84] . IL-10 , IL-12p40 , MIP-1-β and TNF-α were quantified in the supernatants of PBMC cultured as described above using the Human Cytokine LINCOplex Kit ( Millipore Corporation , Billerica , MA ) and the Luminex LX100 ( Luminex Corporation , Austin , TX ) according to manifacturer's instructions . Samples were assayed in duplicate . Freshly isolated PBMC ( 1–5×106 cells ) were immediately placed into RLT lysis buffer ( Qiagen , Valência , CA ) and stored at −80°C . Lysates were further homogenized by passage through QIAshredder columns ( Qiagen ) . Polyadenylated mRNA was extracted using Oligotex Direct mRNA Mini kit ( Qiagen ) . mRNA was reverse transcribed into cDNA using random hexamers and Superscript II Reverse Transcriptase Kit ( all from Invitrogen ) . mRNA and cDNA concentrations were determined using a NanoDrop ND-10 spectrophotometer ( NanoDrop technologies , Wilmington , DE ) . MxA expression was determined by Quantitative Real-Time PCR using AbiPrism 7000 SDS thermocycler ( Applied Biosystems ) using an optimized kit prepared by PrimerDesign Southampton , UK , with the following protocol: enzyme activation ( 95°C for 10 minutes ) , followed by 45 cycles of denaturation ( 95°C for 15 seconds ) and annealing and data collection ( 60°C for 60 seconds ) . Each sample was quantified in duplicate using 1 µg of cDNA in a 20 µl PCR mixture volume containing 10 µl of Platinum Quantitative PCR SuperMix-UDG , 0 , 4 µl ROX Reference Dye 50X ( all from Invitrogen ) , 300 nM of MxA primers/probe mix or 300 nM of GAPDH primers/probe mix ( internal control ) , both from PrimerDesign . Absolute quantities of mRNA product were determined from a standard curve of serial dilutions of known quantities of each specific amplicon ( Primer Design ) . Results are presented as number of copies of MxA mRNA per 1000 copies of GAPDH . Proviral DNA was quantified by real-time PCR based assays that amplify highly conserved regions in HIV-1 and HIV-2 gag using protocols that we have previously described [29] . The detection limit of the assays was 5 DNA copies/106 PBMC . HIV-1 viremia was quantified by RT-PCR ( detection threshold of 40 RNA copies/ml , Roche , Basel , Switzerland ) . HIV-2 viremia was quantified using a RT-PCR-based assay [15] with a detection limit of 200 RNA copies/ml . The cut-off values of the tests were considered for the purpose of the analysis in the cases where detection was below this level . Statistical analysis was performed using GraphPad Prism version 5 . 00 ( GraphPad Software , San Diego , CA ) . The data are presented as arithmetic mean ± SEM and were compared using Mann-Whitney test and Wilcoxon matched pairs test as appropriate . Spearman's correlation coefficient was used to assess the correlation between two variables . P-values <0 . 05 were considered to be significant .
|
Infection by HIV-2 , the second AIDS-associated virus , is considered a unique natural model of attenuated HIV disease . HIV-2 infected individuals exhibit much lower levels of circulating virus ( viremia ) and progress to AIDS at slower rates than HIV-1 infected patients . In this study , we characterized for the first time blood plasmacytoid dendritic cells ( pDC ) , important mediators between innate and acquired immunity , in HIV-2 infection . We observed a profound reduction in circulating pDC levels in HIV-2 infected patients , even in those with undetectable viremia , to levels similar to those found in HIV-1 infection . Moreover , we documented a more differentiated pDC phenotype in both infected cohorts relative to healthy individuals . Despite these similarities between HIV-1 and HIV-2 infections , pDC from HIV-2 patients with undetectable viremia exhibited , upon in vitro stimulation , a better-preserved ability to produce interferon-α ( IFN-α ) , an important anti-viral cytokine with potential to stimulate other immune cells . Overall , our data suggest that the presence of virus in circulation , although not critical for the reduction in pDC number , appears to be central for the impairment of their function . This study of pDC in HIV-2 infection fills a gap in the understanding of their potential role in HIV/AIDS pathogenesis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"virology/immunodeficiency",
"viruses",
"immunology/innate",
"immunity",
"infectious",
"diseases/hiv",
"infection",
"and",
"aids",
"infectious",
"diseases/viral",
"infections",
"immunology",
"immunology/immunity",
"to",
"infections",
"virology/host",
"antiviral",
"responses"
] |
2009
|
Major Depletion of Plasmacytoid Dendritic Cells in HIV-2 Infection, an Attenuated Form of HIV Disease
|
The HIV-1 accessory protein Vpu counteracts tetherin ( BST-2/CD317 ) by preventing its incorporation into virions , reducing its surface expression , and ultimately promoting its degradation . Here we characterize a putative trafficking motif , EXXXLV , in the second alpha helix of the subtype-B Vpu cytoplasmic tail as being required for efficient tetherin antagonism . Mutation of this motif prevents ESCRT-dependent degradation of tetherin/Vpu complexes , tetherin cell surface downregulation , but not its physical interaction with Vpu . Importantly , this motif is required for efficient cell-free virion release from CD4+ T cells , particularly after their exposure to type-1 interferon , indicating that the ability to reduce surface tetherin levels and promote its degradation is important to counteract restriction under conditions that the virus likely encounters in vivo . Vpu EXXXLV mutants accumulate with tetherin at the cell surface and in endosomal compartments , but retain the ability to bind both β-TrCP2 and HRS , indicating that this motif is required for a post-binding trafficking event that commits tetherin for ESCRT-dependent degradation and prevents its transit to the plasma membrane and viral budding zones . We further found that while Vpu function is dependent on clathrin , and the entire second alpha helix of the Vpu tail can be functionally complemented by a clathrin adaptor binding peptide derived from HIV-1 Nef , none of the canonical clathrin adaptors nor retromer are required for this process . Finally we show that residual activity of Vpu EXXXLV mutants requires an intact endocytic motif in tetherin , suggesting that physical association of Vpu with tetherin during its recycling may be sufficient to compromise tetherin activity to some degree .
The downregulation of cell-surface immunomodulatory proteins is a common theme in the evasion of innate and adaptive immune responses by mammalian viruses . One host molecule targeted by diverse enveloped viruses is tetherin ( CD317/BST-2 ) , an interferon-induced dimeric type-II membrane protein , which inhibits the release of nascent virions from the cell surface [1] , [2] . By virtue of an unusual topology that consists of an N-terminal transmembrane domain and a C-terminal glycophosphatidylinositol ( GPI ) -linkage separated by an extended parallel coiled-coil domain [3]–[6] , tetherin partitions into budding virus particles and is thought to directly cross-link the nascent virion to the plasma membrane ( PM ) [7] . Tethered virions may then be endocytosed and degraded in endosomes . Tetherin itself constitutively recycles between the PM , endosomal and trans-Golgi network ( TGN ) compartments through a non-canonical trafficking motif ( YXYXXV ) in its N-terminal cytoplasmic tail , which engages clathrin adaptors AP-1 and AP-2 [8] , [9] . Since tetherin targets a structural component of the virion not encoded by the viral genome , namely the host-cell derived membrane , its potential importance in the innate antiviral response is highlighted by multiple examples of virally-encoded countermeasures ( reviewed in [10] ) . The ability to counteract tetherin is conserved among human and simian immunodeficiency viruses ( HIVs/SIVs ) , although the viral proteins tasked with this activity vary [10] . In HIV-1 the accessory protein Vpu fulfills this role . Vpu , a small integral membrane phosphoprotein , directly associates with tetherin through interactions between the transmembrane domains of both proteins [11]–[14] . Vpu blocks tetherin incorporation into assembling virions [7] and leads to a reduction of tetherin levels on the plasma membrane ( PM ) [2] . Subsequently , tetherin is degraded , most likely in lysosomal compartments , by an ubiquitin-dependent mechanism [15] , [16] . Phosphorylation of Vpu on two conserved serine residues recruits a SCF β-TrCP1/2 E3 ligase complex [17] that ubiquitinates the tetherin cytoplasmic tail on multiple residues [18] , and recent evidence demonstrates that this tetherin degradation is dependent on the ESCRT pathway [19] . However , tetherin degradation is not strictly required for Vpu activity [20] , [21] . While recruitment of the ESCRT-0 subunit HRS by Vpu counteracts tetherin activity [19] , a novel core component of ESCRT-1 , Ubiquitin Associated Protein 1 ( UBAP1 ) , essential for tetherin degradation induced by both Vpu and the KSHV ubiquitin ligase K5 , is dispensable [22] . Coupled with recent evidence that dysregulation of the entire late endosomal compartment by mutants of Rab7a [23] , this suggests an emerging picture that Vpu alters tetherin trafficking to counteract its antiviral activity prior to lysosomal delivery . While HIV-2 and SIV tetherin antagonists Env and Nef promote tetherin internalization through their interactions with AP-2 , [24]–[26] , Vpu does not enhance the rate of tetherin endocytosis [13] , [16] . Rather it is thought that Vpu/tetherin interactions preclude both the recycling of tetherin back to the cell surface and the transit of newly synthesized tetherin to the PM by trapping it in intracellular compartments , notably the TGN [13] . Consistent with this , the ability of Vpu to localize to the TGN correlates with tetherin antagonism [27] , and disruption of the recycling compartment by a dominant Rab11a mutant compromises Vpu activity [28] . Truncations of the Vpu cytoplasmic tail , particularly the second alpha helix , lead to aberrant localization and a reduction in its anti-tetherin activity , suggesting it harbors a domain required for Vpu function [29] . In this study we have examined the role of the second alpha helix of HIV-1 Vpu in tetherin antagonism . We identify a putative sorting signal that is required for post-binding trafficking of Vpu/tetherin complexes and inhibition of antiviral activity in primary CD4+ T cells . While this signal can be functionally replaced by a clathrin adaptor binding peptide derived form HIV-1 Nef , Vpu activity does not require the canonical adaptors AP-1 , AP-2 or AP-3 . Moreover , because residual activity of second helix mutants requires an intact recycling signal in tetherin , we propose that second alpha helix mutants are selectively defective for routing tetherin into an endosomal degradation pathway thereby inhibiting its transit to the PM and incorporation into nascent virions .
Truncations of the Vpu cytoplasmic tail lead to aberrant localization and a reduction in its anti-tetherin activity [29] . To further study the determinants within the second alpha helix of Vpu that account for TGN localization , we performed overlapping triple-alanine scan mutagenesis through the second alpha helix of a codon optimized HIV-1 NL4 . 3 Vpu construct bearing a C-terminal HA tag ( Figure 1A ) . We then assayed these Vpu mutants for their ability to rescue Vpu-defective HIV-1 from tetherin restriction . 293T cells were transfected with wildtype HIV-1 ( HIV-1 wt ) or Vpu defective HIV-1 ( HIV-1 delVpu ) proviruses in combination with fixed doses of a human tetherin expression vector , and 25 ng of Vpu-HA or mutant thereof . 48 h after transfection cell lysates and supernatants were harvested and analyzed for physical viral yield by Western blot ( Figure 1B ) or supernatant infectivity of HeLa-TZMbl indicator cells ( Figure 1C ) . As expected , in the absence of Vpu , both supernatant particle yield and infectivity of HIV-1 delVpu was profoundly reduced in the presence of tetherin , expression of Vpu in trans rescued virus production to wildtype levels . By contrast mutations encompassing either E59 or L63 and V64 but not the intervening or subsequent amino acids displayed defective rescue of HIV-1 delVpu ( Figure 1B and C ) . All Vpu mutants with the exception of Vpu 67-69A-HA were expressed equivalently . Vpu 63-65A-HA appeared to display a dominant interfering activity on HIV-1wt titer , but this was not reflected as apparently in particle yield . Thus these data suggested a functional requirement for E59 and L63/V64 in tetherin antagonism by Vpu . To confirm this we mutated these residues to alanine in the context of an HIV-1 NL4 . 3 provirus ( NL4 . 3 Vpu ELV ) and examined viral release from 293T cells in the presence of increasing expression of tetherin . Because this part of Vpu overlaps with start of the Env open reading frame in the provirus , these mutations were rendered silent in the +1 reading frame and displayed no defect in virus release in the absence of tetherin ( Figure 1D and E ) . In agreement with the virus rescue in trans , NL4 . 3 Vpu ELV release was markedly defective in the presence of increasing tetherin doses , although it did display a residual antagonism of tetherin when compared to the full Vpu-defective deletion ( Figure 1D and E ) . We next examined the phenotypic basis for the defect in tetherin antagonism by Vpu ELV . Vpu stimulates the ubiquitin-dependent degradation of tetherin , most likely in lysozomal compartments . We infected HT1080 cells stably expressing human tetherin bearing an HA-tag in the extracellular domain ( HT1080/tetherin-HA ) with VSV-G-pseudotyped HIV-1 wt , HIV-1 delVpu or HIV-1 Vpu ELV at an MOI of 2 to ensure >90% cell infection . 48 h later the cells were lysed and Western blotting performed for relative tetherin-HA levels ( Figure 2A ) . As expected , cells infected with HIV-1 wt showed reduced steady state levels of tetherin that was not apparent in those infected with HIV-1 delVpu . Similarly , in cells infected with HIV-1 Vpu ELV there was no evidence of tetherin degradation , but interestingly there appeared to be enhanced levels of tetherin , perhaps suggesting stabilization of the protein in the presence of the mutant Vpu . Thus E59 , L63 , V64 mutations abolish the ability of Vpu to induce tetherin degradation . Since this degradation is dependent on Vpu binding to β-TrCP2 via a phosphorylated pair of serines ( S52 and S56 ) [16] , [17] , [30] , we tested whether Vpu ELV mutants were defective for interaction with β-TrCP2 in co-immunoprecipitations from transfected cells ( Figure 2B ) . β-TrCP2 was co-immunoprecipitated with Vpu-HA and Vpu ELV-HA , but as expected , not the phospho-mutant Vpu 2/6-HA , ruling out this defect in Vpu ELV . Recent data suggests that ESCRT-mediated degradation of tetherin in the presence of Vpu is mediated by interaction of Vpu with HRS ( ESCRT-0 ) [19] . We could further show that both Vpu and Vpu ELV also co-precipitated with HA-HRS from transfected 293T cells ( Figure 2C ) indicating that an inability to recruit ESCRT-0 does not explain the defect in Vpu ELV-mediated degradation of tetherin . We then examined the ability of Vpu ELV mutants to downregulate surface tetherin levels . We first transfected HeLa cells ( that express tetherin constitutively ) with Vpu-HA expression vectors in combination with a GFP reporter . 48 h later surface tetherin was assayed by flow cytometry in the GFP positive cells ( Figure 2D ) . As expected , wildtype Vpu expression reduced cell surface tetherin levels . Vpu mutants bearing E59A , LV63 , 64A mutations , or the full ELV mutant all displayed a reduced capacity to downregulate surface tetherin levels ( Figure 2D and S1A ) . To confirm this in a relevant cell-type , we then infected CD4 positive Jurkat T cells with HIV-1 wt , HIV-1 delVpu , HIV-1 Vpu ELV or HIV-1 Vpu 2/6A . 48 h later , the cells were stained for surface tetherin and co-stained for intracellular p24CA as a marker of infection ( Figure 2E ) . To discriminate between truly infected cells , and those that acquired p24+ debris by exposure of cells to high titre ( MOI 1 ) of viral inoculum , we compared to cells exposed to virus in the presence of 50 µM AZT ( Figure S1C ) . Cultures infected with HIV-1 wt showed clear downregulation of tetherin on the surface of p24CA positive cells . By contrast , tetherin was not downregulated from the surface of either HIV-1 delVpu , HIV-1 Vpu 2/6A or HIV-1 Vpu ELV infected cells . Rather , tetherin levels were raised on some infected cells , perhaps reflecting accumulation of tethered virions on the cell surface . A similar result was observed for HeLa cells infected with the same viral stocks ( Figure S1B ) , although in this case we could detect no enhanced tetherin surface expression on cells infected with Vpu mutants , likely due to their endocytic removal from the cell surface [31] . The E59XXXL63V64 motif in Vpu resembles an acidic dileucine sorting signal ( D/E ) XXXL ( L/I/M/V ) found in the cytoplasmic tails of membrane proteins that traffic through endosomal compartments ( reviewed in [32] ) . We therefore addressed whether mutation of this motif affected Vpu subcellular localization . To this end we infected 293T expressing tetherin or not , as well as Jurkat and HeLa cells with VSV-G-pseudotyped HIV-1 NL4 . 3 and NL4 . 3 Vpu ELV and stained them for Vpu in combination with several subcellular markers 48 h later . As expected , in all cells , the predominant localization of wildtype Vpu was in association with the TGN , with between 20–50% of the Vpu immunoreactivity visible in TGN46+ compartments ( Figure 3A–D ) . The proportion of the Vpu ELV mutant in TGN46+ compartments was significantly reduced in 293T/tetherin , Jurkat and HeLa , and appeared as “endosome-like” puncta in the cytoplasm and associated on or near the plasma membrane ( Figure 3A–C and E ) . Interestingly in the parental 293T cells , which lack tetherin expression , Vpu and Vpu ELV localization was indistinguishable , and predominantly associated with TGN46+ compartments ( Figure 3D and E ) . Thus the difference in Vpu ELV localization appeared to be tetherin-dependent . The nature of these extra-TGN compartments was further analyzed in HeLa cells and revealed that Vpu ELV accumulated in EEA1+ early/sorting endosomal compartments , but not CD63+ late endosomes ( Figure 3F ) . Thus mutation of the EXXXLV motif leads to endosomal and surface localization of Vpu consistent with it being required for modulating the trafficking of tetherin . As a putative trafficking signal , the EXXXLV motif may exert its effect on tetherin antagonism in two ways . Firstly , the sequence may be required to permit Vpu to traffic to a compartment where it can interact with tetherin; secondly , the sequence may be required for post-interaction trafficking of Vpu/tetherin complexes such that tetherin is not incorporated into budding virions and expression on the cell surface is reduced . A potential confounding factor in investigating however is that wildtype Vpu induces tetherin degradation , while Vpu ELV does not . Whilst a Vpu 2/6 control maybe applicable , controversies surrounding the nature of its phenotype in terms of tetherin counteraction also make it problematic . To alleviate these issues , we took advantage of our recent observations with novel ESCRT-I component , UBAP1 [22] . UBAP1 contains 3-tandem ubiquitin-binding domains and is found in complex with the core ESCRT-I components TSG101 , VPS28 and VPS36 . However unlike them , UBAP1 is only required for ESCRT-dependent endosomal degradation and not viral assembly or cytokinesis [22] , [33] . Interestingly , UBAP1 is essential for both Vpu and K5-mediated degradation of tetherin , but is not required for Vpu-mediated tetherin antagonism , implying that commitment of tetherin into a degradative pathway by Vpu , but not ESCRT-I function itself , counteracts tetherin activity [22] . We therefore first examined whether Vpu or Vpu ELV interacted with tetherin in immunoprecipitations in the presence or absence of siRNA-mediated silencing of UBAP1 by quantitative Western blotting ( Figure 4A ) . 293T/tetherin cells were transfected twice over 48 h with either UBAP1 or control siRNA , before being infected with HIV-1 wt , HIV-1 Vpu ELV , HIV-1 delVpu , or HIV-1 Vpu A14L/W22A that contains a Vpu transmembrane mutation that abolishes tetherin interaction [14] . As expected , while tetherin levels were reduced in HIV-1 wt infected cells , they were unaffected by Vpu-defective or A14L/W22A mutants and , as in Figure 2 , stabilized in cells infected with the Vpu ELV mutant . Furthermore , as expected , UBAP1 siRNA treatment rescued tetherin levels from Vpu-mediated degradation , but also further enhanced total cellular levels of tetherin [22] consistent with the known role of ESCRT in tetherin's natural turnover [19] . Interestingly , UBAP1 siRNA also enhanced the total cellular content of wildtype Vpu to that of the Vpu ELV mutant ( approximately 4–5 fold ) . Immunoprecipitation of tetherin from the lysates from these cells revealed a similar picture . Wildtype Vpu was detected associated with residual tetherin precipitated from cells , and this was markedly increased upon UBAP1 knockdown . This data strongly suggests that Vpu itself may be co-degraded with tetherin in endosomal compartments . The Vpu ELV mutant efficiently co-precipitated with tetherin irrespective of UBAP1 knockdown , and as expected the A14L/W22A failed to co-precipitate under either condition . Interestingly , however , the ratio of relative band intensities between of Vpu or Vpu ELV precipitated with tetherin in the presence UBAP1 siRNA was equivalent , indicating that Vpu ELV was not defective for physical tetherin interaction , suggesting that the Vpu ELV mutant's defect in tetherin antagonism is due to an inability to mediate post-binding trafficking of Vpu/tetherin complexes into an ESCRT-dependent pathway in which both proteins are degraded . Consistent with this notion , Vpu ELV co-localized with tetherin in infected HeLa and 293T/tetherin , both in peripheral endosomal structures and at the cell surface ( Figure 4B ) . By contrast , the little tetherin visible in cells infected with wildtype virus co-localized with Vpu in perinuclear areas . Vpu is not a constituent of HIV-1 particles . We therefore reasoned that if the EXXXLV motif was required to prevent tetherin trafficking to viral budding sites and commit it for degradation , interaction with tetherin itself might lead to Vpu ELV mutant accumulation in nascent viral particles . To test this hypothesis we took advantage of a tetherin mutant lacking its GPI anchor ( tetherin-delGPI ) , which despite its high surface expression , does not restrict virus particle release , but accumulates in virus particles and is sensitive to Vpu [7] . 293T/tetherin-delGPI cells were mock-transfected or transfected with NL4 . 3 wt , NL4 . 3 Vpu ELV , and NL4 . 3 delVpu . 48 h later cell supernatants were centrifuged through a sucrose cushion and analyzed for tetherin incorporation ( Figure 4C and D ) . As expected , high levels of tetherin-delGPI could be detected in NL4 . 3 delVpu viral pellets , but not in pelleted supernatants from mock-transfected cells . Tetherin-delGPI incorporation was reduced in the wildtype virus consistent with tetherin removal from the cell surface . The level of tetherin incorporation in NL4 . 3 Vpu ELV particles was similar to that of the Vpu-defective mutant . Interestingly , NL4 . 3 Vpu ELV particles contained detectable levels of Vpu ( in this case Vpu appears as a doublet band which we suggest may be due to exposure to active HIV-1 protease in the particle ) . By contrast , no Vpu was detectable in any viral particles derived from 293T cells , indicating that Vpu ELV incorporation into viral particles was tetherin-dependent . Taken together these data indicate that the EXXXLV motif is required for efficient tetherin antagonism , by modulating the trafficking of tetherin such that it cannot become efficiently incorporated into nascent viral particles . Acidic-dileucine based sorting signals , D/EXXXL ( L/I ) , act as binding sites for a hemicomplex of sigma and adaptin subunits of the canonical clathrin adaptors AP-1 , AP-2 and AP-3 , and are required for endocytic and endosomal/Golgi trafficking of these proteins [32] . While the requirement for the acidic and first leucine residues are absolute , the third position is less well conserved , and can be L , I or on occasion V or M . Analysis of the cytoplasmic tails of Vpu sequences from most clades of HIV-1 group M , show that a putative EXXXL ( V/M/I ) is well conserved in the second alpha helix ( Figure S2A ) . In contrast to other HIV-1 subgroups , Clade C and F isolates have an EXXXLL motif juxtaposed to the plasma membrane in helix 1 ( not shown ) , which has been previously suggested to be a determinant of Clade C Vpu localization to the PM [34] . In subgroup B , the V64 position is usually V or M , although occasional I or L residues are found at this position . We mutated position 64 to M , L or I in NL4 . 3 Vpu and found no defect in these proteins' ability to counteract or downregulate tetherin from the surface ( Figure S2B and S2C ) , in agreement with the above data demonstrating that this position is the least important of the three , and consistent with the role of this motif as a sorting signal . The Nef proteins of primate immunodeficiency viruses are also multifunctional adaptor proteins , targeting a variety of immunoregulatory cell surface molecules for downregulation and degradation [35] . Nef interacts promiscuously with AP-1 , AP-2 and AP-3 through a conserved C-terminal EXXXLL motif [36] , [37] . The interaction of AP-2 with this site is essential for Nef targeting of CD4 for ESCRT-dependent lysozomal degradation [37] , but Nef-mediated downmodulation of Class I MHC molecules requires AP-1 [38] . Importantly several SIV Nef proteins are also tetherin antagonists [39] , [40] and again this is dependent on the EXXXLL motif [26] . We therefore asked whether the C-terminus of HIV-1 Vpu could be functionally substituted with a known AP-binding site from these proteins . The EVSALV motif of NL4 . 3 Vpu was first replaced with the core AP-binding site from NL4 . 3 Nef , ENTSLL ( Figure 5A ) . Since this is similar to sites already tested in the previous experiment shown in Figure S2 , this Vpu was as functional as the wildtype protein in virus rescue experiments ( Figure 5B and C ) . We then replaced the entire cytoplasmic tail of Vpu from residue 58 with a 19 amino acid stretch derived from Nef including the ENTSLL and a downstream dual-aspartic acid motif that has been previously shown to stabilize AP-2 interactions [41] . Remarkably , the Vpu/Nef chimeric protein substantially recovered tetherin antagonistic activity ( Figure 5B and C ) . Moreover , this chimera also displayed improved tetherin downregulation from the surface of transfected HeLa cells ( Figure 5D ) . This activity was entirely dependent on the key amino acids required for AP-interaction as a chimera in which E , LL and DD positions were mutated to alanine was unable to counteract tetherin or downregulate it from the surface ( Figure 5B–D ) . Examination of the subcellular distribution of Vpu-Nef or the mutant fused to CherryFP suggested that the mutant was localized more prominently to the PM consistent with a defect in trafficking imparted by the mutation ( Figure 5E ) . Thus Vpu function can be substantially recovered by replacing its entire second alpha helix with a promiscuous AP-binding ( D/E ) XXXL ( L/I ) sorting signal , indicating that linking Vpu directly to the clathrin trafficking machinery can restore its activity in absence of the second alpha helix . The implication of putative clathrin adaptor sites in Vpu-mediated tetherin antagonism led us to test whether inhibition of clathrin function inhibits Vpu activity . Overexpression of the C-terminal fragment of the neuronal adaptor AP180 ( AP180c ) that inhibits clathrin/membrane interactions [42] , which was recently shown to inhibit tetherin downregulation from the surface [43] , specifically blocked Vpu-dependent particle release of HIV-1 wt from 293T cells expressing tetherin ( Figure 6A ) , indicating clathrin-dependent subcellular trafficking is essential for Vpu activity . Infectious yield could not be determined in this experiment because AP180c overexpression inhibits envelope processing and blocks clathrin incorporation into particles , which has been shown to play a role in retroviral particle infectivity [44] , [45] . Interestingly AP180c expression , like UBAP1 siRNA treatment , enhanced total Vpu expression levels , indicating clathrin-dependent transport is involved in the turnover of Vpu . Visualization of Vpu-YFP localization in 293T/tetherin cells overexpressing AP180c showed vesicular rather than peri-nuclear localization similar to that seen in the same cells infected with HIV-1 Vpu ELV ( Figure 6B ) , and this was not apparent in the parental ( tetherin negative ) 293T cells , suggesting again this difference was driven by interaction with tetherin . The similarities between the EXXXLV motif and acidic dileucine sorting signals , and its functional replacement with the promiscuous ENTSLL motif in HIV-1 Nef led us to test whether the major known adaptors involved in trafficking of membrane proteins between the PM , endosomes and Golgi compartments were required for Vpu-mediated tetherin antagonism . Clathrin-mediated endocytosis requires AP-2 , whereas AP-3 controls early-to-late endosomal/lysosomal trafficking . AP-1 plays a role in trafficking of cargo between early endosomes and the TGN , with evidence that it can function in either direction [32] . To this end we examined the effects of siRNA-mediated silencing of AP-1 ( AP-1γ1 ) , AP-2 ( AP-2μ1 ) , AP-3 ( AP-3μ1/AP-3δ1 ) . Depletion of AP-2 by RNAi in 293T/tetherin cells had only minor effects on Vpu-dependent virus particle release , suggesting that unlike SIV Nef and HIV-2 Env , and consistent with the reports that Vpu does not enhance tetherin endocytosis , AP-2 activity is dispensable for Vpu-mediated tetherin antagonism ( Figure 6C ) . Similarly AP-3μ1/AP-3δ1 depletion had no detectable effect on tetherin activity or Vpu-mediated counteraction ( Figure 6D ) . Furthermore , human tetherin could also be downregulated from the surface of mouse fibroblasts defective in AP-3δ1 when transduced to express Vpu [46] ( Figure S3A ) . For AP-1 , siRNA-mediated silencing was inefficient in 293T ( not shown ) . We therefore constructed a HeLa cell line containing a doxycycline-inducible shRNA hairpin against AP-1γ1 . Induction of this hairpin coupled with simultaneous depletion of AP-1γ1 by oligonucleotide transfection led to approximately 95% knockdown efficiency . This treatment had no specific effect on Vpu-dependent virus particle yield in HeLa cells indicating that tetherin-antagonism again was not compromised ( Figure 6E ) , and furthermore tetherin surface downregulation was not defective in mouse AP-1γ1a −/− fibroblasts ( Figure S3B ) [47] . Finally , we examined whether there was any role for the retromer complex , in Vpu-mediated tetherin antagonism . Retromer regulates retrieval and recycling of endosomal proteins to the TGN and is known to act co-operatively or antagonistically with AP-1 ( reviewed in [48] ) . The retromer complex consists of several sorting nexins ( SNX ) , and a core complex containing cargo binding component , Vps34 , and two essential co-factors , Vps26 and Vps29 . We performed siRNA-mediated knockdown of Vps26 ( Figure 6F ) . At levels of knockdown that were sufficient to relocalize the CD8-cation-independent mannose-6-phosphate receptor ( CD8-CI-M6PR ) ( Figure S3C ) , disruption of retromer had no detectable effect of Vpu-mediated HIV-1 release from 293T/tetherin cells . Taken together these data demonstrate that neither depletion of individual cellular adaptor proteins known to bind to ( D/E ) XXXL ( L/I ) motifs , nor disruption of retromer-mediated endosome-to-TGN retrieval , were sufficient to recapitulate the phenotype of the Vpu ELV mutant . Recent data suggests that Vpu blocks both the transit of de novo synthesized tetherin to the cell surface as well as the recycling of tetherin endocytosed from the plasma membrane , with the relative importance of these processes currently a matter of debate . Tetherin recycling requires a dual tyrosine YXYXXV motif in its cytoplasmic tail that acts as a binding site for AP-2 ( for internalization ) and AP-1 ( for recycling via the Golgi ) [8] , [9] . Mutation of this site enhances tetherin's surface expression , but has minor effects on its ability to restrict virus release or its sensitivity to Vpu [11] , [13] . Given that the EXXXLV motif was defective for post-binding inactivation of tetherin , but retained a low residual activity against tetherin in transient transfection assays , we asked whether Vpu ELV was differentially defective against tetherin mutants bearing lesions in its own sorting sequence . We infected 293T/tetherin and 293T/tetherin Y6 , 8A cells with HIV-1 wt , HIV-1 delVpu and HIV-1 Vpu ELV at fixed dose ( MOI 1 ) and measured viral release 48 h later ( Figure 7A and B ) . Vpu-defective viral release was approximately 35-fold reduced from 293T/tetherin cells compared to the wildtype virus , and as expected NL4 . 3 Vpu ELV had an intermediate phenotype in this assay ( 6 fold less release than wt ) . However , in 293T/tetherin Y6 , 8A all residual antagonistic activity of Vpu ELV was abolished with viral release equivalent to that of the Vpu-deleted virus . By contrast the wildtype virus retained the majority of its anti-tetherin activity . This again was not due to a defect of Vpu interaction with tetherin , as immunoprecipitation of tetherin after UBAP1 siRNA treatment demonstrated equivalent levels of Vpu and Vpu ELV co-precipitation from both tetherin and tetherin Y6 , 8A expressing cells ( Figure 7C ) . Thus residual activity of Vpu ELV requires that tetherin retains its capacity to recycle from the PM . These data suggest that Vpu ELV is specifically defective in blocking tetherin transit to the PM , implying that Vpu/tetherin complexes are re-routed in a Golgi-associated compartment into a pathway that ultimately results in tetherin's endosomal destruction . In the absence of the EXXXLV sequence , tetherin/Vpu ELV complexes traffic to the PM . The residual activity of Vpu ELV therefore may be reflective of steric inhibition of tetherin function . The dominant-negative mutant of dynamin 2 ( K44A ) has also been described to have an intermediate effect on tetherin counteraction by Vpu , compared to the complete disruption of HIV-2 Env function which , like SIV Nef , is dependent on AP-2 and endocytosis [43] . We transfected increasing doses of HA-tagged dominant negative dynamin 2 or the wildtype protein along with HIV-1 proviruses into 293T/tetherin and parental cells ( Figure S4A ) . In agreement with Lau et al [43] dominant negative dynamin 2 , but not equivalent levels of the wild type dynamin 2 partially blocked the release of wildtype HIV-1 ( Figure S4A–D ) . Furthermore dominant negative dynamin 2 expression levels in these assays were sufficient to block transferrin uptake in parallel cultures ( Figure S4B ) . Interestingly dominant negative dynamin 2 also blocked residual Vpu ELV-mediated , and even the low level Vpu-defective viral release proportionally ( 7X , 4X and 8X for WT , Vpu-defective or Vpu ELV respectively ) , suggesting that this effect was independent of the ELV motif . Given that dominant negative dynamin 2 inhibits tetherin endocytosis [43] , this data suggests that its effect on restriction may be due more to the build up of tetherin at the cell surface that cannot be turned over rather than a direct effect on Vpu function itself . The results presented hitherto have demonstrated a requirement for the Vpu ELV motif in counteracting tetherin in cells stably expressing it or mutants thereof , and that its is required in constitutively-expressing target cells such as Jurkat to reduce surface tetherin levels . However some studies have cast doubt as to whether tetherin degradation and/or surface reduction is essential for Vpu function . Furthermore , in contrast to 293T/tetherin , release of HIV-1 Vpu ELV from HeLa was only 3-fold less efficient than wildtype in one round release ( Figure S5 ) . We therefore explored whether there was a requirement for the ELV motif in HIV-1 release from physiologically relevant target cells , namely Jurkat or primary human CD4+ T cells , particularly after treatment with type-1 interferon . Infection of Jurkat at an MOI of 1 resulted in a partial defect in release of the HIV-1 Vpu ELV mutant compared to the controls ( Figure 8A and B ) . However induction of higher tetherin expression by overnight treatment with universal type-1 interferon effectively reduced HIV-1 Vpu ELV particle release to levels similar to that of the Vpu-defective control while only reducing the wildtype release moderately . Similarly interferon treatment of purified activated human CD4+ T cells led to a selective defect in the production of cell-free HIV-1 Vpu ELV virions ( Figure 8C ) consistent with a concomitant upregulation of surface tetherin levels ( Figure 8D ) , and surface tetherin downregulation ( Figure 8E ) . Taken together with results presented above , these data demonstrate that the ability to downregulate and degrade tetherin imparted by the EXXXLV motif is required for cell-free virion release from relevant primary HIV target cells , and becomes essential when tetherin expression is enhanced by an antiviral stimulus .
In this study we have identified a determinant within the second alpha helix of the cytoplasmic tail of HIV-1 NL4 . 3 Vpu , E59XXXL63V64 , which is required for efficient antagonism of tetherin . Mutation of this site blocks the ability of Vpu to mediate tetherin downregulation from the cell surface and its ESCRT-dependent degradation , but does not abolish its interaction with tetherin , nor recruitment of β-TrCP2 or the ESCRT-0 component HRS . Importantly , this motif is required to counteract tetherin in CD4+ T cells , particularly after their exposure to type-I interferon . Vpu ELV mutants localize to the cell surface and early/recycling endosomal compartments rather than the TGN by virtue of their interaction with tetherin . This is consistent with a role for this determinant in Vpu-mediated inhibition of the transit of newly synthesized and/or recycling tetherin to the cell surface , its internal sequestration and targeting for endo-lysosomal degradation . However , while we could functionally complement Vpu function by grafting the D/EXXX ( L/I/M ) motif from HIV-1 Nef in place of helix-2 , we could detect no effect of depletion of the canonical clathrin adaptor proteins AP-1 , AP-2 or AP-3 on Vpu-mediated tetherin antagonism . The inhibition by AP180c overexpression , however , implicates clathrin function in Vpu-mediated tetherin antagonism . Finally , residual Vpu ELV activity against tetherin was entirely dependent on an intact recycling motif in tetherin's cytoplasmic tail , suggesting that this motif differentially affects antagonism of newly synthesized tetherin rather than pre-existing pools recycling to the PM . These results are in contrast to the recent study by Lau et al who were unable to demonstrate a phenotype for an LV63 , 64AA mutant despite interfering with Vpu activity with AP180c [43] . Given that we have demonstrated phenotypes in several cellular systems including CD4+ T cells for the Vpu ELV mutant , the reason for this discrepancy is unclear . There has been much debate as to whether the reduction of tetherin levels at the plasma membrane is required to counteract its antiviral activity , particularly in CD4+ T cells ( reviewed in [10] ) . In our hands , tetherin surface levels are reduced in HIV-1 infected primary T cells and Jurkat cells . However , under these conditions , the viral release phenotype of the Vpu ELV mutant is only a few-fold different than the wildtype protein . Interestingly this changes upon treatment of the cells with type-I interferon , which upregulates tetherin expression levels . In this case Vpu ELV mutant release is reduced almost to that of the Vpu-defective virus , but only minor further reductions are observed for the wildtype virus . This therefore indicates that the requirement for surface reduction/degradation of tetherin becomes much more important at higher expression levels of the restriction factor ( something that has been suggested previously by Goffinet et al [49] ) . Recent conflicting data have addressed the effect of tetherin and interferon on cell-cell transmission between T cells [50] , [51] , which when taken together show that if tetherin does restrict this mode of virion transfer , it is far less efficient than its effect on cell-free viral release . Thus , the fact that tetherin antagonism is such a highly conserved attribute amongst primate immunodeficiency viruses implies its importance in vivo . Since interferon treatment of CD4+ T cells magnifies the defective phenotype of a Vpu ELV mutant , the ability to mediate tetherin's surface reduction and target it for endosomal degradation is likely to be essential for the virus to avoid restriction under proinflammatory conditions it is likely to encounter in vivo , particularly during acute infection [52] , [53] . Our results are consistent with the notion that Vpu blocks delivery of tetherin to the plasma membrane [13] , [54] and suggest that Vpu exerts its effect on tetherin trafficking in the TGN/recycling compartment in a manner determined in part by this putative sorting sequence . It is now clear that Vpu-mediated tetherin degradation is ubiquitin-dependent and occurs in lysosomes rather than early reports of proteasomal processing [15] , [16] , and that this process requires the ESCRT pathway [19] . Depletion of both TSG101 and Vps4 inhibits Vpu-mediated tetherin degradation , and recruitment of HRS ( ESCRT-0 ) has been reported to be required for Vpu-mediated tetherin antagonism [19] , as has ubiquitination on multiple residues in the tetherin cytoplasmic tail [18] . However tetherin's ultimate degradation itself is not essential for Vpu or other lentiviral countermeasures to inactivate it . Our recent observations with the novel ESCRT-I subunit , UBAP1 , ( [22] and the results presented herein ) , demonstrate that ESCRT-I function is unlikely to be required for tetherin antagonism , but its commitment to endosomal degradation is . Interestingly , the concomitant enhancement of wildtype Vpu levels , both co-precipitating with tetherin and at steady state , suggest that Vpu is likely co-degraded with its target . Alongside the current literature our data suggests that Vpu interaction with tetherin leads to a differential trafficking of tetherin in the TGN and/or recycling compartments rather than enhancing tetherin internalization ( Figure 9 ) . This inhibits forward trafficking of either recycling or newly synthesized tetherin to the PM , and commits it to a pathway that ultimately targets it to endolysosomal compartments for degradation . We suggest that it is this commitment , regulated by the EXXXLV motif in second helix , rather than degradation per se , that is principally responsible for antagonizing tetherin . Where Vpu interacts with tetherin in the cell is likely to be related temporally to the viral replication cycle , tetherin expression level , and its natural turnover rate . Vpu is expressed “late” in replication from the same mRNA as Env [55] , at the time new virions are being built . Thus Vpu must deal with two pools of tetherin; pre-existing protein in the periphery recycling via the TGN [9] , and de novo synthesized tetherin trafficking through the Golgi en route from the endoplasmic reticulum ( ER ) to the surface . Therefore we predict that Vpu must interact with tetherin in TGN-associated compartments to engage these two pools of tetherin , although our recent data suggests that binding to newly synthesized tetherin may occur prior to this in the ER [14] . Thereafter , a clathrin-dependent sorting event determined by the helix-2 EXXXLV motif precludes Vpu/tetherin complexes transiting to the PM . This intracellular sequestration of tetherin is further coupled to late endosomal targeting and ESCRT-dependent degradation through the recruitment of HRS and tetherin ubiquitination [18] , [19] . In line with this , global disruption of early to late endosomal transition by dominant negative Rab7a also appears to inhibit Vpu activity [23] . Interestingly , ubiquitinated cargo destined for ESCRT-dependent degradation via HRS recruitment has been shown to partition differentially to areas of early/sorting endosomal membranes rich in flat clathrin lattices , thereby anchoring it away from the recycling machinery [56] . Recruitment of tetherin into such structures in the sorting compartment may be sufficient to antagonize its function , offering an explanation as to why clathrin is essential for Vpu activity but dynamin 2 mutants only effect the residual activity of Vpu ELV . Because Vpu activity was insensitive to retromer ( Vps26 ) depletion , which is essential for endosome-to-TGN retrieval and recycling [48] , we suggest that Vpu targets tetherin into this pathway from the TGN to endosomal compartments , and an ensuing swift degradation accounts for the observed sequestration in TGN46 positive compartments . In the absence of the EXXXLV motif , tetherin is not committed to this differential sorting in the TGN , and Vpu/tetherin complexes are targeted to the cell surface and thereafter back into the recycling system , accounting for the localization of Vpu ELV in infected cells and its incorporation into virions . Therefore in addition to acting as an adaptor for the recruitment of ESCRT-0 and E3 ubiquitin ligase activity to tetherin , we propose that through the EXXXLV motif , Vpu directly chaperones associated tetherin molecules into an endosomal compartment from which they cannot recycle to the PM . EXXXLV mutants retain some residual antagonistic activity that is entirely dependent on the AP-1/AP-2-binding motif in tetherin . This would suggest that the action of the EXXXLV motif does not solely account for the inhibition of tetherin by Vpu . Vpu ELV still interacts with β-TrCP2 , suggesting that tetherin ubiquitination , which is important for counteracting its activity [18] , may still take place during the recycling process and effect antiviral function . Also physical interaction between Vpu and tetherin may be sufficient to interfere with some level tetherin function provided tetherin can still associate with the clathrin-dependent endocytic machinery . Since biochemical evidence strongly favors a direct cross-linking model for tetherin's antiviral activity [7] , either of these processes could in principle affect the qualitative nature of tetherin distribution at viral assembly domains at the PM such that tethers do not form efficiently , something that is plausible in the light of Habermann et al's quantitative Cryo-EM analysis of tetherin localization at the PM in infected HeLa cells [57] . The relative efficiency of this “secondary” inhibition of tetherin activity will thus be dependent on cellular tetherin levels and the temporal stage of viral replication ( ie: the expression level of Vpu ) . Since Vpu-defective mutants of HIV-1 in some studies show enhanced cell-to-cell transmission and that tetherin may play a role in virological synapse formation [51] , [58] , [59] , such a differential effect on de novo synthesized versus the recycling pool tetherin may favor viral transmission during early stages of viral production , but effectively antagonize induction of tetherin expression and potent restriction by a pro-inflammatory response . We have so far been unable to demonstrate direct interaction between the Vpu EXXXLV motif and adaptors AP-1 , 2 or 3 that are known to bind acidic dileucine signals via a hemicomplex of their sigma and adaptin subunits [32] . These interactions have been demonstrated in yeast 3-hybrid assays previously for HIV-1 Nef [36] , [37] , but not all interactions are amenable to this method and in our hands the Vpu cytoplasmic tail is a constitutive activator of transgene expression precluding its use . Furthermore , the low affinity of these interactions , their transient nature , and the complex interactions of the adaptor protein with membrane lipids means that adaptor binding is rarely measurable in co-precipitations from cells . The context dependency of ( D/E ) XXXL ( L/I ) motifs that governs which adaptor they bind to in vivo is poorly defined at present , meaning that RNAi-depletion is the most reliable method for identifying the cellular factor involved . While individual depletion of AP-1 , 2 or 3 in our experiments had no effect on Vpu-mediated tetherin antagonism , replacement of the second alpha helix with the promiscuous AP-binding peptide from HIV-1 Nef did recover function . This site in Nef is essential for several functions including CD4 and MHC Class I downregulation , the former being AP-2-dependent [37] , the latter requiring AP-1 [38] . Importantly , binding of AP-2 to this site in SIV Nef proteins is required for their counteraction of non-human primate tetherins [26] . Thus there are several possibilities . The EXXXLV motif in Vpu might be a promiscuous AP-binding site and adaptor usage may be redundant . The subcellular localization of tetherin/Vpu ELV complexes in EEA1 positive endosomes and the partial effect of dominant negative dynamin 2 would argue against a significant role for AP-2 , the major regulator of clathrin-mediated transport from the PM . However toxicity associated with simultaneous knockdown of multiple adaptors has precluded us from addressing this possibility so far . Alternatively , replacement of the second alpha with the Nef peptide recovers function because it confers AP-2 binding to Vpu , thus allowing Vpu to counteract tetherin in a manner similar to SIV Nef and HIV-2/SIV Envs that both require AP-2 interactions [24]–[26] , [60] . However , these data do underscore that linking Vpu to the clathrin trafficking machinery promotes its ability to counteract tetherin . If Vpu-mediated tetherin antagonism is clathrin-dependent , but independent of AP-1 , AP-2 or AP-3 ( for which the importance of clathrin is debated ) , what other adaptors might be important ? Two further heterotetrameric adaptors , AP-4 and AP-5 , have been identified , but understanding of their role in subcellular trafficking is limited and at present they are not known to bind to acidic dileucine motifs or to control clathrin-mediated transport [32] , [61] . The monomeric GGA ( Golgi-associated , γ-ear containing , ARF-cofactors ) 1–3 proteins that also function as clathrin adaptors that regulate Golgi-to-endosome transport are potentially attractive candidates , whose dysregulation has been reported to inhibit HIV-1 assembly [62] . However , known GGA-binding motifs in cellular cargoes correspond to a DXXLL consensus where the spacing of the leucines from the acidic residue is thought critical . The ( E/D ) XXXL ( V/M/I ) motif is conserved in Vpu proteins from HIV-1 M group subtypes A , B , D , G and H ( Figure S2 ) , as well as in Group O ( www . hiv . lanl . gov ) . While Group O Vpus cannot antagonize tetherin [63] , this maps to defective TM domain-mediated interaction and the membrane proximal hinge region . When replaced by those from a Group M Vpu these are sufficient to confer tetherin inactivation implying that C-terminal determinants retain function [64] . In Group M clades C and F , the equivalent position is ( D/E ) XXXL ( S/A ) respectively , suggesting that the site may not be functional in those Vpus , with the caveat that the second position in the dileucine motif is less important . Intriguingly both these subtypes bear EYXXL ( L/I ) motifs in the membrane proximal region of their cytoplasmic tails that encompass both a putative tyrosine and dileucine-based sorting sequence . Evidence from Ruiz et al [34] has shown that a model subtype C Vpu localizes to the PM rather than the TGN and that this is in part due to this motif . While LL mutations in this subtype C Vpu confer T-cell line replication phenotypes suggestive of a failure to downmodulate tetherin , no direct experiments on the role of this motif in tetherin-mediated HIV-1 release have been thus far performed , nor whether this site can bind known clathrin adaptors . In summary our data implicate a trafficking determinant in the Vpu cytoplasmic tail that is required for tetherin downregulation , degradation and efficient antagonism , and suggest that it governs differential sorting of Vpu/tetherin complexes in the TGN to prevent forward transit of tetherin to the PM and viral budding sites .
HEK293T , HeLa and Jurkat cells were obtained from ATCC ( American Tissue Culture Collection ) . 293T/tetherin and 293T/tetherin-delGPI and HT1080/tetherin-HA are cell lines stably expressing human tetherin or mutant thereof , with or without a hemagglutinin ( HA ) epitope tag inserted at nucleotide 463 , which has been previously described [1] , [65] . The reporter cell line HeLa-TZMbl , was kindly provided by John Kappes through the NIH AIDS Reagents Repository Program ( ARRP ) . All adherent cells were maintained in Dulbecco's modified Eagle medium ( DMEM ) ( Invitrogen , UK ) supplemented with 10% fetal calf serum and Gentamycin; T-cell lines were grown in Roswell Park Memorial Institute medium ( RPMI ) supplemented with 10% fetal calf serum and Gentamycin . Murine fibroblasts from pearl and AP1γ1a deficient mice and derivative in which AP3μ or AP1γ1a were re-expressed were kindly provided by Andrew Peden [46] and Peter Schu [47] respectively . These cells were transduced to express HA-tagged human tetherin using pLHCX-THN-HA463 [65] and maintained in hygromycin selection . Wildtype HIV-1 NL4 . 3 ( obtained from NIH-ARRP ) , a Vpu-defective counterpart and pCR3 . 1 Vpu-HA containing a modified codon optimized NL4 . 3 Vpu has been described previously [31] . All second alpha helix mutants of Vpu and mutations in the NL4 . 3 proviral genome were generated by Quick-change site-directed mutagenesis PCR according to standard protocols using Phusion-II polymerase ( New England Biolabs ) . Vpu-Nef chimeras and corresponding mutants were made with long reverse PCR primers encoding Nef clathrin adaptor binding sites , cloned into pCR3 . 1 expression vectors encoding tagged or untagged tetherins which have been described elsewhere [1] . β-TrCP2 was cloned from HeLa cDNA and inserted into pCR3 . 1 with a C-terminal myc-tag . pCR3 . 1 HA-HRS was kindly provided by Juan Martin-Serrano [66] . pCR3 . 1 dynamin 2-HA and dominant negative dynamin 2-HA have been previously described by [31] . Primary human CD4+ T cells were isolated from fresh venous blood drawn from healthy volunteers . CD4+ T cells were purified from total peripheral blood mononuclear cells ( PBMC ) isolated by lymphoprep ( AXIS-SHIELD ) gradient centrifugation using a CD4+ T cell Dynabeads isolation kit ( Invitrogen ) . T cells were then activated for 48 h using anti-CD3/anti-CD28 magnetic beads ( Invitrogen ) . The beads were then removed cells were then maintained in rhIL-2 ( 20 U/ml ) ( Roche ) . For full-length HIV-1 stocks pseudotyped with the Vesicular Stomatitis Virus Glycoprotein ( VSV-G ) , 293T cells were transfected with 2 µg of proviral plasmid and 200 ng of pCMV VSV-G . 48 h post-transfection , viral stocks were harvested and endpoint titers were determined on HeLa-TZMbl cells as described below [25] . For transient-transfection-based virus release assays , subconfluent 293T cells were plated on 24 well plates and transfected with 500 ng proviral clone , in combination with 50 ng of tetherin and 25 ng of Vpu-HA or mutants using 1 µg/ml polyethyleneimine ( Polysciences ) . The medium was replaced 5 h and 16 h post-transfection , cells were harvested after 48 h . The infectivity of viral supernatants was determined by infecting HeLa-TZMbl , 48 h later cells were assayed for β-galactosidase activity using the chemiluminescence Tropix GalactoStar kit ( Applied Biosystems ) . For biochemical analysis of virus particle release , supernatants were filtered ( 0 . 22 µm ) and pelleted through a 20% sucrose/PBS cushion at 20 , 000 g for 90 min at 4°C , and pellets were lysed in SDS-PAGE loading buffer . Virion and cell lysates were then subjected to SDS-PAGE and Western blotted for HIV-1 p24CA ( monoclonal antibody 183-H12-5C; kindly provided by B Chesebro through the NIH ARRP ) , rabbit anti-Hsp90 ( Santa Cruz Biotechnologies ) , monoclonal mouse anti-HA . 11 ( Covance ) , polyclonal rabbit anti-HA ( Rockland ) and/or Vpu ( rabbit polyclonal; kindly provided by K . Strebel through the NIH ARRP [67] , and visualized by LiCor apparatus using fluorophores conjugated secondary antibodies ( IRDye 800 Goat anti-rabbit , IRDye 680 Goat anti-mouse ) . 5×105 cells ( 293T/tetherin , Jurkat or CD4+ T cells ) were infected with VSV-G-pseudotyped HIV-1 wt , HIV-1 delVpu or HIV-1 Vpu ELV at an MOI of 0 . 5–1 . 16 h post infection medium was replaced and cells ( treated or not with 5000 U/ml of universal type-1 interferon ( PBL InterferonSource ) ) were cultured for a further 24 h . The cells harvested , the infectivity of viral supernatants was determined by infecting HeLa-TZMbl and biochemical analysis of virus particle release was performed as in Virus release assay . For examining tetherin degradation HT1080 cells stably expressing tetherin-HA were infected with VSV-G-pseudotyped HIV-1 wt , HIV-1 delVpu or HIV-1 Vpu ELV virus stocks at a multiplicity of infection ( MOI ) of 2 to ensure that approximately 90% of the cells were infected . The medium was replaced 4 h after infection . 48 h post infection cell lysates were harvested and processed as described above . HeLa cells were transfected with 400 ng of pCR3 . 1 GFP and 400 ng of pCR3 . 1 Vpu-HA or indicated mutants . 48 h post transfection the cells were harvested and stained for surface tetherin using a specific anti-BST2 monoclonal IgG2a antibody ( Abnova ) and goat-anti-mouse IgG2a-Alexa633 conjugated secondary antibody ( Molecular Probes , Invitrogen , UK ) . Tetherin expression on GFP positive cells was then analyzed using a FacsCalibur flow-cytometer ( Becton Dickinson ) and the FlowJo software . Murine fibroblasts were transduced with the pMigR1-based retroviral vector pCMS28-IRES-eGFP or a derivative expressing NL4 . 3 Vpu . 48 h after transduction the cells were stained for surface HA versus GFP expression . Jurkat or CD4+ T cells were infected with VSV-G-pseudotyped HIV-1 wt , HIV-1 delVpu or HIV-1 Vpu ELV at an MOI of 1 . 48 h post infection cells were stained for surface tetherin expression as above , then fixed and permeabilized for 20 minutes ( Cytofix/cytoperm Fixation/Permeabilization kit , BD Biosciences ) and stained for intracellular HIV-1 p24CA using the KC57 antibody conjugated to PE ( Beckman- Coulter ) . Cells were grown on coverslips and infected with VSV-G-pseudotyped HIV-1 wt or HIV-1 Vpu ELV , 48 h later cells were fixed in 4% paraformaldehyde/PBS , washed with 10 mM glycine/PBS , and permeabilized in 1% bovine serum albumin/0 . 1% Tritin-X100/PBS for 15 min . The infected cells were stained using anti-rabbit polyclonal Vpu in combination with sheep anti-human TGN46 ( AbD Serotec ) , mouse anti-EEA1 ( BD Biosciences ) , mouse anti-CD63 ( Developmental Studies Hybridoma Bank , University of Iowa ) or mouse polyclonal anti-BST-2 ( Abnova ) followed by the appropriate secondary antibodies conjugated to Alexa 488 or 594 fluorophores ( Molecular Probes , Invitrogen ) . The cells were then mounted on glass slides using ProLong AntiFade- 4′ , 6-diamidino-2-phenylindole ( DAPI ) mounting solution ( Molecular Probes , Invitrogen ) . Cells were visualized with a Leica DM-IRE2 confocal microscope . Images were analyzed using Leica Confocal Software and ImageJ . 293T cells stably expressing tetherin were transfected twice over 48 h with siRNA oligonucleotide against UBAP1 targeting CTCGACTATCTCTTTGCACAT or Non-targeting siRNA was used as control ( Dharmacon ) . The cells were then infected with VSV-G-pseudotyped HIV-1 wt , HIV-1 delVpu , HIV-1 Vpu ELV or HIV-1 Vpu A14L , W22A at an MOI of 2 . 48 h post infection the cells were lysed on ice for 30 min in buffer containing 50 mM Tris-HCL pH 7 . 4 , 150 mM NaCl , complete protease inhibitors ( Roche ) and 1% digitonin ( Calbiochem ) . After removal of the nuclei , the supernatants were immunoprecipitated with 5 µg/ml mouse monoclonal anti-BST2 antibody ( eBiosciences ) for 1 . 5 h at 4°C . Sepharose-protein G beads were washed in lysis buffer before they were added to the samples and incubated for further 3 h . The beads were washed extensively in lysis buffer containing 0 . 1% digitonin and resuspended in SDS-PAGE loading buffer . Cell lysates and immunoprecipitates were subjected to SDS-PAGE , and Western blot assays were performed using a rabbit anti-Vpu antibody ( kindly provided by K Strebel through the NIH ARRP ) , polyclonal rabbit anti-tetherin antibody ( kindly provided by K Strebel through the NIH ARRP ) and polyclonal rabbit anti-UBAP1 antibody ( Proteintech ) , and visualized by ImageQuant using corresponding HRP-linked secondary antibodies ( New England Biolabs , UK ) . For HRS/Vpu coIP , 293T cells were co-transfected with 700 ng of pCR3 . 1 HA-HRS and pCR3 . 1 Vpu-YFP , pCR3 . 1 Vpu ELV-YFP or pCR3 . 1 YFP expression plasmids . 48 h post transfection the cells were lysed in buffer containing 0 . 1 M MES-NaOH pH 6 . 5 , 1 mM magnesium acetate , 0 . 5 mM EGTA , 200 µM sodium ortho-vanadate , 10 mM NEM , complete protease inhibitors ( Roche ) and 1% digitonin . After removal of the nuclei , the supernatants were immunoprecipitated with 5 µg/ml monoclonal mouse anti-HA . 11 antibody ( Covance ) . Immunoprecipitation was performed as described above and Western blot assays were performed using a polyclonal rabbit anti-HA antibody ( Rockland ) and an anti-Vpu antibody . 293T cells were co-transfected with 700 ng of pCR3 . 1 myc β-TrCP 2 and pCR3 . 1Vpu-HA , pCR3 . 1 Vpu ELV-HA , pCR3 . 1 Vpu 2/6A-HA or pCR3 . 1 YFP expression plasmids . 48 h post transfection , Crosslinking Immunoprecipitation was performed as previously described [68] . Cell lysates and immunoprecipitates were subjected to SDS-PAGE , and Western blot assays were performed using a rabbit anti-Vpu antibody and mouse anti-myc antibody ( kindly provided by M . Malim ) , and visualized by ImageQuant ( GE ) using corresponding HRP-linked secondary antibodies ( New England Biolabs ) . 293T cells stable expressing tetherin or HeLa cells were seeded at a density of 2×105 cells per well in a 12 well plate . After 3 h , the first transfection was performed . For each well , 3 µl Oligofectamine ( Invitrogen ) was added to 10 µl of Opti-MEM ( Life Technologies ) , this solution was added to 5 µl of 20 µM siRNA in 85 µl of Opti-MEM according to manufactures protocol . For AP-1 knockdown , HeLa cells stably expressing doxycycline-inducible pTRIPZ shRNA against AP-1γ1 ( OpenBiosystems ) were used in combination with siRNA oligonucleotide against AP-1γ1 targeting AAGAAGATAGAATTCACCTTT . For AP-2 knockdown , SMARTpool siRNA targeting the AP-2 µ1 subunit was used ( Dharmacon ) . For AP-3 knockdown , SMARTpool siRNA targeting the AP-3 µ1 subunit was used in combination with SMARTpool siRNA targeting the AP-3δ1 subunit ( Dharmacon ) . For Vps26 knockdown , siRNA targeting the AACCACCTATCCTGATGTTAA sequence was used ( Qiagen ) . On Non-targeting siRNA was used as control ( Dharmafect ) . The cells were reseeded into a 24 well plate on day 2 and a second transfection was performed according to manufactures protocol . The cells were infected 3 h post transfection with VSV-G-pseudotyped HIV-1 wt , HIV-1 delVpu at an MOI of 0 . 8 . The infectivity of viral supernatants was determined by infecting HeLa-TZMbl as described above . Cell lysates and viral particles were subjected to SDS-PAGE , and Western blot assays were performed using a mouse monoclonal AP-1γ1 antibody ( Sigma ) , mouse anti-AP50 ( AP-2μ1 ) and mouse anti-AP-3δ1antibodies ( BD Bioscience ) , polyclonal rabbit AP-3μ1 antibody ( kindly provided by M . S . Robinson ) and rabbit polyclonal Vps26 antibody ( Abcam ) . Ethical approval for the drawing of blood and preparation of leukocyte subsets from healthy donors following written informed consent was obtained through the King's College London Infectious Disease BioBank Local Research Ethics Committee ( under the authority of the Southampton and South West Hampshire Research Ethics Committee – approval REC09/H0504/39 ) , approval number SN-1/6/7/9 .
|
Tetherin inhibits the release of several diverse enveloped viruses from infected cells and is counteracted by the HIV-1 accessory gene Vpu . Vpu prevents tetherin's incorporation into nascent viral particles , promotes its downregulation from the cell surface and targets tetherin for degradation . Here we identify a determinant that resembles an acidic-dileucine-based sorting sequence in the Vpu cytoplasmic tail that is required for efficient counteraction of tetherin activity , particularly in CD4+ T cells treated with type-1 interferon . Mutation of this motif prevents cell-surface downregulation and degradation of Vpu/tetherin complexes but does not affect their interaction . Rather , in its absence , Vpu accumulates in early endosomes and at the cell surface where it becomes incorporated into assembling virions with tetherin , indicating that this motif modulates sub-cellular trafficking of tetherin . Furthermore Vpu activity is clathrin-dependent and can be reconstituted by replacing a portion of the cytoplasmic tail encompassing this motif with one derived from HIV-1 Nef that is known to bind several clathrin adaptors . Finally , we demonstrate that residual function of the mutant Vpu requires a trafficking motif in tetherin , suggesting that physical interaction of tetherin with Vpu during its recycling to the cell-surface can interfere with its function to a variable extent .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"infectious",
"diseases",
"immunology",
"biology",
"molecular",
"cell",
"biology"
] |
2012
|
A Cytoplasmic Tail Determinant in HIV-1 Vpu Mediates Targeting of Tetherin for Endosomal Degradation and Counteracts Interferon-Induced Restriction
|
Protein misfolding and aggregation are typically perceived as inevitable and detrimental processes tied to a stress- or age-associated decline in cellular proteostasis . A careful reassessment of this paradigm in the E . coli model bacterium revealed that the emergence of intracellular protein aggregates ( PAs ) was not related to cellular aging but closely linked to sublethal proteotoxic stresses such as exposure to heat , peroxide , and the antibiotic streptomycin . After removal of the proteotoxic stress and resumption of cellular proliferation , the polarly deposited PA was subjected to limited disaggregation and therefore became asymmetrically inherited for a large number of generations . Many generations after the original PA-inducing stress , the cells inheriting this ancestral PA displayed a significantly increased heat resistance compared to their isogenic , PA-free siblings . This PA-mediated inheritance of heat resistance could be reproduced with a conditionally expressed , intracellular PA consisting of an inert , aggregation-prone mutant protein , validating the role of PAs in increasing resistance and indicating that the resistance-conferring mechanism does not depend on the origin of the PA . Moreover , PAs were found to confer robustness to other proteotoxic stresses , as imposed by reactive oxygen species or streptomycin exposure , suggesting a broad protective effect . Our findings therefore reveal the potential of intracellular PAs to serve as long-term epigenetically inheritable and functional memory elements , physically referring to a previous cellular insult that occurred many generations ago and meanwhile improving robustness to a subsequent proteotoxic stress . The latter is presumably accomplished through the PA-mediated asymmetric inheritance of protein quality control components leading to their specific enrichment in PA-bearing cells .
Proper protein folding and maintenance of proteome integrity are essential for cell function and viability [1 , 2] . Nevertheless , the generation of nonnative protein conformations is inevitable to some extent because of the inherent stochastic nature of protein folding [3 , 4] and is often even aggravated by genetic ( e . g . , mutations [5] ) , physiological ( e . g . , cellular aging [6] ) , and environmental ( e . g . , heat [7] or antibiotics [8] ) conditions . Such aberrant protein conformations typically expose hydrophobic residues and regions normally buried within their native structure , which drive nonfunctional intermolecular interactions leading to the formation of larger insoluble structures termed protein aggregates ( PAs ) [9] . In prokaryotes , the occurrence of PAs is regarded as a strictly adverse phenomenon . In the model bacterium E . coli , for example , intracellular PAs are perceived as naturally occurring , inevitable , and constantly accruing “garbage bins” [10–12] . This view is based on single cell–level observations of growing E . coli cells , in which the yellow fluorescent protein ( YFP ) -fused inclusion body binding protein ( IbpA-YFP ) was employed as a reporter for the presence of aggregated proteins [10 , 13] . Such PAs were observed to appear randomly during growth and to segregate asymmetrically upon cell division due to their nucleoid-enforced polar localization [10 , 14–16] . Moreover , PA accumulation in cells with older cell poles was shown to slow down their growth in a dosage-dependent manner , thereby giving rise to the previously observed pattern of aging and rejuvenation in growing microcolonies [10 , 17] . It was even theorized that this asymmetric damage segregation strategy was superior to damage repair strategies in most environmental conditions and could thus have evolved as an optimal damage riddance strategy [18–21] . This negative perception contrasts with emerging findings in eukaryotes that illustrate the potential functionality of stress-induced misfolded and aggregated proteins . The proteome-wide characterization of aggregation tendencies during thermal stress in the yeast Saccharomyces cerevisiae identified heat-induced aggregation and stress-granule formation as specific , reversible , and promoting cellular adaptation and survival [22 , 23] . Similarly , nutrient depletion of yeast cells leads to the formation of metabolite-specific , reversible protein assemblies that have been proposed to function as storage depots but at the same time might also enhance metabolic efficiency during nutrient stress [24–26] . Detailed characterization and quantification of the Caenorhabditis elegans proteome along its lifespan showed extensive proteome remodeling and suggested the sequestration of proteins in insoluble aggregates to be a protective strategy directed toward maintaining proteome integrity during aging [27] . In fact , the term quinary structures has been coined to describe such functional ( stress-induced ) large protein assemblies , as they lack the fixed stoichiometry specific to quaternary assemblies [28 , 29] . However , direct phenotypic evidence of their beneficial impact has remained limited . In addition , it remains unclear to what extent the formation of such structures affects and shapes future behavior , after alleviation of the stress conditions . Given these emerging examples of functional protein aggregation in eukaryotes , we set forward to revisit the occurrence and potential impact of ( stress-induced ) protein aggregation in prokaryotes using E . coli as a model system . Using validated fluorescent PA reporters , we could demonstrate that intracellular PAs only emerge as a response to sublethal proteotoxic stresses , after which these structures become asymmetrically segregated and gradually disaggregated . Remarkably , rather than being fitness compromised , we found that cells asymmetrically inheriting an ancestral PA became significantly more resistant to a subsequent stress compared to their isogenic , PA-free siblings . Our results therefore reveal that PAs can serve as epigenetically inheritable memory elements that enable long-term cross-generational reminiscence of previous cellular insults .
The E . coli IbpA protein belongs to the conserved family of ATP-independent small heat shock proteins that readily associate with misfolded and aggregated proteins [30–32] . Previously , E . coli strains expressing an IbpA-YFP fusion protein have been used to microscopically visualize intracellular PAs , which appeared as punctate and polarly located fluorescent foci inside the cytoplasm of healthy , living cells [10 , 15 , 33] . However , given the potential of many commonly used fluorescent proteins to cause label-induced oligomerization and trivial foci formation , this IbpA-YFP reporter has been suggested to yield a biased view on intracellular PA dynamics [34 , 35] . In order to examine this bias more closely and look for a more reliable PA reporter , we created a set of nearly identical E . coli MG1655 strains only differing in the fluorescent reporter that was translationally fused to the 3′-end of the native chromosomal ibpA gene ( Fig 1A , S1A Fig , and S1 Table ) . The observation that IbpA fusion proteins constructed with monomeric fluorescent protein derivatives ( i . e . , IbpA-mVenus , IbpA-mCherry , IbpA-monomeric cerulean [mCer] , and IbpA-monomeric superfolder green fluorescent protein [msfGFP] ) gave rise to lower fractions of punctate cellular fluorescence compared to reporters constructed with nonmonomeric forms ( i . e . , IbpA-YFP and IbpA-Venus ) indeed confirms that the self-aggregating tendency of these fusion proteins can lead to an overestimation of the natural PAs present inside the cell ( Fig 1A–1C ) . In the same vein , inclusion body binding protein B ( IbpB ) —the other small heat shock protein encoded in the ibp operon that associates with PAs through its interaction with IbpA [32]—displayed similar variability in localization when fused to different fluorescent proteins ( S1D–S1F Fig ) , albeit to a lesser degree . Interestingly , the IbpA-msfGFP reporter ( i . e . , the fusion equipped with the most monomeric fluorescent label; [35 , 36] ) displayed a strictly diffuse cytosolic fluorescence in exponential phase ( Fig 1A–1C ) and only exhibited a punctate localization in some cells ( 12 . 6% ) of a stationary phase population ( S1A–S1C Fig ) . Furthermore , this reporter also displayed wild-type expression levels of the ibpA promoter ( PibpA; which is highly sensitive toward perturbations of protein homeostasis [37–39] ) , while PibpA activity was markedly increased in the biased MG1655 IbpA-YFP reporter ( Fig 1D ) . Nevertheless , exposure of the MG1655 IbpA-msfGFP reporter to a sublethal heat shock led to significantly increased ibpA expression ( Fig 1E ) and the emergence of multiple foci ( mainly localized in the polar and mid-cell regions ) , confirming its ability to sense and report protein aggregation ( Fig 1F–1H and S2A and S2B Fig ) . Moreover , sublethal treatment of this reporter strain with streptomycin—conditions known to induce mistranslation and subsequent protein misfolding [8 , 10]—induced a similar response in which emerging IbpA-msfGFP foci colocalized with the streptomycin-induced inclusion bodies ( visible in the phase contrast images; S2C–S2E Fig ) . Inclusion body formation driven by the production of recombinant protein also instigated a similar up-regulation and colocalization phenotype ( S2F–S2I Fig ) , further forwarding IbpA-msfGFP as a reliable reporter for monitoring native intracellular PA dynamics . After validating IbpA-msfGFP as a truthful marker for the presence of natively occurring PAs , we set forward to further probe their occurrence and physiological impact . To this end , we exposed exponentially growing MG1655 ibpA-msfgfp cells to a range of sublethal heat shocks ( temperatures higher than 50 °C led to the inactivation of a substantial fraction of cells and were subsequently not considered in this analysis ) and monitored their response and subsequent outgrowth by time-lapse fluorescence microscopy ( TLFM ) . While both IbpA up-regulation and IbpA-msfGFP foci formation originally increased with the severity of the heat shock , this coordinated behavior displayed remarkable alterations at higher sublethal temperatures . The number of visible PA foci per cell leveled off at temperatures higher than 45 °C , and IbpA expression reached a maximum at 47–48 °C ( Fig 2A–2C ) . Given the observed foci account for most of the cellular fluorescence , this indicates that in between this temperature range ( 45–48 °C ) , larger but not more PAs were formed . After exposures to higher temperatures ( 49–50 °C ) , both the expression level and the number of PAs declined , and a significant delay in cellular growth resumption could also be observed ( Fig 2D ) . Within populations exposed to the same temperature , a large variability in the extent of protein aggregation ( both in the number of visible PAs as in the amount of IbpA-msfGFP fluorescence ) could be observed with cells containing 0 to 5 distinct IbpA-msfGFP foci at cell division after exposure to the sublethal heat treatment ( Fig 2C ) . Populations exposed to higher sublethal temperatures ( 48–49 °C ) displayed a remarkable bifurcation phenotype ( Fig 2E and 2F ) . Although a significant fraction of cells behaved similarly as recorded for lower sublethal temperatures , displaying clear IbpA up-regulation ( which persisted during initial growth resumption ) and foci formation , others exhibited neither of these characteristics ( Fig 2F ) . This fraction of cells did display belated IbpA up-regulation but not to the same extent as their PA-forming counterparts . Moreover , an initial “patchy” IbpA-msfGFP localization pattern could be observed in these cells , which appeared to be resolved concomitantly with IbpA up-regulation as cells resumed growth and division ( Fig 2F ) . Interestingly , whereas the PA-forming fraction of cells readily resumed growth , this other non-PA-forming fraction displayed significantly longer resuscitation times ( Fig 2E and 2F ) . Whereas only a limited number of cells ( 10 . 5% ) displayed this behavior after exposure to 48 °C , this fraction , as well as their average resuscitation time , became significantly larger ( 22 . 5% ) in populations exposed to 49 °C ( Fig 2D and 2E ) . At higher temperatures ( 50 °C ) , this bifurcation disappeared , and the behavior of ( surviving ) cells shifted completely toward attenuated IbpA up-regulation without PA formation combined with longer resuscitation times ( Fig 2D ) . Although the mechanisms underlying the observed bifurcation and phenotypic shift remain elusive , our observations highlight an apparent plasticity in the formation of PAs . This suggests that PA formation might be more than an inevitable artefact of exposures to elevated temperatures , in which case PAs would be observed in surviving cells after exposure to all temperatures above a given threshold . To further examine the impact of PAs on cellular physiology , we exposed cells to a sublethal heat treatment leading to maximal PA formation while minimally compromising physiology and subsequent proliferation ( 47 °C , 15 min ) . We quantitatively characterized the growth of these cells over multiple generations in detail with TLFM . As the cells resumed growth , PAs typically remained intact , became localized in the cell poles , and segregated asymmetrically while cells grew out into microcolonies ( Fig 3A–3D , S1 and S2 Movies ) . Although the existence of so-called cellular aging could be detected in these microcolonies ( 9 . 02% decrease in cellular growth rate of cells inheriting the oldest cell pole compared to the rest of the population; S2J Fig ) , stochastic partitioning of PAs during the first generation post heat treatment led to only 6 out of a total of 40 observed oldest cells containing a PA , suggesting both phenomena might not be associated with each other . Subsequent examination of aging in PA-free cells indeed confirmed this phenomenon to occur independently of protein aggregation ( 9 . 36% decrease in cellular growth rate; S2J Fig ) . Moreover , a permutation test revealed that the limited average fitness defect of PA-bearing cells does not differ from that of PA-free cells with a similar age structure ( p-value = 0 . 56; Fig 3E–3G ) , indicating PAs themselves impart no significant growth defect on their individual host cells . This finding was further strengthened by the lack of correlation between cellular growth rate and average cellular GFP concentration for PA-bearing cells ( Fig 3E ) . To assure that the phenomenon of asymmetrically inherited PAs indeed represents native behavior and is not a trivial consequence of the fluorescent labeling of IbpA with a monomeric fluorescent protein , we equipped wild-type MG1655 cells ( thus containing an unlabeled native copy of ibpA ) with a vector ( pBAD33-ibpA-msfgfp ) in which expression of the IbpA-msfGFP fusion protein was placed under an arabinose-inducible promoter . These cells were grown in repressing conditions ( 0 . 2% glucose ) to exponential phase , exposed to a sublethal heat treatment ( 47 °C , 15 min ) , and subsequently monitored on arabinose-containing ( 0 . 15% ) agarose pads by TLFM ( S3 Fig ) . As cells grew under these inducing conditions , a gradual increase in cellular fluorescence could be observed , followed by the appearance of one or multiple distinct foci that remained present during the rest of recording ( S3A Fig ) . While a similar increase in cellular fluorescence could also be observed in control cells not subjected to a sublethal heat treatment , this increase was not accompanied by the manifestation and stable inheritance of distinct fluorescent foci ( S3B Fig ) . Together , these findings indicate that the presence of fluorescently labeled IbpA is not a prerequisite for the formation of asymmetrically segregating PAs ( which become apparent upon induction of the IbpA-msfGFP fusion protein ) , demonstrating that the latter is indeed a natively occurring phenomenon after exposure to a sublethal heat treatment . Remarkably , higher induction levels of the fusion protein led to the irregular formation ( i . e . , continuously , not only initially when cellular fluorescence increased ) of multiple fluorescent foci in both heat-treated and control cells ( S3C Fig ) . This suggests that increased expression of IbpA itself , in line with previous observations describing the emergence of PAs upon overproduction of IbpA [40 , 41] , might contribute to the formation of these structures in proteotoxic stress conditions ( as these are known to induce ibpA expression ) . We noticed that the asymmetric segregation of PAs was accompanied by an IbpA-msfGFP concentration gradient stemming from the PA-bearing cell ( s ) and progressively diminishing in its closest relatives ( Fig 3A–3D ) . To examine whether this concentration gradient could be the consequence of increased ibpA expression in PA-bearing cells , we performed an identical experiment with MG1655 ibpA-msfgfp pSG1 cells , which allow the concurrent monitoring of IbpA promoter activity , concentration , and localization ( S4A Fig ) . Although these cells behaved similarly after exposure to a sublethal heat shock in terms of PA formation , localization , and asymmetric segregation ( S4B Fig ) , no increased mCherry signal could be detected in PA-bearing cells ( as compared to that of their PA-free counterparts; S4C Fig ) . This indicates that the observed concentration gradient is likely not the consequence of a transcriptional ibpA response to the presence of these PAs but presumably finds it origin in the gradual disaggregation of the existing PA . In agreement , a small fraction of intracellular PAs was even observed to disappear after a certain amount of time , presumably because of their complete disaggregation . Intriguingly , the above described phenomenon was not only instigated by a sublethal heat treatment . Exposure to other sublethal environmental stressors impacting proteostasis , such as streptomycin or hydrogen peroxide , gave rise to similar PA-associated phenotypic behavior ( S5 Fig ) . As such , the formation and subsequent asymmetric segregation of these structures appears to be a conserved phenomenon throughout a wide range of sublethal environmental conditions affecting cellular proteostasis . Since PAs thus appear as asymmetrically segregating remnants of a previously encountered proteotoxic stress , we wondered to what extent the presence of these structures ( and the resulting concentration gradient ) could influence survival upon a subsequent stress exposure . To this end , we challenged a total of 38 lineages stemming from sublethally heat-shocked cells ( 47 °C , 15 min ) to a subsequent semilethal heat shock ( 51 °C , 7 min; Fig 4A ) , an assay in which we have previously described cellular survival to behave as a stochastic trait free of predispositions [42] . The heterogeneity in PA formation gave rise to microcolonies harboring a variable number of PAs , with an average of 1 . 41 PA foci per microcolony at the moment of the second heat treatment . Average cellular survival within these microcolonies ( i . e . , 55 . 6% ) did not significantly differ from that within microcolonies stemming from unstressed cells not exposed to a prior sublethal heat shock ( Fig 4A–4C ) , indicating the previously mounted heat shock response had faded to levels unable to affect the overall survival frequency [43 , 44] . Interestingly , however , comparison of the average survival of PA-free cells to that of PA-bearing cells clearly revealed the latter to display a significantly higher survival frequency ( Student t test , p-value = 1 . 33 × 10−3; Fig 4D ) . Moreover , by binning the frequency of survival by average cellular IbpA-msfGFP fluorescence before the second heat shock , it also became obvious that cells bearing the highest initial IbpA concentrations ( i . e . typically containing a larger PA ) tend to display significantly higher survival probabilities ( from an approximately 1/2 to 4/5 chance to survive; Fig 4E ) . Even within sister-cell pairs consisting of a PA-free and PA-bearing cell , a similar differentiation could already be observed ( survival frequencies of 58 . 3% and 74 . 4% for PA-free and PA-bearing cells , respectively ) , further underscoring the close link of this increased robustness to PA inheritance . Another potential explanation for these observations could stem from a confounding factor underlying both increased robustness and PA inheritance . Given the asymmetric nature of the latter phenomenon , increased cell age could be such a biasing property . Although PA-bearing cells were , on average , indeed older than the remainder of the population ( average cell age of 1 . 80 versus 3 . 44 generations for PA-free and PA-bearing cells , respectively ) , no overall tendency for older cells to display increased survival probabilities could be detected ( S6A Fig ) . Moreover , by exploiting the stochastic partitioning of PAs during the first generation post sublethal heat treatment , which gave rise to PA-bearing and PA-free old pole cells , we could directly disentangle the effect of both phenomena on survival . In line with our previous findings in microcolony-level semilethal survival assays [42] , we found no evidence for any age-related bias in survival frequencies of PA-free cells ( S6B Fig ) , further strengthening the role of PAs as deterministic factors that increase robustness of individual cells in an otherwise stochastic assay . In line with our previous observations with more severe heat shocks ( Fig 2 ) , no new , distinct PAs emerged in cells surviving the semilethal heat shock , although in cells already containing a PA , this structure itself mostly remained present ( Fig 4A ) . Again , an initial “patchy” localization pattern and a belated up-regulation of IbpA together with its apparent dissolution could be observed ( Fig 4A ) . A similar observation could be made in surviving cells of microcolonies consisting of unstressed cells not exposed to a prior heat shock ( Fig 4B ) . Taken together , these results indicate that heat-induced PAs are able to increase the survival probability of individual cells upon exposure to a subsequent heat shock . As such , PAs appear to function as asymmetrically segregating , epigenetic memory elements ( i . e . , reminiscing previous torments ) that impose phenotypic heterogeneity in isogenic microcolonies by improving the robustness of the PA-inheriting siblings . Moreover , detailed lineage tracing after the semilethal heat shock revealed that the increase in robustness might not only manifest itself in terms of higher survival probabilities but also in terms of decreased resuscitation times . Surviving PA-bearing cells appeared to resume growth and division faster than other surviving cells , leading to an enrichment of their progeny in the emerging population ( Fig 4F ) . To independently confirm that intracellular PAs constitute an asymmetrically transmissible form of epigenetic memory that drives heat resistance , we set forward to develop a novel synthetic PA model system that alleviates the need for an external stress factor to induce PA formation . More specifically , we started with a fragment of the lambda prophage repressor protein ( cI ) , a protein known for its potential to misfold and aggregate by the introduction of a limited number of mutations [45] . To ensure the inertness of the repressor within the E . coli cytoplasm , its N-terminal DNA-binding domain was first removed [46] . The obtained fragment , dubbed cI78WT ( as it does not include the first 77 amino acids of the full-length repressor protein ) , was fused N-terminally to the mCerulean3 fluorescent protein ( yielding mCer-cI78WT ) and placed under the control of an isopropyl β-D-1-thiogalactopyranoside ( IPTG ) -inducible promoter on a pTrc99A expression vector ( yielding pTrc99A-mCer-cI78WT ) . This construct was subsequently transformed into an MG1655 ΔlacY strain to allow the titratable expression of the fusion protein [47] . Upon induction , this fluorescent fusion protein displayed a diffuse cytosolic fluorescence , indicating the protein remained completely soluble ( Fig 5A ) . To isolate constitutively aggregating mCer-cI78 variants , we screened for mutations in cI78 , introduced by error-prone PCR , that resulted in an alteration of this diffuse fluorescent localization pattern . As such , we were able to identify 1 mutant ( cI78EP8; harboring 1 synonymous and 3 nonsynonymous mutations ) displaying strict punctate and polarly located fluorescence ( Fig 5B–5F ) , characteristic of disrupted folding and subsequent aggregation . To further validate this isolated mutant as an aggregating protein , we examined whether it displayed other typical properties of prokaryotic PAs . First , we investigated whether its polar localization was nucleoid-enforced by equipping our MG1655 ΔlacY pTrc99A-mCer-cI78EP8 strain with a hupA-Venus fusion , allowing the fluorescent visualization of the nucleoid [48] , as well as by examining the behavior of mCer-cI78EP8 in the absence of a nucleoid ( by employing a ΔrecA derivative of MG1655 ΔlacY pTrc99A-mCer-cI78EP8 , which occasionally gives rise to anucleate cells [49] ) . In agreement with characteristic PA behavior , mCer-cI78EP8 foci localized to nucleoid-free regions of the cell and were retained in the cell poles by nucleoid occlusion ( evidenced by foci freely roaming throughout the entire cytoplasm in anucleate cells; S7A–S7C Fig ) . Second , we equipped mCer-cI78WT- and mCer-cI78EP8-expressing strains with pSG1 and confirmed that ibpA expression under inducing conditions indeed increased in the latter ( S7D Fig ) . In addition , we equipped MG1655 ibpA-msfgfp cells with a pTrc99A-mCherry-cI78EP8 construct ( in which the mCerulean3 fluorescent protein was replaced by mCherry because of its spectral incompatibility with msfGFP ) and examined whether IbpA was able to recognize the misfolding and aggregating cI78EP8 variant . From this , a clear colocalization pattern between the IbpA-msfGFP and mCherry-cI78EP8 foci could be observed , indicating the latter indeed represents a misfolded and aggregating protein species ( recognized by IbpA; S7E and S7F Fig ) . We subsequently investigated whether the production of the mCer-cI78EP8 aggregating protein , as is the case with other misfolding protein species [10 , 50] , affected cellular fitness . Whereas no difference in fitness could be detected in uninduced conditions ( without the addition of IPTG ) , cells actively producing the mCer-cI78EP8 variant grew significantly slower than the cells producing the mCer-cI78WT variant , with the fitness disadvantage attributable to PAs increasing with the amount of insoluble protein produced ( S7G Fig ) . Moreover , in fully induced conditions ( 1 mM IPTG ) , a clear negative correlation between microcolony fluorescence ( indicative of the amount of insoluble protein present ) and microcolony growth rate could be observed for cells expressing the aggregating mCer-cI78EP8 variant ( S7H Fig ) . A similar negative correlation could not be detected in cells expressing the soluble cI78WT variant ( S7I Fig ) , indicating the fitness defect was indeed attributable to misfolding and aggregation of the mCer-cI78EP8 protein . Please note that the observed fitness defect under these conditions , in which the misfolding protein is actively being produced , is fundamentally different from the previously reported absence of a fitness defect for host cells of asymmetrically segregating PAs after exposure to a sublethal heat treatment ( Fig 3G ) . Whereas the former represents cells that are continuously challenged by misfolding proteins , environmental proteotoxic stress conditions have been relieved in the latter , and the PAs only remain present as remnants of a previously encountered sublethal proteotoxic insult . Using this newly developed and validated model system , we subsequently examined whether synthetic PAs ( consisting of an aggregating , inert E . coli protein ) indeed confer increased heat resistance . In a first step , MG1655 ΔlacY pTrc99A-mCer-cI78EP8 cells were grown to exponential phase in AB medium with 0 . 2% glycerol in the presence of 1 mM IPTG ( to induce the production of mCer-cI78EP8 ) , harvested , and washed into AB medium with 0 . 2% glucose ( impeding further induction of mCer-cI78EP8 production ) , after which their growth was monitored by TLFM . Similarly as in MG1655 ibpA-msfgfp cells after a sublethal heat shock , the synthetic PAs not only segregated asymmetrically throughout the emerging microcolonies but were also accompanied by a mCer-cI78EP8 concentration gradient originating from the PA-bearing cells ( Fig 6A and 6B and S3 Movie ) . In this case , given that production of the fluorescent protein has been halted , it is clear that the emerging concentration gradient is a consequence of protein disaggregation rather than a specific response to the presence of PAs . Although PA-bearing cells , on average , displayed a significant fitness defect ( 16 . 3%; Fig 6A–6D ) , this effect could not easily be discerned from cellular aging , as both phenomena were similar in magnitude ( 16 . 1% decrease in cellular growth rate of cells inheriting the oldest cell poles compared to the rest of the population; S2K Fig ) , and a larger fraction of old pole cells also contained a PA ( 13 out of 24 observed old pole cells ) . Nevertheless , a direct comparison of the growth rates of PA-bearing old pole cells with their PA-free counterparts yielded no significant difference ( Student t test , p-value = 8 . 25 × 10−2 ) , indicating that the contribution of PAs to the observed cellular aging was again negligible . Upon exposure of these growing microcolonies to a heat shock killing approximately half of the population ( 52 °C , 7 min ) , the synthetic PAs were found to confer a similar protective effect as seen with the IbpA-msfGFP-labeled stress-induced PAs ( Fig 6E–6I ) . While average cellular survival did not significantly differ between PA-harboring ( MG1655 ΔlacY pTrc99A-mCer-cI78EP8 ) and PA-free ( MG1655 ΔlacY pTrc99A-mCer-cI78WT ) microcolonies ( Fig 6G ) , PA-bearing cells displayed a significantly higher survival frequency within the former group ( Fig 6H and 6I ) . Remarkably , this synthetic PA-mediated protection already manifested itself after one division ( Fig 7A and 7B ) , underscoring the apparent speed with which this PA-mediated differentiation occurs . Upon further analysis , this experimental setup clearly revealed the dual effect of PA presence on stress management , as PA-bearing cells not only displayed a higher survival frequency , but surviving PA-bearing cells also displayed a significantly reduced resuscitation time as compared to their surviving PA-free counterparts ( Fig 7C ) . Synthetic PAs appeared as non-diffraction-limited fluorescent spots , which allowed the accurate determination of their size ( instead of relying on indirect , and potentially biased , fluorescence intensity measurements , see Materials and methods; Fig 7D and 7E and S8 Fig ) . These measurements revealed that aggregate size did not appear to correlate with cellular survival , indicating that the amount of aggregated protein species is , at least within the sampled range , irrelevant for the observed phenotype ( Fig 7F ) . Moreover , these measurements allowed us to estimate the copy number of the mCer-cI78EP8 molecules in each aggregate . Based on its amino acid sequence , the fusion protein has an approximate molecular weight of 45 kDa and is estimated to occupy a volume of 5 . 71 × 10−8 μm3 [51] . Assuming spherical aggregate shape and a pure mCer-cI78EP8 composition , individual synthetic aggregates are subsequently roughly estimated to be composed of around 4 × 105 protein molecules . Cells actively producing misfolded and aggregating protein species became more susceptible to heat stress than those producing similar amounts of soluble protein ( S9 Fig ) . This was illustrated by the decreased survival of MG1655 ΔlacY pTrc99A-mCer-cI78EP8 populations as compared to that of MG1655 ΔlacY pTrc99A-mCer-cI78WT populations after exposure to a semilethal heat treatment in inducing conditions ( S9 Fig ) . The protective effect of PAs thus appears limited to cases in which the original conditions giving rise to their emergence have been relieved . Together with the previously described negative impact of PA production on cellular fitness ( S7G–S7I Fig ) , this sensitization toward proteotoxic stress during PA production likely impedes PA formation from being an efficient way to increase average population-level robustness in benign conditions ( without a previous sublethal proteotoxic exposure ) . As such , PAs likely function as true epigenetic memory elements , in which cells must have encountered an environmental condition licensing their emergence . In order to further substantiate and characterize the observed PA-mediated robustness in a more versatile fashion , we used an alternative approach to determine survival frequencies of large numbers of PA-bearing and PA-free cells . In essence , after transient induction of synthetic ( mCer-cI78EP8 ) PA production , the population continued growth and concomitant asymmetric PA segregation in liquid culture , after which the resulting PA-free and PA-bearing cells were stress-challenged and only then monitored on a single-cell level by TLFM ( Fig 8A ) . By postponing the mounting and incubation of cells under the microscope , this setup allowed us to easily ( i ) increase the experimental throughput , ( ii ) vary the number of generations/segregations between PA production and stress challenge , ( iii ) expose the cells to other stresses than heat , and ( iv ) expose the cells more homogeneously to a stress as well . First , we varied the intensity of the heat challenge and exposed clonal populations of PA-free and PA-bearing siblings to heat shocks ranging from 51 to 53 . 5 °C . As such , we found that PA-bearing cells consistently displayed higher survival frequencies than PA-free cells ( Fig 8A ) , confirming the protective effect of PAs and indicating a wide protective range . Moreover , the relative increase in survival that could be attributed to PA presence generally increased with increasing heat intensity . For example , after an exposure to 53 . 5 °C ( for 15 min ) , PA-bearing cells displayed an almost 6-fold higher survival frequency than PA-free cells ( 11 . 6% versus 2 . 0% ) . Even after the mildest exposure ( 51 °C for 15 min ) , PA-bearing survivors displayed significantly shorter resuscitation times than their PA-free counterparts ( Fig 8B ) , again underscoring the dual protective effect of PAs . Subsequently , we used the above-described approach to investigate PA-mediated robustness over longer timescales of asymmetric PA inheritance . While we previously employed timescales corresponding to around 4 generations or fewer ( Figs 4 , 6 and 7 ) , we could now allow PA-containing populations to grow for longer periods of time before applying a heat challenge ( 52 °C for 15 min ) and quantify the resulting survival of clonal PA-free and PA-bearing siblings . Although the ongoing asymmetric segregation of PAs during prolonged growth inevitably leads to a progressively decreasing fraction of PA-bearing cells , a sufficient number could still be observed to irrefutably demonstrate that the protective effect of PAs persists over longer timescales ( Fig 8C ) . Even after 200 min of growth ( corresponding to approximately 7–8 generations of growth ) , the presence of these structures was associated with a similar increase in survival frequency ( Fig 8C ) . Finally , this setup also enabled us to investigate whether the PA-mediated increase in cellular robustness extended to proteotoxic stresses other than heat . To this end , we exposed clonal populations of PA-free and PA-bearing siblings to semilethal concentrations of either hydrogen peroxide ( Fig 8D ) or the ribosome-targeting antibiotic streptomycin ( Fig 8E ) . In line with the previous heat shock experiments , we found that cells harboring a PA consistently displayed higher survival frequencies for both challenges ( Fig 8D and 8E ) , suggesting that PAs can confer robustness to a wide variety of proteotoxic stresses . In a subsequent step , we sought to provide some insight into the molecular mechanisms underlying the observed increase in robustness of PA-bearing cells . Given the apparent absence of a direct link between the identity of the aggregated protein species and the memory encoded by them ( as is the case for other examples of protein-based inheritance [52 , 53] ) , we chose to characterize and examine the potential role of protein quality control components , as previous work has demonstrated their physical association with ( disaggregating ) PAs [14 , 54–56] . In a first step , we constructed deletion strains of ibpA , ibpB , and the entire ibpAB operon and equipped these with our synthetic PA system . We subsequently probed for the existence of PA-mediated robustness in these strains by using a similar setup as in Fig 6 , in which growing and PA-containing microcolonies were exposed to a semilethal heat shock ( 52 °C , 7 min ) . None of the deletions , however , appeared to affect the increase in heat resistance of PA-bearing cells ( S10 Fig ) , suggesting these small heat shock proteins do not play a role in establishing PA-mediated asymmetry . In contrast to the small heat shock proteins , deletion of many other protein quality components is known to severely compromise heat survival [57 , 58] . This is further exemplified by the lack of detectable survivors in a ΔclpB strain in our survival assay . ( S10 Fig; ClpB is a heat shock protein 100 [Hsp100] AAA+ chaperone involved in protein disaggregation [59] ) . This reduction in survival frequency impedes direct interpretations concerning the potential role of ClpB ( and other protein quality control components ) in establishing PA-mediated robustness . To overcome this confounding issue and obtain some insight into the potential role of the proteostasis network , we resorted to the use of fluorescent fusion proteins . We constructed both transcriptional and translational msfGFP fusions to a variety of chaperones and proteases and equipped these strains with the mCherry version of our synthetic PA system ( to ensure spectral compatibility with msfGFP ) . As we expected nonfunctional fusions to decrease heat survival frequency [57 , 58] , we first validated the fusion proteins by exposing the corresponding strains to a heat treatment and comparing their inactivation levels to those of a control without any chaperone or protease fusions ( S11A and S11B Fig ) . This analysis revealed that only the transcriptional PclpP-msfgfp fusion led to a significantly increased inactivation ( S11A and S11B Fig ) and thus likely exerts a polar effect that perturbs wild-type cellular physiology . We further validated the transcriptional fusions by exposing these strains to a sublethal heat shock and verifying the expected increase in promoter activity ( [60]; S11C Fig ) . Whereas most transcriptional fusions indeed displayed a significant up-regulation directly after heat treatment ( S11C Fig ) , the transcriptional PclpP-msfgfp fusion did not , further supporting the lack of functionality of this fusion . A similar increase in concentration after exposure to a sublethal heat shock could often not be detected on a protein level for these chaperones and proteases ( S11D Fig ) , suggesting the potential existence of posttranscriptional regulation , slow protein folding and maturation , and/or high rates of protein turnover . In line with the previously noted absence of a heat shock response to the presence of asymmetrically segregating PAs ( S4 Fig ) , we found that none of the tested protein quality control components displayed a transcriptional up-regulation in the presence of PAs ( S12A and S12B Fig ) . Their expression level ( indirectly measured through the average cellular GFP concentration ) often even appeared lower than that of PA-free cells , although this likely is a consequence of the impermeability of ( synthetic ) PAs to other cytosolic components ( S12C Fig ) . The presence of PAs therefore appears to lead to an extra addition of cell volume containing no fluorescence , in turn leading to a lower apparent concentration . On the protein level , however , specific protein quality control components ( the Hsp70 chaperone DnaK , the Hsp40 chaperone DnaJ , the Hsp100 AAA+ chaperone ClpB , and the serine protease ClpP ) displayed an increased concentration in PA-bearing cells ( Fig 9A ) . TLFM revealed that the increased concentration was a direct consequence of a remarkable colocalization of these proteins with asymmetrically segregating PAs that persisted over multiple generations ( Fig 9B ) . This observation appears to be in line with the similarly ongoing PA-disaggregation observed earlier ( Figs 3A , 3C , 4A and 6A ) . Other protein quality components ( the AAA+ protease subunit ClpX , the serine protease Lon , the AAA+ protease subunit HslU , and the Hsp90 chaperone HtpG ) did not display any colocalization or an increased concentration ( Fig 9A and S13 Fig ) . In fact , in similar fashion as for the transcriptional reporters and presumably because of the same PA-mediated addition of nonfluorescent cell volume , a lower apparent concentration was often measured in PA-bearing cells for these chaperones and proteases . Asymmetric segregation of PAs thus appears to drive the specific enrichment of protein quality control components in their host cells . This enrichment might in turn be responsible for the increased robustness of PA-bearing cells toward proteotoxic stresses .
In contrast to the governing view on prokaryotic PAs as inevitably debilitating structures [10–12 , 14] , our findings in the E . coli model system reveal the potential of PAs to improve cellular robustness , both in terms of increased survival frequencies and decreased resuscitation times , independent of their origin . Moreover , because of their asymmetric segregation and limited disaggregation , these structures resist dilution during cytoplasmic inheritance and are therefore able to persist for many generations as a functional physical relic of an ancestral insult . As such , these observations reveal the existence of stress-induced , long-term epigenetic memory in prokaryotes . More specifically , such cellular memory appears to be installed through ( sublethal ) proteotoxic exposures that lead to initial up-regulation of the heat shock response and colocalization of specific chaperones and proteases with the emerging PAs . Growth after relief of the proteotoxic exposure coincides with asymmetric segregation of this PA and its associated chaperones and proteases , while additional heat shock proteins do not seem to be induced upon the mere inheritance of this structure . Whereas the selective redistribution of quality control components to polar PAs in E . coli during and directly after proteotoxic stress has been observed before [8 , 10 , 13 , 14] , our data indicate that this association persists over multiple generations and leads to their specific enrichment in PA-bearing cells . In fact , the slowly ongoing PA-disaggregation during outgrowth might be a consequence of a prolonged association of these components with the PA . Importantly , previous studies have reported that the association of several chaperones and proteases with PAs is dynamic [14] , indicating that these proteins are not necessarily trapped within these intracellular structures and might become available as additional protein folding aids that increase cellular robustness in times of stress . Interestingly , this putative mechanism might even extrapolate to other ( eukaryotic ) microorganisms , since S . cerevisiae mother cells harboring an age-associated PA were previously reported to clear heat-induced aggregates faster than daughter cells without such an aggregate [61] . However , further research is required to establish a direct causal relation between PA-associated chaperones and proteases and the observed cellular robustness . While the use of individual deletion mutants does not seem a good option because of their severely compromised heat survival [57 , 58] ( S10 Fig ) , obtaining tunable amounts of different protein quality control components and subsequently assessing to what extent this gradually affects PA-mediated robustness might be a more fruitful strategy to establish causality . Furthermore , it will also be interesting to see whether or not PA-mediated protection is consequently limited to those stresses that require protein folding aids to be alleviated . In essence , epigenetic memory refers to hysteretic behavior in which the physiological state of a cell is determined by its ( or its ancestor’s ) past experience rather than being dictated by genomic mutations or the current environment it resides in [62 , 63] . Nonmutational mechanisms allowing the long-term inheritance of a previously acquired physiological state are typically based on the passage of specific protein activities or interactions from one generation to the next . As such , E . coli cells already expressing the LacY permease in intermediate inducer concentrations will continue doing so for subsequent generations because their ability to import the inducer will lead to continued production of the permease [64 , 65] . Similarly , cells inheriting a self-propagating prion seed transmit this infectious protein conformation to their progeny , in which this process will repeat itself [66–69] . In contrast to these mechanisms , intracellular PAs essentially lack a self-proliferative effect ( i . e . , it is the original PA that is passed on from one cell to another ) but nevertheless remain inheritable for many generations because the asymmetric segregation and limited disaggregation prevent their dilution . This relates to the mnemons concept posed by [52] , in which the conditional self-aggregation of a specific protein ( the Whi3 mRNA binding protein of S . cerevisiae ) compromises its functionality and proper segregation to daughter cells . However , since the exact PA origin in our experiments appears to be irrelevant for its phenotypic consequences , prokaryotic PA-mediated memory seems to comprise a completely novel type of protein-based inheritance . Our single cell–level observations also revealed other intriguing aspects of the process and impact of in vivo protein aggregation . PAs , for example , seem not to emerge by default after exposures to a proteotoxic stress . Instead , their formation appears limited to less severe , sublethal stressful encounters . Although we cannot currently provide a conclusive answer , we hypothesize that this observation could reflect different , nonmutually exclusive , cellular strategies . One possibility is that the dosage of misfolded proteins determines the choice between aggregation and refolding/degradation , given that a largely aggregated proteome would trivially lead to defects in cellular function . Another possibility is that the formation of these large intracellular structures requires a certain factor or multiple factors ( in line with the observation in fission yeast that Hsp16 mediates PA fusion [70] ) , but this factor is allocated to repair or prevention of aggregation of other critical cellular components under more severe stress conditions . Alternatively , PA formation could pose too much of a risk under severe heat stress , as aggregation could potentially lead to coaggregation and trapping of other critical cellular components [71] , themselves destabilized and misfolded under these more severe stress conditions . Aside from their remarkable pattern of emergence , the slow and limited poststress disaggregation of PAs over multiple generations also represents cellular behavior that previously remained undetected using population-level determinations of aggregated protein fractions [8 , 54 , 72] . Although many of the molecular players and steps involved in the disaggregation process have already been identified and studied in vitro [54 , 56 , 73] , our observations underscore that the in vivo implementation and kinetics of this process require further study . The resulting persistence and asymmetric inheritance of PAs over multiple generations gives rise to a progressively decreasing fraction of memory cells as population numbers increase . This heterogeneity is reminiscent of a bet-hedging strategy typically employed by microbial populations to mitigate costs or trade-offs associated with the installment of given phenotypes . However , PA-mediated fitness defects could not be detected in terms of cellular growth rate , in agreement with recent findings in fission yeast [74] , suggesting such trade-offs either remain undetected in our setup ( and are yet to be discovered ) or are simply not present . In this light , the formal distinction between protein aggregation ( believed to be deleterious ) and quinary assembly formation ( believed to be adaptive ) [22 , 23] becomes more vague as well . Importantly , microorganisms can encounter many forms of proteotoxic stress throughout their habitats , and our data indicate that sublethal exposures to streptomycin and hydrogen peroxide not only give rise to similar emergence and inheritance of PAs but that the PA-mediated robustness extends toward these other proteotoxic stresses as well . These stressors are of specific importance given their roles in curbing and fighting microbial infections in humans . Not only do many antibiotics specifically target protein homeostasis [8]; the mechanism behind the activity of bactericidal antibiotics has also been linked to the production of reactive oxygen species ( ROS ) [75 , 76] . Moreover , the generation of ROS has been implicated in microbial killing by phagocytes [77 , 78] . Consequently , our findings might have broader implications in the context of understanding how bacteria cope with and evade antimicrobial therapeutics and host defense mechanisms . In conclusion , intracellular prokaryotic PAs appear to be more than inevitable and detrimental cellular garbage bins , as these structures encode epigenetic long-term memory of previous proteotoxic torments that becomes vertically transmitted for a number of generations while conferring an increased robustness to its carrier cell . As such , the formation and conservation of PAs in prokaryotes resembles an adaptive process that aids cellular survival and adaptation in fluctuating environments . Moreover , this paradigm further suggests that other types of cytoplasmically inheritable damaged biomolecules could likewise serve as functional “scar tissue” that improves cellular robustness over multiple generations .
Bacterial strains , plasmids , and primers used in this study are listed in S2 , S3 and S4 Tables , respectively . For culturing of bacteria , mostly lysogeny broth ( LB ) medium was used as either a broth or solid medium after the addition of 2% agarose ( for agarose pads intended for microscopy ) . In indicated cases , LB medium was replaced by AB medium , supplemented with 0 . 2% of a carbon source ( glucose or glycerol ) , 0 . 2% casamino acids , 10 μg/ml thiamine , and 25 μg/ml uracil . Stationary-phase cultures were obtained by growing E . coli overnight for approximately 15 h in LB broth at 37 °C under well-aerated conditions ( 200 rpm on an orbital shaker ) . Exponential phase cultures were in turn prepared by diluting stationary phase cultures 1/100 in fresh prewarmed broth and allowing further incubation at 37 °C until an OD600 of 0 . 2–0 . 6 was reached . When appropriate , the following chemicals ( Applichem , Darmstadt , Germany , and Sigma-Aldrich ) were added to the medium at the indicated final concentrations: kanamycin ( 50 μg/ml ) , ampicillin ( 100 μg/ml ) , chloramphenicol ( 30 μg/ml ) , 4′ , 6-diamidino-2-phenylindole ( DAPI; 1 μg/ml ) , IPTG ( 1–1 , 000 μM ) , glucose ( 0 . 2% ) , L-arabinose ( 0 . 1%–0 . 2% ) , and glycerol ( 0 . 2% ) . The construction of mutants in E . coli is greatly facilitated by lambda Red-mediated recombination . The protein products of the red genes ( Gam , Exo , and Beta ) enable the highly efficient recombination of PCR products ( containing a selectable antibiotic cassette , optionally flanked by other genes of interest such as those encoding fluorescent proteins ) flanked by short ( 50 bp ) nucleotide sequences , homologous to the target sequence [79] . These antibiotic cassettes are usually flanked by frt sites , which allow the excision of the cassette by site-specific recombination between two frt sites . All constructed mutants were initially confirmed by PCR with primer pairs attaching outside of the region where homologous recombination occurred . Correct deletion or integration of PCR products was further verified by sequencing ( Macrogen , the Netherlands ) . The different C-terminal translational fusions of IbpA and IbpB were constructed by recombineering PCR fragments obtained from plasmid pDHL1029 and its derivatives pDHL-venus , pDHL-mVenus , pDHL-mCherry , and pDHL-mCer ( the original pDHL1029 plasmid contains a msfgfp-frt-nptI-frt cassette [36]; its derivatives were constructed during this work and contain sequences encoding other fluorescent proteins: Venus , mVenus [monomerized Venus by introducing V206K mutation] , mCherry , mCerulean3 ) using primer pairs SG1-2 ( IbpA ) and SG3-4 ( IbpB ) , in MG1655 creating a C-terminal fusion of IbpA and IbpB to the different fluorescent proteins . The resistance cassette was subsequently excised by transiently equipping this strain with plasmid pCP20 expressing the Flp site-specific recombinase [80] , resulting in the desired strain containing a fluorescent IbpA/IbpB fusion protein . C-terminal translational fusions of protein quality control components ( DnaK/DnaJ/ClpB/ClpP/ClpX/HtpG/HslU/Lon ) were constructed in similar fashion using primer pairs SG23-38 . The transcriptional fusions probing their expression level were constructed by first obtaining a msfgfp-frt-nptI-frt amplicon from plasmid pDHL1029 using primer pairs SG39–SG54 . The respective fluorescent transcriptional reporter strains were then created by recombineering this amplicon 5 bp after the stop codon of the gene of interest , creating an artificial operon and ensuring cotranscription . To maximize cotranslational activity , the gene encoding msfgfp was preceded by a strong synthetic ribosome binding site ( BBa_B0034; sequence AAAGAGGAGAA [81] ) . The resistance cassette was subsequently excised by transiently equipping this strain with plasmid pCP20 expressing the Flp site-specific recombinase [80] , resulting in the desired fluorescent transcriptional reporter strains . The C-terminal translational fusion of HupA to Venus was constructed using primer pair SG5-6 and pGBKD-venus as a template . Deletion strains were constructed using an amplicon prepared on pKD13 using the primers listed in the study by Baba and colleagues [79 , 82] for recombineering . This procedure replaced the genes of interest with an frt-flanked kanamycin resistance cassette , which could subsequently be excised by transiently equipping this strain with plasmid pCP20 [80] , resulting in the desired deletion strain . All constructed plasmids were verified by both PCR and sequencing ( Macrogen , the Netherlands ) . Plasmids were introduced into their respective host strains by transformation and selection for antibiotic resistance encoded by the plasmid . Plasmids pGBKD-mCherry and pGBKD-venus were constructed by integrating an mCherry/venus amplicon , generated from pDHL-mCherry and pDHL-venus with primer pairs SG7-8 ( mCherry ) and SG9-10 ( venus ) , into pGBKDparSpMT1 [83] using EcoRI and BamHI restriction sites . In addition to adding the respective restriction sites to the end of the amplicon , these primer pairs also add a flexible linker ( encoding GSGSGS; [84] ) , facilitating folding of fluorescent fusion proteins constructed with these sequences . Plasmid pSG1 was constructed by first inserting an amplicon—obtained from pGBKD-mCherry using primer pair SG11-12—into MG1655 , creating a transcriptional ibpA fusion in which the native ibpA gene was completely replaced by mCherry . From this , an amplicon ( using primer pair SG55-56 ) was obtained containing the entire 5′-upstream region of ibpA ( including its promoter and 5′-UTR ) in front of mCherry , which was subsequently blunt-end ligated into a pACYC184 vector ( [85]; opened by PCR amplifying the entire plasmid using primer pair SG13-14 ) . The resulting plasmid , pSG1 , was initially transformed to MG1655 , in which it functioned as a transcriptional reporter for ibpA expression . The plasmid was validated by exposing MG1655 pSG1 cells to a sublethal heat treatment that , as expected , led to an increase in cellular mCherry fluorescence . Plasmid pTrc99A-mCer-cI78WT was constructed by first making an mCer amplicon from pDHL-mCer with primer pair SG15-16 . These primers amplified the mCer encoding gene without a start or a stop codon and added a NcoI and BamHI restriction site to the ends of the amplicon . Digestion of the amplicon and pTrc99A vector [86] with these restriction enzymes allowed the subsequent ligation of the amplicon in the expression vector . After being verified by both PCR and sequencing , the created construct was digested again with BamHI and SalI restriction enzymes . Another amplicon containing a fragment of the lambda prophage repressor protein was generated from an in-house E . coli strain harboring the lambda prophage , using primer pair SG17-18 . These primers amplify a fragment of the lambda repressor protein named cI78 ( as it does not contain the first 77 amino acids of the normal full-length protein ) and added a BamHI and SalI restriction site as well as a flexible linker ( coding for GSGS ) to the end of the amplicon . Subsequent digestion of this amplicon allowed its ligation into the previously constructed and digested construct . The resulting plasmid , pTrc99A-mCer-cI78WT , expresses a C-terminal fusion of mCerulean3 to cI78 under control of an IPTG-inducible promoter . Plasmid pTrc99A-mCer-cI78EP8 is a plasmid expressing a misfolding and aggregating cI78 mutant , identified in a screen specifically aimed at obtaining such mutants . To find mutations that contribute to misfolding , we first performed random error-prone PCR mutagenesis [87] on cI78 ( dubbed cI78WT ) and ligated the obtained mutant library behind mCer in pTrc99A ( in similar fashion as cI78WT ) . We screened for misfolding cI78 variants by examining individual mutants under the microscope for alterations of the normally diffuse fluorescent localization pattern . From this , a cI78 mutant , cI78EP8 ( the term “EP” stemming from the error-prone PCR method used for its construction , and 8 stemming from the number of the isolated mutant ) , was identified , which displayed strict punctate and polarly located fluorescence , characteristic of disrupted folding and subsequent aggregation . cI78EP8 harbors 1 synonymous ( 180C>T ) and 3 nonsynonymous mutations ( 38A>T , 135T>A , and 343T>C ) . Plasmid pTrc99A-mCherry-cI78EP8 was created by replacing the mCer gene by mCherry . Plasmid pBAD33-ibpA-msfgfp was constructed by first generating an ibpA-msfgfp amplicon from MG1655 ibpA-msfgfp cells using primer pair SG19-20 . These primers amplify the entire gene fusion and add a strong synthetic ribosome binding site ( BBa_B0034; sequence AAAGAGGAGAA [81] ) preceding the ibpA start codon . The amplicon was subsequently blunt-end ligated into a pBAD33 vector ( [88]; opened by PCR amplifying the entire plasmid , using primer pair SG21-22 ) . The resulting plasmid was then transformed to MG1655 , in which it was used to induce expression of the fusion protein to further probe native IbpA ( i . e . , unlabeled ) behavior ( of which the resulting strain thus also carries a copy ) . For TLFM , cell suspensions were diluted appropriately , transferred to agarose pads ( containing the appropriate medium ) , placed on a microscopy slide , and mounted with a cover glass . A Gene Frame ( Thermo Scientific ) was used to hold the cover glass on the microscopy slide . TLFM was performed with a temperature controlled ( Okolab , Ottaviano , Italy; 37 °C ) Ti-Eclipse inverted microscope ( Nikon , Champigny-sur-Marne , France ) equipped with a 60× objective , a TI-CT-E motorized condenser , a YFP filter ( Ex 500/24 , DM 520 , Em 542/27 ) , a CFP filter ( Ex 438/24 , DM 458 , Em 483/32 ) , a GFP filter ( Ex 472/30 , Dm 495 , Em 520/35 ) , an mCherry filter ( Ex 562/40 , Dm 593 , Em 641/75 ) , a DAPI filter ( Ex 377/50 , DM 409 , Em 447/60 ) , and a CoolSnap HQ2 FireWire CCD-camera . Images were acquired at user-chosen time intervals using NIS-elements software ( Nikon ) . During the acquisition of TLFM recordings , care was taken to prevent potential photobleaching of fluorescent molecules ( i . e . , the photochemical alteration of a fluorophore molecule by , for example , the prolonged exposure light of excitation wavelengths , such that it permanently is unable to fluoresce ) by minimizing excitation light intensity and enlarging time intervals in between acquisition of fluorescent images . The resulting images were further handled with open source software ImageJ . Characteristics ( e . g . , length , area , fluorescence ) of individual cells growing in/into microcolonies ( and of the microcolonies themselves ) were acquired using the MicrobeTracker software [89] . In order to obtain robust results , manual curation was necessary to improve automatic segmentation and tracking . The data generated by this analysis were fed into a relational database enabling its subsequent transformation ( e . g . , calculation of certain cellular characteristics [growth rate] , establishment of genealogical relationships between cells ) and mining . The distribution of fluorescent PA foci was obtained by detecting their relative localization along the cell axis using the SpotFinder tool of MicrobeTracker [89] . Heat maps displaying the distribution of intracellular PA localization were generated using MicrobeJ , an ImageJ plugin [90] . Given the relative constancy of cell width during cell cycle progression in a given environment , cell length was employed to quantify cellular growth of individual cells . Growth rates of individual cells were determined by exponential fits of cell length over time . To examine the effect of the presence of PAs on cellular growth , only growth rates of fourth- and fifth-generation cells were considered , as these cells have grown a sufficient number of generations after the removal of the corresponding PA-inducing agent ( heat or 1 mM IPTG to induce expression of cI78EP8 ) but are not yet suffering from any ( local ) nutrient-depletion effects . To examine the effect of cI78EP8 production on cellular fitness , microcolony growth rates were determined in different induction regimes and compared to those of microcolonies producing cI78WT . Growth rates of microcolonies were determined by exponential fits of microcolony area over time ( time interval: 1–3 h after beginning of recording ) . The presence of cellular aging was examined by quantifying the fitness defect of the oldest cells within the fourth and fifth generation compared to all other cells of those generations . Cell age was inferred from old pole generations as introduced by Stewart and colleagues [17] . Consequently , each generation within a microcolony contained 2 oldest cells ( i . e . , the cells inheriting the cell original cell poles of the “founder” cell of the microcolony ) . Violin plots illustrating the extent of this effect were generated using a custom script in R . Given that incorporation of fluorescent proteins into a PA potentially compromises their structure and fluorescence [91] , we determined PA sizes directly using the ObjectDetection module within the Oufti software [92] , which allows the detection of non-diffraction-limited fluorescent regions . The mCerulean3 fluorescent protein emits at 475 nm , leading to an Abbe diffraction limit of 475 / ( 2 * 1 . 4 ) = 169 . 6 nm ( the value 1 . 4 corresponds to the numerical aperture of the microscope objective ) . The minimum size of circular spots that can subsequently be reliably detected is π * ( 169 . 6/2 ) 2 = 22591 . 3 nm2 ( or approximately 0 . 02 μm2 ) , which is significantly smaller than the smallest measured size of PAs ( 0 . 037 μm2 ) . From this , it is clear that PAs , produced by the synthetic mCer-cI78EP8 model system , occur as non-diffraction-limited spots within individual cells . These measurements also allowed us to directly investigate the potential correlation between PA size and average cellular fluorescence ( S8 Fig ) . Although an overall good correlation could be observed , this correlation appears nonlinear and noisy , especially for smaller and larger PAs . The permutation test to investigate potential contributions of PAs to the observed fitness defect of their host cells ( next to the age of these cells ) was performed by randomly sampling PA-free cells of similar age structure ( 100 , 000 times/permutations ) and comparing their average fitness to that of all other PA-free cells . A p-value was subsequently calculated as the proportion of sampled permutations in which the absolute difference in fitness was greater than or equal to the observed fitness defect of PA-bearing cells ( as compared to all PA-free cells ) . For the synthetic PA system , a similar approach could unfortunately not be employed . The large fraction of old cells also bearing a PA ( more than half the number of old cells ) made it impossible to disentangle the potential individual contribution of both phenomena to cellular fitness defects using this approach . Fifty μl of exponential phase cells was transferred aseptically to a sterile PCR tube and heat treated for 15 min at indicated temperatures in a thermocycler ( Westburg , Leusden , the Netherlands ) . Control samples were also transferred to PCR tubes but were kept at room temperature for 15 min . After heat treatments , samples were aseptically retrieved from the PCR tubes and subjected to TLFM as described above . To examine the effect of PAs on survival frequency in heat shock experiments , the same cells were microscopically examined before and after heat treatment . To accomplish this , cells were first mounted on a microscopy slide as described above and allowed to grow for an indicated number of generations while their spatial coordinates on the slide were noted . Subsequently , the slide as a whole was subjected to a heat shock ( for indicated times at indicated temperatures; heat shock durations and intensities were chosen so to inactivate approximately half of the cellular population ) by taping the slide to the lid of a thermocycler ( Westburg , Leusden , the Netherlands ) , after which the spatial coordinates were used to trace back and microscopically follow up the same cells on the heat-treated slide . For experiments in which cells were precultured and challenged in liquid cultures , and survival , together with potential presence of PAs , was examined microscopically , MG1655 ΔlacY pTrc99A-mCer-cI78EP8 cells were grown to exponential phase ( OD600 = 0 . 2–0 . 6 ) in AB medium with 0 . 2% glycerol . PA production was subsequently induced by adding 1 mM IPTG to the medium for 1 . 5 h , after which the cells were harvested and washed into AB medium with 0 . 2% glucose . After washing , cells were incubated at 37 °C for 75 min ( or longer times when indicated ) and exposed to heat , peroxide , or streptomycin stress for the indicated period of time . The fractions of surviving PA-bearing and PA-free cells were subsequently determined at the single-cell level by TLFM ( after washing away the peroxide or streptomycin in cases in which these stressors were employed ) . As PA production appeared to occasionally ( 10%–20% ) give rise to likely anucleate , PA-bearing cells ( observable as small cells with PAs filling almost their entire cytoplasm; S14 Fig ) , the fraction of surviving cells for both cellular classes ( PA-bearing and PA-free ) was compared to the fraction of outgrowing cells in unstressed control conditions to determine the relative fraction of surviving cells for each class . To examine the effect of ongoing PA production on stress sensitivity , MG1655 ΔlacY pTrc99A-mCer-cI78WT and MG1655 ΔlacY pTrc99A-mCer-cI78EP8 cells were exposed to a semilethal heat shock ( 49 °C , 15 min ) during mCer-CI78WT ( soluble ) and mCer-CI78EP8 ( i . e . , PA ) production ( AB medium with 0 . 2% of glycerol in the presence of 1 mM IPTG ) . The surviving fraction of cells was determined through spot-plating experiments in which the appropriate dilutions of a sample were prepared in PBS and subsequently spot-plated ( 5 μl ) on LB agar . After 24 h of incubation at 37 °C , the plates were counted , and the number of survivors in CFU per ml was determined . Streptomycin ( 10 μg/ml for 60 min or 15 μg/ml , final concentrations , for 30 min ) and H2O2 ( 6 mM , final concentration , for 90 min ) were directly added to exponentially growing MG1655 ibpA-msfgfp cultures . After treatments , streptomycin and H2O2 were washed away , and samples were diluted and prepared for microscopy as described above . Cellular viability ( i . e . , the relative number of cells surviving the heat sock ) was determined by TLFM . Cells that could be observed to grow and divide within an 8 h time frame after heat treatment were scored as surviving cells . Cell meshes generated by the MicrobeTracker program were used to determine resuscitation times of individual cells , as described previously [15] . Since bacterial cells typically only elongate in the longitudinal direction , resuscitation times were measured by looking at the length increase of individual cells over time . First , an initial length was calculated as the mean of the first 3 measurements for each individual cell . The length of that cell in the subsequent frames was then compared to this initial length , and the resuscitation time was defined as the time corresponding to the frame in which cell length had increased over 10% compared to its initial length , plus the time between the end of the heat treatment and the beginning of microscopy recording ( typically around 10 min ) . This 10% increase in initial length was taken as a threshold to prevent random measurement fluctuations from influencing the results and ensure that only resuscitation times of cells that had fully committed to growth were measured . In addition , only resuscitation times of surviving cells were measured , i . e . , cells that subsequently committed to growth and division . As we expected nonfunctional fusions to compromise cellular heat survival , the respective fusion proteins were validated by exposing PA-containing populations ( MG1655 ΔlacY pTrc99A-mCherry-cI78EP8 cells , with the respective fusions ) to a heat treatment and comparing inactivation levels to that of PA-containing populations of unlabeled cells ( without any additional fluorescent fusions ) . For this , cells were first grown to exponential phase ( OD600 = 0 . 2–0 . 3 ) in AB medium with 0 . 2% glycerol . PA production was subsequently induced by adding 1 mM IPTG to the medium for 1 . 5 h , after which the cells were harvested and washed into AB medium with 0 . 2% glucose . After washing , cells were exposed to a heat shock ( 52 °C , 15 min ) , and inactivation was determined through spot-plating experiments . In these experiments , the appropriate dilutions of a sample were prepared in 0 . 85% KCl and subsequently spot-plated ( 5 μl ) on AB glucose agar . After 24 h of incubation at 37 °C , the number of survivors was scored , compared to that of untreated controls , and total inactivation was determined . We further validated the transcriptional and translational fusions by exposing them to a sublethal heat shock ( 47 °C , 15 min ) , conditions known to activate the heat shock regulon of which the investigated proteins are a part of [93] . We quantified promoter activity and protein concentrations for the tested chaperones and proteases before and after heat treatment . For the transcriptional fusions , average cellular GFP fluorescence ( corresponding to promoter activity ) was determined directly after heat treatment . For the translational fusions , average cellular GFP fluorescence ( corresponding to protein concentration ) was determined after 30 min of incubation at 37 °C after heat treatment ( allowing additional time for fusion protein folding and maturation ) . Closer examination of the translational DnaK-msfGFP reporter strain revealed that this fusion protein occasionally displayed small and transient foci in unstressed PA-free cells ( Fig 9B ) . Whether these foci reflect native , functionally relevant DnaK behavior or are artefactual ( label-induced , but without severely compromising DnaK functioning ) remains to be established . To assess the variability in surviving cellular fractions , the original sample size was bootstrapped ( sampled with replacement ) 3 , 000 times , and the mean fraction of surviving cells was calculated for each of these samples . The bootstrapped estimation of the standard error of the mean fraction of surviving cells was subsequently calculated by taking the standard deviation of the bootstrapped means .
|
Since accurate protein folding is crucial for cellular viability , misfolded and aggregated proteins have typically been thought of as detrimental structures with potentially harmful physiological effects . In this report , we show that this general paradigm does not appear to hold in the model bacterium Escherichia coli . We find that the emergence of protein aggregates is mainly linked to sublethal , proteotoxic exposures ( e . g . , heat stress ) and that asymmetric partitioning of these aggregates among daughter cells may have benefits beyond mere waste disposal . In fact , cells that inherited an ancestral protein aggregate ( formed during stress exposure many generations before ) were better able to cope with subsequent exposure to proteotoxic stress . Our observations therefore reveal the existence of stress-induced , long-term , and protein aggregate–mediated “memory” in prokaryotes and highlight the potential role of protein aggregation in aiding cellular survival and adaptation in fluctuating environments .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"cell",
"physiology",
"fluorescence",
"imaging",
"cellular",
"stress",
"responses",
"medicine",
"and",
"health",
"sciences",
"engineering",
"and",
"technology",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"manufacturing",
"processes",
"green",
"fluorescent",
"protein",
"light",
"microscopy",
"luminescent",
"proteins",
"microscopy",
"heat",
"treatment",
"research",
"and",
"analysis",
"methods",
"heat",
"shock",
"response",
"imaging",
"techniques",
"proteins",
"fluorescence",
"microscopy",
"critical",
"care",
"and",
"emergency",
"medicine",
"biochemistry",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"resuscitation",
"cell",
"fusion"
] |
2018
|
Protein aggregates encode epigenetic memory of stressful encounters in individual Escherichia coli cells
|
Insulin signaling has a profound effect on longevity and the oxidative stress resistance of animals . Inhibition of insulin signaling results in the activation of DAF-16/FOXO and SKN-1/Nrf transcription factors and increased animal fitness . By studying the biological functions of the endogenous RNA interference factor RDE-4 and conserved PHD zinc finger protein ZFP-1 ( AF10 ) , which regulate overlapping sets of genes in Caenorhabditis elegans , we identified an important role for these factors in the negative modulation of transcription of the insulin/PI3 signaling-dependent kinase PDK-1 . Consistently , increased expression of pdk-1 in zfp-1 and rde-4 mutants contributed to their reduced lifespan and sensitivity to oxidative stress and pathogens due to the reduction in the expression of DAF-16 and SKN-1 targets . We found that the function of ZFP-1 in modulating pdk-1 transcription was important for the extended lifespan of the age-1 ( hx546 ) reduction-of-function PI3 kinase mutant , since the lifespan of the age-1; zfp-1 double mutant strain was significantly shorter compared to age-1 ( hx546 ) . We further demonstrate that overexpression of ZFP-1 caused an increased resistance to oxidative stress in a DAF-16–dependent manner . Our findings suggest that epigenetic regulation of key upstream signaling components in signal transduction pathways through chromatin and RNAi may have a large impact on the outcome of signaling and expression of numerous downstream genes .
The role of RNA interference ( RNAi ) in the silencing of transposons and other repetitive elements is well documented [1] , [2] , while the knowledge of its impact on endogenous genes and signaling pathways is limited . In this article we investigate the remarkable and similar effects of the Caenorhabditis elegans RNAi-promoting factors RNAi DEficient 4 ( RDE-4 ) [3] and Zinc Finger Protein 1 ( ZFP-1 ) on the expression of stress-related genes . We focus on the key gene regulated by RDE-4 and ZFP-1 , pdk-1 , which encodes 3-phosphoinositide-dependent kinase-1 ( PDK-1 ) [4] , a component of a conserved insulin-signaling pathway . We describe a functional connection between zfp-1 , rde-4 and insulin signaling in genetic epistasis experiments and demonstrate the significance of pdk-1 regulation by zfp-1 and rde-4 for C . elegans fitness . ZFP-1 , a Plant Homeo Domain ( PHD ) zinc finger protein , was first identified as a factor promoting RNAi interference in C . elegans [5]–[7] . It is a homolog of mammalian AF10 ( Acute Lymphoblastic Leukemia 1-Fused gene from chromosome 10 ) [8] and plays a key role in leukemias caused by Mixed Lineage Leukemia MLL-AF10 fusion due to the recruitment of histone methyltransferase Dot1 by the AF10 portion of the fusion protein [9] . The developmental and physiological roles of AF10 are largely unknown . RDE-4 is a double-stranded RNA ( dsRNA ) -binding protein and a component of the Dicer complex responsible for the production of short interfering RNAs ( siRNAs ) from exogenous dsRNA [10] . The rde-4 ( ne299 ) null mutation was discovered in a screen for RNAi resistant mutants [3] . rde-4 ( ne299 ) does not have obvious developmental abnormalities , but shows synthetic phenotypes when combined with the null mutant in C . elegans Retinoblastoma gene lin-35 [11] and appears to be less healthy at elevated temperatures [12] . Also , rde-4 mutants were reported to have a slightly reduced lifespan [13] . The effects of rde-4 loss-of-function are likely to be related to recently identified endogenous siRNAs ( endo-siRNAs ) , which perfectly match thousands of genes in C . elegans either in sense or antisense orientation [14]–[17] . Indeed , the expression of some endo-siRNAs is diminished in the absence of rde-4 [14] , [18] . Our recent genome-wide mRNA expression study has revealed that ZFP-1 and RDE-4 affect the transcript levels of close to 250 overlapping genes [19] . Furthermore , putative target genes of endo-siRNAs [16] showed a significant enrichment among genes upregulated in the rde-4 ( ne299 ) null [3] and zfp-1 ( ok554 ) [20] loss-of-function mutant larvae [19] . We proposed that ZFP-1 and endo-siRNAs produced in an rde-4-dependent manner cooperate in the repression of target genes in the nucleus . Here , we confirm a direct repressive effect of ZFP-1 on endo-siRNA targets by comparing gene expression changes in zfp-1 ( ok554 ) and rde-4 ( ne299 ) with genome-wide localization of ZFP-1 . Moreover , using functional analysis of misregulated genes we predict a role for RDE-4 and ZFP-1 in modulating insulin signaling and further demonstrate that regulation of pdk-1 transcription by ZFP-1 and endogenous RNAi underlies the oxidative stress sensitivity and short lifespan of zfp-1 ( ok554 ) and rde-4 ( ne299 ) mutants .
In order to elucidate the common biological roles of ZFP-1 and endogenous RNAi we analyzed gene sets misregulated in zfp-1 ( ok554 ) and rde-4 ( ne299 ) mutants [19] . We found that genes with lowered expression in the mutants compared to the wild type were enriched in metabolic , oxidative stress-related and anti-pathogenic factors present in the intestine ( Table S1 ) . Since insulin signaling mutations lead to increased expression of factors important for defense against oxidative stress and pathogens [21]–[23] , we decided to compare the lists of genes downregulated in zfp-1 ( ok554 ) and rde-4 ( ne299 ) with longevity-promoting “Class 1” genes found upregulated in the daf-2 mutant in a daf-16-dependent manner [23] . Insulin-like signaling in C . elegans via the DAF-2 insulin receptor and phosphatidylinositol 3-kinase ( PI3K ) negatively regulates the DAF-16/FOXO [24] , [25] and SKN-1/Nrf [26] transcription factors . When insulin signaling is reduced , the enhanced DAF-16 and SKN-1 activities contribute to longer lifespan and stress resistance in worms due to concerted regulation of many of their targets [21]–[23] , [27] , [28] . DAF-16 and SKN-1 are negatively regulated in part at the level of their nuclear localization; therefore , mutants in this pathway are long-lived due to a higher level of the active nuclear DAF-16 and SKN-1 and appropriate transcriptional activation or repression of their direct targets . Our analyses revealed that genes downregulated in the zfp-1 and rde-4 mutants significantly overlapped with “Class 1” longevity promoting genes upregulated in the daf-2 mutant ( a condition when DAF-16 and SKN-1 are activated ) [23] ( Figure 1 , Table 1 , Table S1 ) . Examples of genes whose expression is negatively regulated by daf-2 and positively regulated by zfp-1 and/or rde-4 include glutathione transferases gst-4 and gst-38 , and aquaporin ( aqp-1 ) ( Table 1 , Figure 2A ) . Since RNAi is a gene-silencing phenomenon and gene sets expressed lower in zfp-1 ( ok554 ) and rde-4 ( ne299 ) are not enriched in endo-siRNA targets [19] , we predict that “Class 1” longevity-promoting genes are regulated by ZFP-1 and RDE-4 indirectly . Consistently , genome-wide localization data showed no enrichment of ZFP-1 at longevity-promoting genes ( Figure 1 ) . We considered the possibility that a direct target gene negatively regulated by rde-4 and zfp-1 would be de-repressed in the mutants to account for the reduced expression of the secondary targets , which may therefore be regulated by these factors indirectly . Indeed , a component of the insulin-signaling pathway , the kinase PDK-1 , was among the most upregulated genes in zfp-1 and rde-4 [19] ( Figure 2A ) . Although our microarray study was performed on zfp-1 and rde-4 mutant larvae ( L1–L2 ) , we found that pdk-1 expression was increased in these mutants at other developmental stages as well ( Figure 2A ) . The zfp-1 gene was shown to be a direct target of DAF-16 by chromatin immunoprecipitation ( ChIP ) combined with sequencing [29] and , more recently , using chromatin profiling by DNA adenine methyltransferase identification ( DamID ) [30] . However , it was not clear whether DAF-16 had a significant role in the regulation of zfp-1 . We found that zfp-1 mRNA expression in the daf-2 mutant background was influenced by daf-16 and was 2-fold lower in the daf-2; daf-16 double mutant strain ( Figure 2B ) . Therefore , DAF-16 appears to enhance transcription of zfp-1 , although not nearly to the same extent as other prominent DAF-16 targets , such as sod-3 ( Figure 2B ) . The analyses of gene expression described above suggest a model where ZFP-1 and RDE-4 modulate the insulin-signaling pathway by repressing pdk-1 and that a DAF-16-dependent enhancement of zfp-1 expression under conditions of low insulin signaling may contribute to a positive-feedback loop enhancing the effect of DAF-16 on other targets ( Figure 2C ) . Next , we determined a molecular lesion in the weak loss-of-function pdk-1 allele sa709 [4] and tested whether the pdk-1 ( sa709 ) mutant mRNA was still regulated by ZFP-1and RDE-4 . We found that sa709 affects pdk-1 mRNA splicing and leads to the incorporation of intron three into the mature pdk-1 mRNA with a very low expression level of the correctly spliced mRNA in the mutant ( Figure 3A , 3B ) . We combined pdk-1 ( sa709 ) with zfp-1 ( ok554 ) and found the level of mutant pdk-1 mRNA expression to be elevated in the double mutant compared to pdk-1 ( sa709 ) alone ( Figure 3C ) . The pdk-1 ( sa709 ) mRNA expression was also elevated in rde-4 ( ne299 ) ; pdk-1 ( sa709 ) ( Figure 3C ) . Therefore , regulation of pdk-1 ( sa709 ) mRNA expression by ZFP-1 and RDE-4 was similar to that of wild type pdk-1 mRNA . Since loss-of-function mutations in insulin-signaling components lead to increased nuclear localization of DAF-16::GFP [31] , we tested the pdk-1 ( sa709 ) allele in this assay and found that DAF-16::GFP had more prominent nuclear localization in pdk-1 ( sa709 ) , while it was mostly cytoplasmic in wild type , zfp-1 ( ok554 ) and rde-4 ( ne299 ) worms ( Figure 3D , 3E ) . Nuclear localization of DAF-16::GFP persisted in pdk-1; zfp-1 and pdk-1; rde-4 double mutant animals ( Figure 3D , 3E ) . These results demonstrate that pdk-1 ( sa709 ) is epistatic to zfp-1 ( ok554 ) and rde-4 ( ne299 ) and support a model where ZFP-1 and RDE-4 affect expression of DAF-16 targets through regulation of pdk-1 . Since longevity-promoting genes have lower expression in the zfp-1 and rde-4 mutants , it is expected that they may live shorter than wild type worms . Indeed , a decrease in lifespan of zfp-1 ( ok554 ) [29] and rde-4 ( ne299 ) [13] has been reported , with the zfp-1 mutant exhibiting a stronger phenotype than rde-4 . In order to test whether upregulation of PDK-1 and therefore increased insulin signaling may contribute to the short lifespan of zfp-1 , we conducted epistasis experiments with a reduction-of-function mutation in the PI3 kinase AGE-1 , age-1 ( hx546 ) [32] , [33] . We found that the short lifespan phenotype of zfp-1 ( ok554 ) was suppressed by age-1 ( hx546 ) ( Figure 4A ) , i . e . the reduction in lifespan of the mutant was dependent on the active insulin signaling . Also , the extended lifespan of age-1 ( hx546 ) was dependent on ZFP-1 function , consistent with the possibility that enhanced PDK-1 dosage may suppress the defect in signaling conferred by the non-null age-1 mutation ( Figure 4A ) . Indeed , increased pdk-1 dosage suppresses the constitutive dauer phenotype of age-1 ( mg44 ) [4] . In order to show that high levels of pdk-1 expression contributed to the short lifespan of zfp-1 ( ok554 ) we attempted to combine zfp-1 ( ok554 ) with a strong loss-of-function mutation pdk-1 ( sa680 ) [4] for genetic suppression analyses . We were not able to recover zfp-1 ( ok554 ) ; pdk-1 ( sa680 ) and assume that this double mutant is not viable . Therefore , all epistasis analyses described below were performed with the sa709 allele . zfp-1 and rde-4 affect expression of multiple target genes , and some phenotypes of zfp-1 ( ok554 ) , such as dauer promotion [29] , are similar rather than opposite to the phenotypes of insulin-signaling mutants . We have found that zfp-1; pdk-1 and rde-4; pdk-1 double mutants display some egg-laying deficiency , which complicates the longevity assays that we conduct in the absence of drugs inducing sterility . However , although zfp-1; pdk-1 and rde-4; pdk-1 worms were undoubtedly sicker than zfp-1 or rde-4 single mutants , we found that the reduction of pdk-1 function significantly suppressed the decreased lifespans of zfp-1 ( ok554 ) and rde-4 ( ne299 ) ( Figure 4B , 4C ) . These results further support the idea that ZFP-1 and RDE-4 affect insulin signaling through the negative regulation of pdk-1 . The gene expression signatures of zfp-1 and rde-4 mutants suggested that they could be deficient in oxidative stress response . We induced oxidative stress by soaking L4 animals in 100mM paraquat and found that the zfp-1 ( ok554 ) mutant strain was much more sensitive to this treatment compared to the wild type ( Figure 4D ) , similarly to daf-16 ( mu86 ) ( Figure 4D ) , while rde-4 ( ne299 ) showed moderate sensitivity ( Figure 4D , Figure S1A ) , and age-1 ( hx546 ) and pdk-1 ( sa709 ) were more resistant than wild type ( Figure 4D ) . We found that zfp-1; age-1 , zfp-1; pdk-1 and rde-4; age-1 , rde-4; pdk-1 double mutants were less sensitive to oxidative stress than zfp-1 and rde-4 , respectively ( Figure 4D ) , indicating that the stress sensitivity of zfp-1 ( ok554 ) and rde-4 ( ne299 ) was due to active insulin/PI3K signaling . In order to determine whether increased pdk-1 expression may be sufficient to cause a stress sensitivity phenotype , we tested the SP940 strain , which contains the free duplication mnDp ( II;X;f ) that includes the pdk-1 locus . We found that pdk-1 mRNA levels are increased about 2 . 5-fold in this strain ( Figure S1C ) , close to that observed in rde-4 ( ne299 ) , and it shows a comparable sensitivity to paraquat ( Figure S1A , S1B ) . These data are consistent with the idea that regulating pdk-1 dosage is important for animal fitness . We generated transgenic lines expressing ZFP-1::GFP and ZFP-1::FLAG fusion proteins by introducing tags into the C-terminal region of ZFP-1 through fosmid recombineering in bacteria [34] . The resulting genes are expressed from the 30kb fosmid and are subject to the same regulatory inputs as the endogenous zfp-1 locus; the transgenes fully rescued the stress sensitivity and reduced lifespan phenotypes of the zfp-1 mutant ( Figure 5A , 5B ) . zfp-1 mRNA expression was about two-fold greater in ZFP-1 transgenic lines compared to the control ( Figure S2 ) . Moreover , we found that these ZFP-1 overexpressing lines were more resistant to oxidative stress compared to the control line generated by a similar technique of unc-119 mutant rescue but not containing the ZFP-1 fosmid ( Figure 5A ) . The stress resistance of ZFP-1 overexpressing lines was dependent on DAF-16 function ( Figure 5A ) . This is consistent with the repression of insulin signaling and therefore indirect activation of DAF-16 by ZFP-1 . We have not observed lifespan extension in the ZFP-1 overexpressing lines ( Figure 5B ) , which indicates that a higher level of ZFP-1 may be advantageous only in acute stress situations . An example of an acute stress response is the response of animals to pathogens . The human pathogenic bacterium Pseudomonas aeruginosa ( PA14 ) inhibits DAF-16 nuclear localization and therefore downregulates the expression of defense factors that are dependent on DAF-16 [35] . We tested the effect of the loss of ZFP-1 function on innate immunity by assaying the survival of zfp-1 ( ok554 ) animals . Upon exposure to PA14 under the standard infection assay conditions [36] , we observed that the zfp-1 ( ok554 ) mutants were significantly more susceptible to P . aeruginosa infection- mediated killing ( Figure 6A , 6B ) . The pathogen sensitivity seen in zfp-1 ( ok554 ) mutants was due to loss of ZFP-1 function as was confirmed using a ZFP-1::GFP transgene that rescued the mutant phenotype ( Figure 6B ) . Next , we tested whether the increased susceptibility of zfp-1 ( ok554 ) to PA14 was dependent on insulin signaling . We confirmed that age-1 ( hx546 ) was more resistant to the infection ( Figure 6C ) and tested age-1; zfp-1 double mutants . The results were similar to those found in the longevity assays: age-1 and zfp-1 suppressed each other's phenotypes ( Figure 6C ) . The survival of the double mutant was closer to that of zfp-1 ( ok554 ) than age-1 ( hx546 ) , although age-1 significantly suppressed the sensitivity of zfp-1 to PA14 killing . We conclude that PI3K signaling contributes to the pathogen-sensitivity of zfp-1 ( ok554 ) . Consistent with our expression and genetic epistasis data suggesting a direct role of ZFP-1 in repressing pdk-1 transcription , a strong peak of ZFP-1 localization was found at the pdk-1 promoter in ChIP/chip experiments conducted by the modENCODE ( model organism ENCyclopedia Of DNA Elements ) project ( Figure 7A ) . We confirmed ZFP-1 localization to the pdk-1 promoter by ChIP/PCR with antibodies specific to endogenous ZFP-1 ( Figure 7B ) as well as with anti-FLAG antibodies in experiments with ZFP-1::FLAG transgenic lines ( Figure 8A ) . ZFP-1 was not localized to the promoters of other genes of the insulin signaling pathway ( daf-2 , age-1 , akt-1 , sgk-1 ) ; it was also not present at DAF-16 target genes that have reduced expression in zfp-1 ( ok554 ) and appear to be positively regulated by this factor , as discussed earlier ( Figure 1 , Figure 7B , Figure 8A , and Table S1 ) . There was no enrichment in direct ZFP-1 targets among the longevity-promoting genes ( P-value 0 . 83 ) ( Figure 1 ) . Therefore , ZFP-1 is likely to target directly only genes whose transcription it inhibits ( Table S1 ) . We reported previously a very significant overlap between genes negatively regulated by zfp-1 and rde-4 and endogenous siRNA target genes [19] . Consistently , we find that direct ZFP-1 target genes are overrepresented among genes expressed higher in the rde-4 mutant ( Figure 7C ) . pdk-1 is repressed by ZFP-1 and is also negatively regulated by RDE-4 , which is a dsRNA-binding protein required for the biogenesis of siRNAs in the exogenous RNAi pathway [10] and contributing to the biogenesis of some endo-siRNAs [14] , [18] . Knowing this , we searched available deep sequencing data [37]–[40] for endo-siRNAs mapping to the pdk-1 locus . There were few endo-siRNAs corresponding to the coding region of pdk-1 , and more siRNAs mapped to the promoter region of the gene ( 5kb upstream of the transcription start site ) , including a predicted open reading frame , H42K12 . 2 . ( Figure 7A ) . However , for this open reading frame , no transcriptional evidence exists , neither from EST collections nor from deep sequencing runs undertaken in the context of the modENCODE project , and it therefore appears to be a mis-annotated gene [41] , [42] . We were able to detect ∼100–250 nt transcripts at the pdk-1 promoter produced from both the plus and minus DNA strands , consistent with the possibility of dsRNA production and processing by RDE-4 and Dicer ( Figure 7D and Text S1 ) . Moreover , we detected an elevated level of this RNA in the rde-4 mutant ( Figure 7E ) , further supporting the possible involvement of RDE-4 in the dsRNA processing . Unfortunately , pdk-1 promoter-specific endo-siRNAs are expressed at a very low level , and we were not able to reliably detect them by RT-qPCR . Nevertheless , additional evidence for pdk-1 regulation by endogenous RNAi comes from the observation that pdk-1 mRNA levels are increased in drh-3 ( ne4253 ) , a loss-of-function mutant in dicer-related helicase 3 [38] , ( Figure 8B ) . DRH-3 is thought to participate in multiple branches of endogenous RNAi in C . elegans [38] . We have shown that both zfp-1 and rde-4 affect the longevity of C . elegans and its ability to resist oxidative stress and that pdk-1 mRNA levels are elevated in zfp-1 ( ok554 ) and rde-4 ( ne299 ) ( [19] and Figure 1 and Figure 3 ) . Furthermore , we have demonstrated that ZFP-1 binds the pdk-1 promoter and that endogenous siRNAs also have a potential to regulate pdk-1 directly . Next , we analyzed RNA polymerase II ( Pol II ) occupancy at the pdk-1 promoter and coding region by ChIP in wild type , zfp-1 ( ok554 ) and rde-4 ( ne299 ) L3-L4 animals and found it to be significantly increased in both mutants ( Figure 8C , 8D ) . Consistent with transcriptional regulation , pdk-1 pre-mRNA levels were elevated in both mutants as well ( Figure 8E ) . RDE-4 , and therefore rde-4-dependent endo-siRNA production , did not affect ZFP-1 localization to the pdk-1 promoter ( Figure 8A ) . It is possible that ZFP-1 and the RNAi machinery are independently recruited to the same targets and cooperate in inhibiting their transcription . Alternatively , ZFP-1 may help stabilize downstream RNAi factors at the endo-siRNA target genes . Pol II levels increased only at the promoters , but not at the coding regions of indirect target genes expressed lower in zfp-1 ( ok554 ) and rde-4 ( me299 ) ( Figure 8C , 8D ) , a signature consistent with a slower rate of transition from transcriptional initiation to elongation [43] . This finding reflects the lower expression of these genes in the mutants , although they are not regulated directly by ZFP-1 and do not belong to the group of prevalent endo-siRNA targets ( [19] and Table S1 ) . We have previously described a very significant overlap between genes misregulated in zfp-1 ( ok554 ) and genes misregulated in rde-4 ( ne299 ) and noted that the level of expression of zfp-1 mRNA did not change in rde-4 ( ne299 ) and vice versa [19] . Since the rde-4 mutation has milder effects on gene expression than zfp-1 ( ok554 ) , they could potentially be due to zfp-1 misregulation . Therefore , we further confirmed that protein levels of ZFP-1 are not decreased in rde-4 ( ne299 ) ( Figure S3 ) . ZFP-1 localizes to the pdk-1 promoter and both the pdk-1 mRNA level and Pol II occupancy at the pdk-1 gene are increased in zfp-1 ( ok554 ) . These findings strongly suggest that transcription of pdk-1 is directly and negatively modulated by ZFP-1 . Our genetic and molecular data also clearly demonstrate that rde-4 has a role in the transcriptional regulation of pdk-1 . Several lines of evidence provide correlative support for a possible direct role of endo-siRNAs in pdk-1 regulation: endo-siRNAs match the pdk-1 promoter in a region also targeted by ZFP-1 , dsRNA production is detected at the promoter and is increased in rde-4 ( ne299 ) , and pdk-1 mRNA levels are elevated in at least two RNAi pathway mutants . However , since the endo-siRNAs targeting pdk-1 are not very abundant , we were not able to determine whether they change in rde-4 ( ne299 ) , and there is a possibility that rde-4 affects pdk-1 transcription indirectly . In either case , RDE-4 is most likely involved in gene regulation through endo-siRNA production since this is the only known molecular function of this protein . The relatively more abundant endo-siRNAs matching the pdk-1 promoter ( Figure 7A ) are not unique and correspond to the repeat sequences . The Argonaute proteins that bind endo-siRNAs and work downstream in the RNAi pathways have been described and include at least two separate branches: the CSR-1 branch [37] and the WAGO branch [38] . Although CSR-1-bound endo-siRNAs are enriched in sequences antisense to protein-coding genes , they also include endo-siRNAs matching repeats [37] , while the WAGO system appears to preferentially target repeats and pseudogenes [38] . Both the WAGO and CSR-1 systems have been shown to have some connection to the RDE-4-regulated genes [38] , [44] , and ZFP-1 ChIP/chip targets are enriched in both WAGO and CSR-1-dependent endo-siRNA target gene sets ( G . Cecere , M . Jensen , et al . , manuscript in preparation ) . Therefore , we think that regulation of some endogenous genes , such as pdk-1 , which contain simple repeats in their promoters , may have evolved to depend on the RNAi surveillance system , either WAGO or CSR-1-based .
This work has revealed a direct repression of pdk-1 transcription by C . elegans AF10 homolog ZFP-1 and the significance of this transcriptional regulation in modulating insulin signaling . We have demonstrated that overexpression of ZFP-1 leads to enhanced resistance to oxidative stress in nematodes in a DAF-16-dependent manner . The role of DAF-16/FOXO in longevity and stress response is conserved in animals [45] , and it would be interesting to see whether AF10 has a role in promoting stress resistance through the activation of FOXO . FOXO proteins have been shown to cause a neuroprotective effect in C . elegans , Drosophila and mammalian models of neurodegeneration [46] . Another transcription factor involved in the antioxidant response , Nrf2 – a homolog of C . elegans SKN-1 - has been implicated in the neuroprotection of motor neurons in a mouse model of ALS [47] , while SKN-1 was shown to be important for protection of dopamine neurons against methylmercury-induced degeneration in C . elegans [48] . Since both DAF-16 and SKN-1 are negatively regulated by insulin/PI3K signaling in C . elegans [45] ( Figure 2C ) , perhaps inhibition of this signaling pathway in mammalian neurons will lead to activation of both FOXO3a and Nrf2 . Our work suggests that the homolog of ZFP-1 , AF10 , may have a neuroprotective effect by indirectly activating FOXO3a and Nrf2 if the regulation of pdk-1 by ZFP-1/AF10 is conserved in animals . RNAi was discovered in C . elegans as a response to exogenously introduced dsRNA [3] , [49] and was considered to be primarily an anti-viral mechanism also directed against repetitive elements [50] , especially since the first RNAi-resistant mutants did not have obvious developmental phenotypes [3] . The discovery of mutants in RNA-dependent RNA polymerase ( RdRP ) genes that displayed developmental phenotypes [51] , [52] and were either RNAi-resistant [51] or more sensitive to exogenous RNAi [52] , highlighted the possibility that RNAi may be used for regulating endogenous genes . Indeed , endogenous siRNAs antisense to protein-coding genes and similar to those produced during exogenous RNAi were discovered first in the worm [15] and then in other animals [53] . It became apparent that in mutants lacking specific endo-siRNAs , corresponding mRNAs become upregulated [14] , [54] , [55] , and microarray and deep sequencing approaches have been used for identifying genes that change expression in the RNAi mutants [13] , [14] , [19] , [37]–[39] , [44] , [56]–[58] . However , the significance of misregulation of specific genes for the biology of the worm has not been clearly demonstrated and phenotypes described for RNAi-related mutants [13] , [37] , [51] , [54]–[61] were not connected to specific targets by functional epistasis experiments . This study interprets the microarray signature of zfp-1 and rde-4 mutants , demonstrates short lifespan and stress sensitivity phenotypes consistent with the signature , and provides functional evidence that pdk-1 is a major target responsible for these phenotypes through genetic epistasis , RNA expression and ChIP analyses . RNAi in C . elegans has the potential to cause both post-transcriptional [49] and transcriptional [6] , [62] gene silencing . It is possible that endogenous RNAi utilizes multiple mechanisms and that some genes are subject mostly to post-transcriptional regulation while others are regulated at the transcriptional level; the latter are likely to have fewer matching endo-siRNAs to the coding region and relatively more promoter-specific endo-siRNAs , like pdk-1 . We surveyed the genes upregulated in rde-4 ( ne299 ) for an endo-siRNA signature similar to that of the pdk-1 locus and found a number of examples ( Figures S4 and S5 ) . Interestingly , most of these types of genes , including pdk-1 , have repetitive elements at the promoters and endo-siRNAs matching them . It appears that a modulating effect of RDE-4 on the transcription of some endogenous genes is linked to the control of repetitive elements . RNAi-dependent silencing of long terminal repeats ( LTR ) and non-coding RNA genes located in euchromatic regions that functions with trace amounts siRNAs has been described recently in S . pombe [63] . The lack of abundant siRNA species was remarkable , considering that Dicer and RdRP interacted physically with the loci and that LTR transcript levels were significantly elevated in the dcr1 , ago1 and rdp1 mutants . This type of RNAi-based regulation appears to be very similar to that operating on the pdk-1 gene in C . elegans that we describe here . Examples of genes regulated by RNAi through repetitive elements in promoters already exist in Arabidopsis and include the FWA gene , which affects flowering time [64] , [65] and , more recently , an extracellular peroxidase Ep5C gene [66] . High levels of Ep5C promote susceptibility to Pseudomonas syringae and mutation in the Argonaute 4 gene was recovered in an unbiased screen for increased susceptibility to infection [66] . It is interesting that both in plants and animals regulation of endogenous genes by RNAi has evolved to promote fitness .
Strains were maintained at 20°C unless otherwise noted , using standard methods [67] . The following mutants were used: LGI: daf-16 ( mu86 ) , daf-16 ( mgDf50 ) , LGII: age-1 ( hx546 ) , LGIII: daf-2 ( e1370 ) , rde-4 ( ne299 ) , zfp-1 ( ok554 ) , LGX: pdk-1 ( sa709 ) . Compound mutant strains and transgenes used are as follows: CF1595: daf-16 ( mu86 ) I; daf-2 ( e1370 ) III , AGK138: zfp-1 ( ok554 ) III; pdk-1 ( sa709 ) X , AGK241: rde-4 ( ne299 ) III; pdk-1 ( sa709 ) X , AGK25: age-1 ( hx546 ) II; zfp-1 ( ok554 ) III , AGK264: age-1 ( hx546 ) II; rde-4 ( ne299 ) III , AGK72: daf-16 ( mgDf50 ) I; armEx5 , TJ356: zIs356 IV , AGK30: zfp-1 ( ok554 ) III; zIs356 IV , AGK262: zfp-1 ( ok554 ) III; zIs356 IV; pdk-1 ( sa709 ) X , AGK377: rde-4 ( ne299 ) III; zIs356 IV , AGK 265: rde-4 ( ne299 ) III; zIs356 IV; pdk-1 ( sa709 ) X , AGK267: zfp-1 ( ok554 ) unc-119 ( ed3 ) III; armIs5 , AGK248: rde-4 ( ne299 ) zfp-1 ( ok554 ) unc-119 ( ed3 ) III; armIs5 , AGK260: zIs356 IV; pdk-1 ( sa709 ) X , SP940: unc-52 ( e444 ) II; unc-1 ( e538 ) X; mnDp11 ( II;X;f ) . Transgenic worms were created by microparticle bombardment using a PDS-1000 Hepta Apparatus ( Bio-Rad ) [68] . All strains were made by co-bombardment of both a fosmid of interest and plasmid pMM016b ( AddGene ) for unc-119 ( ed3 ) III rescue . Strains created are as follows: AGK29: armIs2 Is[unc-119+] – control strain , AGK128: armIs5 Is[ZFP-1::FLAG , unc-119+] , AGK26: armEx5 Ex[ZFP-1::GFP , unc-119+] . The WRM0629bD09 fosmid containing the ZFP-1 locus was obtained from the C . elegans fosmid library generated by C . elegans Reverse Genetics Core Facility , Vancouver , B . C . , Canada . http://www . lifesciences . sourcebioscience . com/clone-products/genomic-dna-clones . aspx We generated derivative fosmid constructs to express recombinant ZFP-1 protein tagged with GFP or FLAG at the C-terminal portion of the protein by a fosmid recombineering method as described by [34] . Paraquat sensitivity assays were done essentially as described by [69] . L4 animals were transferred from NGM agar plates into 24-well plates ( 10 per well ) containing 300 µL of 100 mM paraquat dissolved in M9 . Worms were then incubated at 20°C and scored for survival after 20 hours . Dead animals were scored by their continuous absence of swimming movements and pharyngeal pumping . A t-test between two means was used to calculate statistical significance . Assays were performed as described by [70] . Worms were kept at 20°C on NGM plates ( 10 animals per plate ) . Day of hatching was used as the first time point . Dead animals were scored as dead when they refused to move after repeated prodding with a pick . Animals that crawled away from the plate , exploded , or contained internally hatched worms were excluded from the analysis . Life spans were determined in parallel for all strains shown together on graphs . Statistical significance was determined by a log-rank analysis using Prizm software . C . elegans survival assays were performed as described earlier [36] . To avoid the confounding effects of varying brood sizes , egg laying rates and progeny hatching within the infected worms on worm mortality , we used worms rendered sterile by RNAi of pos-1 , loss of which results in inviable embryos [71] , [72] . Worms that died due to desiccation on the walls of the Petri dish or due to bursting vulva were censored from further analysis . Statistical analysis was performed using Kaplan-Meier non-parametric survival analysis using the software Statview ( Version 5 . 0 . 1 SAS Institute Inc . ) . P<0 . 001 was considered significantly different than wild type . Since the addition of the DAF-16::GFP transgene to the zfp-1 ( ok554 ) ; pdk-1 ( sa709 ) double mutant strain led to a penetrant dauer phenotype at 20°C , all DAF-16::GFP strains were maintained at 16°C . L4 and adult stage worms were used for scoring nuclear localization . Worms were mounted on agarose pads and DAF-16::GFP localization was assessed in 10–20 worms at a time using 200X magnification on a Zeiss AxioImager Z1 immediately , higher magnification images of DAF-16::GFP localization in intestinal cells were done at 630X . Synchronous populations of animals were grown at 20°C on NGM plates seeded with OP50 E . coli at a density of approximately 100 , 000 animals per 15 cm Petri dish and harvested at specific stages of development . The harvested animals were washed three times with M9 buffer and the pellet was frozen in dry ice with TRI Reagent ( MRC , Inc . ) . After five times of freeze and thaw , total RNA was isolated according to the TRI Reagent protocol . Ten micrograms of the total RNA sample was digested with 2U of Turbo DNase ( Ambion ) at 37°C for 1hr followed by phenol-extraction and ethanol-precipitation . cDNA was generated from 2 µg of total RNA , using oligo-dT primer and RevertAid Reverse Transcriptase ( Fermentas ) . Quantitative PCR was performed on the Mastercycler ep realplex ( Eppendorf ) using the QuantiFast SYBR Green PCR Kit ( Qiagen ) . Thermocycling was done for 40 cycles in a two-step cycling , according to the manufacturer's instructions , with 25 µl of reaction containing 12 . 5 µl SYBR master mix , 0 . 15 µl of 100 µM primers , 5 µl of diluted cDNA , and 7 . 2 µl dH2O . Each PCR reaction was performed in triplicate . We used the ΔΔCt method to quantify the change in mRNA expression in the mutant samples compared to wild type and act-3 mRNA was used as a reference gene . The primers used were as follows: Forward CACGAGACTTCTTACAACTCC and Reverse GCATACGATCAGCAATTCCT for act-3 mRNA detection , Forward AGCCATCAACACCGTCTAAC and Reverse CGAATTGGCGCGTGGTGC for pdk-1 mRNA detection , Forward GCTAGGATGTCAGGTGGTC and Reverse CCAAGAGAAGCCACGAAAGC for aqp-1 mRNA detection , Forward ATGCTCGTGCTCTTGCTGAG and Reverse GACTGACCGAATTGTTCTCCAT for gst-4 mRNA detection , Forward TACCGATGAGGAGTGGGAGA and Reverse CGAATTCCCGAGCAAGATAA for gst-38 mRNA detection , Forward TTTCAGAATCACAGAGCAACAC and Reverse TGCGATACATGTTCAGAAGAG for zfp-1 mRNA detection , Forward ACACTATTAAGCGCGACTTCG and Reverse AGTTGGCAATCTTCCAAATAGC for sod-3 mRNA detection , Forward pdk-1 ex2-ex3 junction CCTACAGCCAGGTATTCCG and Reverse pdk-1 intron 3 ACAAGTGGATTTTGATGGGTTC for detecting the mutant sa709 pdk-1 mRNA and pre-mRNA and Reverse pdk-1 ex3-4 junction GATCACGAAATAAATTCTAGCCTGG-for detecting the wild-type pdk-1 mRNA . For detection of bi-directional transcription at the pdk-1 promoter the primers used were as follows . Region 1 RT primers: detecting ( − ) strand transcript CCGAGGTTATAATTTTGGCTAAACTT; detecting ( + ) strand transcript ATCAAGAGATACAGCGGGAG . Region 1 PCR primers: forward- CGGAGTTATAACCAAGCAACCA reverse- GTGTCAACTGGATATGAATCCGAA Region 2 RT primers: detecting ( − ) strand transcript CTCCCGCTGTATCTCTTGAT detecting ( + ) strand transcript GTACGGTTGTTATCGCTTTCAGG . Region 2 PCR primers: forward - GAATGTTCAAAGCCTTAAAGC reverse – AGGGATAATTGGAGTGACATGG . Chromatin immunoprecipitation was performed following the modENCODE Protocol from the Lieb Lab with the following modification: 2 . 5–3mg of cross-linked extract from L3 or adult worms was incubated for 1h at 4°C with the specific antibody and the immune complexes were then incubated with 60 µl IgG Dynabeads ( Invitrogen ) for 1h at 4°C . DNA was cleaned up with the Qiagen PCR purification kit . For the FLAG ChIP , we incubated the cross-linked extract with ANTI-FLAG M2 Affinity Gel ( Sigma ) for 2h at 4°C and , after the washing steps , eluted with 300 µg/ml of FLAG peptide ( Sigma ) for 30min at 4°C . The other antibodies used were anti-ZFP-1 ( generated by the Lieb Lab ) and anti-Pol II 8WG16 ( Covance ) . The immunoprecipitated DNA was quantified by qPCR using the ΔΔCt method to calculate the percentage of immunoprecipitation relative to the input . We used the following specific primers: Forward AAACAACACATAGACTTGTGCC and Reverse GTACGGTTGTTATCGCTTTCAG to amplify the promoter region of the pdk-1 gene; Forward pdk-1 ex2 GCAAGTGAATCGGAGAACAG and Reverse pdk-1 ex2 TGAAGAAACATGAAGTGCTTGG to amplify the coding region of the pdk-1 gene; Forward TTTCAGAACTATCATGCCACG and Reverse TCTCTGAGCACACTTTGAGG to amplify the promoter region of the aqp-1 gene; Forward aqp-1 ex5 TTGCCAGTTATCCATCTCCA and Reverse aqp-1 ex5 CTCTCATCAATAACAACGCAG to amplify the coding region of the aqp-1 gene; Forward TTAGATAGAGAATTGGCGAGAG and Reverse CAAGTAGCAAAGCGATAAACC to amplify the promoter region of the gst-4 gene; Forward gst-4 ex4 TGAAGTTGTTGAACCAGCC and Reverse gst-4 ex4 CCCAAGTCAATGAGTCTCCA to amplify the coding region of the gst-4 gene . To investigate the function of ZFP-1 with ChIP we first developed an antibody ( termed JL00006_ZFP1 ) to the C-terminal portion of the protein . Alternative transcription start sites give rise to two ZFP-1 protein isoforms with identical C-terminal domains . As expected , both isoforms are recognized by the JL00006_ZFP1 antibody . The protocols used for generating ZFP-1 ChIP/chip data are described at http://www . modencode . org/Lieb . shtml . C . elegans genes ( refSeq id ) from genome build CE4 ( ws170 ) were extracted from the UCSC genome browser's refGene table . A gene was called bound by ZFP-1 if the center base pair of a ZFP-1 peak overlapped the ORF or the 1 , 500 bp upstream region . Overlap calls were done using the Galaxy web tool . Of the total 24 , 901 genes , 3 , 598 were bound by ZFP-1 . Genome-wide ZFP-1 localization data are available at modENCODE: http://intermine . modencode . org/ .
|
Reduced activity of the insulin-signaling pathway genes has been associated with a longer lifespan and increased resistance to oxidative stress in animals due to the activation of important transcription factors , which act as master regulators and affect large networks of genes . The ability to manipulate insulin signaling and reduce its activity may allow activation of oxidative-stress response programs in pathological conditions , such as neuronal degeneration , where oxidative stress plays a significant role . Here , we describe a new way of inhibiting insulin signaling that exists in the nematode Caenorhabditis elegans . We find that transcription of one of the insulin-signaling genes is inhibited by mechanisms involving chromatin and RNA interference , a silencing process that depends on short RNAs . We demonstrate that mutants deficient in RNA interference are more susceptible to stress due to increased insulin signaling and that increased dosage of a chromatin-binding protein repressing insulin signaling and promoting RNA interference leads to better survival of nematodes grown under oxidative stress conditions . Since there is a clear homolog of this chromatin-binding protein in mammals , it may also act to promote resistance to oxidative stress in human cells such as neurons .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"model",
"organisms",
"genetics",
"biology",
"genomics",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] |
2011
|
A Conserved PHD Finger Protein and Endogenous RNAi Modulate Insulin Signaling in Caenorhabditis elegans
|
Genome-wide association studies have revealed an association between coronary heart disease ( CHD ) and genetic variation on chromosome 13q34 , with the lead single nucleotide polymorphism rs4773144 residing in the COL4A2 gene in this genomic region . We investigated the functional effects of this genetic variant . Analyses of primary cultures of vascular smooth muscle cells ( SMCs ) and endothelial cells ( ECs ) from different individuals showed a difference between rs4773144 genotypes in COL4A2 and COL4A1 expression levels , being lowest in the G/G genotype , intermediate in A/G and highest in A/A . Chromatin immunoprecipitation followed by allelic imbalance assays of primary cultures of SMCs and ECs that were of the A/G genotype revealed that the G allele had lower transcriptional activity than the A allele . Electrophoretic mobility shift assays and luciferase reporter gene assays showed that a short DNA sequence encompassing the rs4773144 site interacted with a nuclear protein , with lower efficiency for the G allele , and that the G allele sequence had lower activity in driving reporter gene expression . Analyses of cultured SMCs from different individuals demonstrated that cells of the G/G genotype had higher apoptosis rates . Immunohistochemical and histological examinations of ex vivo atherosclerotic coronary arteries from different individuals disclosed that atherosclerotic plaques with the G/G genotype had lower collagen IV abundance and thinner fibrous cap , a hallmark of unstable , rupture-prone plaques . A study of a cohort of patients with angiographically documented coronary artery disease showed that patients of the G/G genotype had higher rates of myocardial infarction , a phenotype often caused by plaque rupture . These results indicate that the CHD-related genetic variant at the COL4A2 locus affects COL4A2/COL4A1 expression , SMC survival , and atherosclerotic plaque stability , providing a mechanistic explanation for the association between the genetic variant and CHD risk .
Coronary heart disease ( CHD ) is a multifactorial disorder caused by both genetic and life-style factors . Genome-wide association studies ( GWASs ) have revealed a relationship between the disease and genetic variation on chromosome 13q34 [1 , 2] . The lead CHD-associated single-nucleotide-polymorphism ( SNP ) in this genomic region was rs4773144 located in the third intron of the COL4A2 gene , with the G allele of this SNP associating with increased CHD risk [1 , 2] . The molecular and cellular mechanisms underlying this genetic association have , however , remained unclear . Further investigations into such mechanisms are required . The COL4A1 and COL4A2 genes reside next to each other in the head-to-head orientation on chromosome 13q34 and share common transcriptional regulatory sequences [3–9] . These two genes encode the collagen IV protein α1 and α2 chains , respectively [3–9] . Collagen IV is the major constituent of the basement membrane and is essential for its integrity and functionality [10] . In the blood vessel wall , the basement membrane underlies the endothelium and surrounds smooth muscle cells ( SMCs ) [11] . The basement membrane not only serves as an extracellular scaffold but also regulates cell behavior [10 , 12] . Abnormalities of vascular endothelial cells ( ECs ) and SMCs play important roles in the pathogenesis of atherosclerosis , the vascular pathology underlying CHD [13] . In this study , we sought to investigate the functional effects of the CHD-associated SNP rs4773144 and found that it has an impact on COL4A2/COL4A1 expression , vascular SMC survival , and coronary atherosclerotic plaque stability .
To investigate if SNP rs4773144 genotype had an effect on COL4A2 and COL4A1 expression , we carried out quantitative RT-PCR assays using total RNA samples extracted from primary cultures of vascular SMCs and ECs from different individuals ( n = 148 and n = 137 , respectively ) . The assays showed a genotypic effect of rs4773144 on COL4A2 RNA levels in both cell types , with the G allele associating with lower expression level in an additive fashion ( Fig 1A and 1B ) , and a similar effect on COL4A1 RNA levels ( Fig 1C and 1D ) . The above results suggest that rs4773144 per se , or a SNP in linkage disequilibrium with it , influences COL4A2 and COL4A1 expression . The COL4A1 and COL4A2 genes , situated on chromosome 13q34 , are in a head-to-head arrangement , being separated by a bidirectional promoter that drives the transcription of COL4A1 and COL4A2 in opposite directions [3–6] . Studies have shown that transcription of these two genes is modulated by regulatory sequences located in the first intron of each gene and the third intron of COL4A2 [5 , 7–9] . SNP rs4773144 resides in the third intron of COL4A2 , and is in strong linkage disequilibrium ( r2>0 . 8 ) with 3 other SNPs ( rs4773143 , rs7986871 and rs3809346 ) which are also located in intron 3 of COL4A2 . A bioinformatics analysis showed that these SNPs reside in a genomic region that has important transcriptional regulatory features including H3k27Ac marks and DNase I hypersensitivity , identified by the ENCODE and Roadmap Epigenomics project ( S1 Fig and S2 Fig ) . Therefore , we investigated whether SNP rs4773144 affects gene transcription . To this end , we carried out chromatin immunoprecipitation to capture transcriptionally active chromatins using an antibody against RNA polymerase II ( Pol II ) and then performed allelic imbalance analyses of the immunoprecipitated chromatins from cells that were heterozygous for rs4773144 . The analyses showed that in both SMCs and ECs , the ratio of the rs4773144 G allele versus the A allele was lower in chromatin samples precipitated by the anti-Pol II antibody than in non-precipitated chromatin samples ( Fig 2A and 2B ) , indicating that the G allele has lower transcriptional activity than the A allele . An analogous analysis using an anti-H3k27Ac antibody showed a similar trend ( S3 Fig ) . To further test if the DNA sequence at the rs4773144 site can modulate gene transcription , we performed a luciferase reporter gene assay . In this experiment , SMCs were transfected with a firefly luciferase gene plasmid containing either the rs4773144 G or A allele sequence as well as a renilla luciferase gene plasmid to serve as a transfection efficiency reference . The experiment showed that G allele plasmid transfectants had lower firefly luciferase levels than A allele plasmid transfectants ( Fig 3A ) , suggesting that the G allele sequence had lower activity in driving gene transcription . Further experiments using the electrophoretic mobility shift assay technique showed binding of a nuclear protein to oligonucleotide probes corresponding to the DNA sequence at the rs4773144 site , with lower binding efficiency for the G allele [the DNA-protein complex band had a higher intensity in assays with labeled A allele probe ( lane 2 ) than in assays with labeled G allele probe ( lane 9 ) , and was reduced more readily by unlabeled A allele probe ( lane 4 ) than by unlabeled G allele probe ( lane 6 ) ] ( Fig 3B ) . A bioinformatics analysis showed that the DNA sequence at the rs4773144 site ( TTCACGGGA[A/G] ) shared similarity with the consensus binding element ( TTCNNNNGAA ) of the transcription factor STAT3 ( S4 Fig ) . To investigate if the protein in the DNA-protein complex mentioned above was STAT3 , we performed electrophoretic mobility super-shift assay using an anti-phospho-STAT3 antibody; however , the formation and mobility of the DNA-protein complex mentioned above was unaffected by the antibody ( S5 Fig ) , suggesting that it may involve a different protein . Collagen IV is the major constituent of the basement membrane and participates in cell-matrix and cell-cell communication [14] . Studies have suggested that collagen IV binds to integrins and activates integrin-mediated intracellular signaling , consequently promoting endothelial cell proliferation and inhibiting apoptosis [15–19] . Integrins promote cell survival and inhibit apoptosis , in part by up-regulating the expression of the anti-apoptotic protein BCL2 [20 , 21] . In agreement , we found that knockdown of either COL4A2 or COL4A1 in SMCs and ECs resulted in increased apoptosis with a decrease of BCL2 ( S6–S9 Figs ) . Importantly , an analysis of primary cultures of SMCs from different individuals showed an influence of rs4773144 on the rate of apoptosis , with the G allele associating with higher apoptotic rates ( Fig 4A ) . Similarly , there was an association between the G allele and lower levels of the anti-apoptotic protein BCL2 in primary cultures of ECs from different individuals ( Fig 4B ) . Further to the assays of primary cultures of vascular cells described above , we investigated if there was a relationship between rs4773144 genotype and collagen IV levels in atherosclerotic plaques . In this investigation , atherosclerotic coronary arteries from different individuals were genotyped for rs4773144 and subjected to immunohistochemical analyses using antibodies against COL4A2 and the smooth muscle cell marker SMA ( smooth muscle alpha-actin ) , respectively . The analyses showed a genotypic effect of rs4773144 on the percentages of COL4A2 positive areas in atherosclerotic plaques , with the G allele associating with lower percentages ( Fig 5A and 5B ) . It is well established that an atherosclerotic plaque typically contains a lipid core covered by a fibrous cap which is primarily composed of SMCs and extracellular matrix proteins [22] . SMC apoptosis can reduce the SMC content and the thickness of the fibrous cap , rendering the plaque prone to rupture , which can trigger thrombosis [23 , 24] . Coronary thrombosis is the most common cause of acute coronary ischemic events such as myocardial infarction ( MI ) [25 , 26] . Since our experiments described earlier showed that rs4773144 genotype influenced SMC apoptosis , we investigated whether rs4773144 genotype had an effect on atherosclerotic plaque cap thickness . Histopathological examination of atherosclerotic coronary arteries from individuals of different genotypes for rs4773144 showed a genotypic effect of rs4773144 in plaque cap thickness , with the G allele associating with thinner plaque cap and lower cap/intima ratio ( Fig 5C and 5D , S10 Fig and S11 Fig ) . Since atherosclerotic plaques with a thin cap are unstable and prone to rupture [23 , 24] and coronary atherosclerotic plaque is the primary cause of MI [25 , 26] , we investigated if there was an association between rs4773144 and MI in CHD patients . We studied a group of patients with angiographically documented coronary disease with >50% luminal stenosis , and observed a genotypic effect of rs4773144 , with the G allele associating with higher rates of MI , which remained after adjusting for age , sex , low-density-lipoprotein cholesterol , hypertension , and diabetes mellitus ( Table 1 ) . There are several novel findings from this study on the effect of the CHD-associated SNP rs4773144 located in the COL4A2 gene . At the molecular and cellular levels , a genotypic effect on COL4A2 and COL4A1 expression in vascular ECs and SMCs and on cell survival was detected . From the examination of ex vivo coronary atherosclerotic plaques , we found a relationship between the CHD risk genotype and histological features of plaque instability . Furthermore , in the study of patients with angiographically documented CHD , we observed an association between the risk allele and occurrence of MI . These findings provide a mechanistic explanation for the association between CHD risk and genetic variation at the COL4A1/COL4A2 locus . Each collagen IV molecule is composed of three α chains forming a triple helical structure , with the classic isoform containing two α1 ( IV ) chains and one α2 ( IV ) chain [27] . The COL4A1 and COL4A2 genes , which encode the α1 ( IV ) and α2 ( IV ) chains respectively , reside next to each other on chromosome 13q34 and are transcriptionally co-regulated [3–9] . Severe rare mutations in either COL4A1 or COL4A2 have been reported to cause vascular lesions and hemorrhagic stroke in humans [14 , 28–31] , suggesting that defects of either α1 ( IV ) or α2 ( IV ) can result in similar vascular phenotypes . In agreement , our study shows that SNP rs4773144 genotype affects the expression of both COL4A1 and COL4A2 , and the siRNA experiments demonstrate that knockdown of either COL4A1 or COL4A2 in vascular SMCs or ECs induces cell apoptosis ( which is line with a reported finding that a frame shift mutation in the COL4A2 gene increases rates of apoptosis of fibroblasts isolated from an individual carrying the mutation [31] ) . Apart from providing a mechanistic explanation for the association between rs4773144 and CHD , these results suggest that preserving adequate production of both of these two collagen IV genes can be a potential strategy for developing new therapeutics for the disease . Data from the ENCODE and Roadmap Epigenomics projects show DNase I hypersensitivity and histone modifications surrounding rs4773144 in a number of different types of cells and tissues including SMCs , ECs , monocytes , T-cells , B-cells , adipose , heart , skeletal muscle , brain , thymus , etc . It is possible that rs4773144 genotype may affect COL4A2/COL4A1 expression not only in vascular SMCs and ECs as demonstrated in this study but also in other cells , and that additional functional mechanisms involving other tissues may also contribute to the association between rs4773144 and CHD . Over 50 genomic loci have hitherto been identified by GWASs to be associated with CHD risk [32] . However , for many of these loci , the functional mechanisms leading to the genetic effect remain unknown . Functional characterization of these genetic variants can aid the understanding of the underlying biological mechanisms and may facilitate the translation of the genetic discoveries to therapeutic development . The findings of our present study on the CHD-related genetic variant at the COL4A1/COL4A2 locus are pertinent in this context .
NRES Committee London–City & East ( approval number: 08/H0704/140 ) and Shantou University Medical College Ethics Committee approved this research . SMCs were isolated from arteries of umbilical cords from different individuals; ECs were isolated from umbilical cord veins . Isolated SMCs and ECs were subjected to immunocytochemical examinations of the SMC marker SMA , the EC marker von Willebrand factor ( vWF ) , and the fibroblast marker discoidin domain receptor-2 ( DDR2 ) , which verified that SMCs were SMA-positive but vWF- and DDR2-negative and that ECs were vWF-positive but SMA- and DDR2-negative . Primary cultures of SMCs and ECs , up to passage 5 , were used in experiments of this study . Genomic DNA was extracted from cultured SMCs and ECs or from sections of formaldehyde-fixed paraffin-embedded blocks of atherosclerotic coronary arteries using the Wizard SV Genomic DNA Purification System ( Promega ) . rs4773144 genotypes were determined with the use of the TaqMan SNP genotyping assay . Accuracy of the genotyping results was verified by sequencing of a random selection of the samples . Total RNA samples were prepared from primary cultures of SMCs and ECs , with the use of the SV Total RNA Isolation System ( Promega ) . RNA was reverse transcribed into cDNA using random primers ( Promega ) and M-MLV reverse transcriptase ( Promega ) . The resultant cDNA was subjected to real-time polymerase chain reactions for COL4A1 , COL4A2 , and β-actin , respectively , with the use of TaqMan Gene Expression Assays . The 2-∆∆CT method [33] was used to ascertain differences between genotypes in COL4A1 and COL4A2 levels standardized against the reference gene β-actin . SMCs and ECs , heterozygous for SNP rs4773144 , were crosslinked by incubation in formaldehyde and then incubated with glycine to quench formaldehyde . Subsequently , cells were lysed and chromatin sheared to 200–1 , 000bp in length by sonication . Aliquots of the samples were incubated with protein G-agarose beads and then further with an anti-human RNA polymerase II antibody ( Santa Cruz Biotechnology , sc-9001 ) . DNA-protein-antibody complexes bound to protein G-agarose beads were precipitated by centrifugation and de-crosslinked . Sheared chromatin samples and immunoprecipitated DNA samples were subjected to allelic imbalance analyses of SNP rs4773144 , with the use of the TaqMan method to determine the Ct values for the A allele ( detected by a VIC fluorescein-labeled probe ) and G allele ( detected by a FAM fluorescent dye-labeled probe ) , respectively . Additionally , heterozygous ECs were subjected to chromatin immunoprecipitation using an anti-human H3k27Ac antibody ( Abcam ab4729 ) and then TaqMan assay to determine the G allele versus A allele ratio . A 350 base pair DNA sequence encompassing the rs4773144 site of the A and G alleles , respectively , was amplified by PCR and inserted into the pGL3-promoter vector ( Promega ) containing a firefly luciferase reporter gene . The resultant construct containing the inserted DNA sequence corresponding to either the A or G allele of rs4773144 was mixed with a plasmid ( pRL-TK , Promega ) containing a renilla luciferase gene and transfected into cultured vascular SMCs . At 48 hours after transfection , the transfectants were lysed , and the activities of firefly luciferase and renilla luciferase in the lysates were measured . The ratio of firefly luciferase activity to renilla luciferase activity was used as a measurement of the transcription modulating activity of the inserted DNA sequence encompassing the rs4773144 site of the A and G alleles , respectively . Biotin-labelled , double-stranded 25-mer oligonucleotides corresponding to the sequences at and surrounding the SNP rs4773144 site were used as probes . The probe sequences were: 5'-CCTTTCACGGGAACTGGGAACTTAA-3' ( A allele ) and 5’-CCTTTCACGGGAGCTGGGAACTTAA-3’ ( G allele ) , respectively . The probes were individually incubated with nuclear extracts of cultured ECs , in the presence or absence of unlabelled oligonucleotide competitors in molar excess or an anti-phospho-STAT3 antibody ( Cell Signaling Technology , #9131 ) , followed by non-denaturing polyacrylamide gel electrophoresis . Free probes and probe-protein complexes were detected using a LightShift Chemiluminescent EMSA kit ( Pierce Biotechnology , 20148 ) . Three independent experiments were carried out . SMCs and ECs were transfected with either COL4A2 siRNA ( ThermoFisher Scientific , 4457308 ) , COL4A1 siRNA ( ThermoFisher Scientific , AM16708 ) , or control siRNA ( ThermoFisher Scientific , 4390843 ) , with the use of Lipofectamine RNAiMAX transfection reagent ( Invitrogen , 13778150 ) . COL4A2 or COL4A1 knockdown was verified by immunoblotting analysis . Cell lysates were prepared by incubating cells with a lysis buffer containing a protease inhibitor cocktail . An aliquot of 20μg proteins from each sample was subjected to Tris-glycine , sodium-dodecyl-sulfate , polyacrylamide gel electrophoresis , followed by standard immunoblotting analysis with an anti-COL4A2 antibody ( Abcam , ab69782 ) , an anti-COL4A1 antibody ( Abnova , PAB17326 ) , an anti-BCL2 antibody ( Abcam , ab32124 ) , an anti-β-actin antibody , or an anti-HSC70 antibody ( Santa Cruz , sc7298 ) . SMCs ( same number ) or ECs ( same number ) transfected with either the COL4A2 siRNA , COL4A1 siRNA or control siRNA , or untransfected primary SMCs ( same number ) from different individuals , were cultured for 72 hours , and then detached and counted . Additionally , cells were subjected to apoptosis assays with the use of Annexin V-FITC apoptosis detection kit ( Beyotime Institute of Biotechnology , C1062 , for transfected cells ) or Cell Death Detection ELISAPLUS kit ( Roche , 11774425001 , for untransfected cells ) to quantify histone-complexed DNA fragments . Formaldehyde-fixed paraffin-embedded sections of atherosclerotic coronary arteries from autopsies were deparaffinised , rehydrated , and incubated in sodium citrate for antigen retrieval . The sections were subjected to peroxidase immunostaining with a mouse anti-human smooth muscle α-actin ( SMA ) antibody ( Dako , M-0635 ) . A subset of the collection was subjected to peroxidase/alkaline phosphatase double immunostaining with the mouse anti-human smooth muscle α-actin ( SMA ) antibody ( Dako , M-0635 ) and a rabbit anti-human COL4A2 antibody ( Abcam , ab69782 ) . Chromagens were diaminobenzidine and nitrobluetetrazolium/bromo-chloro-indolyl phosphate , respectively . Images of the sections were captured using a microscope with an imaging system and analyzed using Image-Pro software to determine the sizes of positive immunostain areas , atherosclerotic plaque cap thickness , and intima thickness . We studied 1125 consecutive patients undergoing diagnostic or interventional coronary angiography in the First Affiliated Hospital of Shantou University Medical College . All subjects were Chinese . We collected demographic and clinical data including age , sex , plasma levels of total cholesterol , low-density-lipoprotein cholesterol , high-density lipoprotein cholesterol and triglycerides , coronary angiographic findings , incident or prevalent MI diagnosed according to the World Health Organization criteria , systolic and diastolic blood pressure , and the presence or absence of diabetes mellitus . Of the 1125 subjects , 655 had significant angiographically documented CHD as having >50% diameter stenosis in ≥1 major epicardial coronary artery . Among them , a total of 227 subjects had incident or prevalent MI . The study was approved by the appropriate research ethics committee . The data were analyzed anonymously . Variables not in normal distribution were normalized by logarithmic transformation . Linear regression analyses were performed to test differences between genotypes in COL4A1 and COL4A2 expression levels , apoptosis assay result , BCL2 immunoblotting band intensity ( standardized against HSC70 band intensity ) , percentage of COL4A2 stain areas in total atherosclerotic plaque area , atherosclerotic plaque cap thickness , and atherosclerotic plaque cap/intima ratio , in an additive genetic model . Student’s t-tests were used to ascertain allelic differences in the allelic expression imbalance assays and the chromatin immunoprecipitation assays . Student’s t-tests were also used to test differences between cells transfected with either the rs4773144 A allele plasmid or G allele plasmid in firefly luciferase activity after standardized against renilla luciferase activity in the luciferase assays , and between cells transfected with COL4A2 siRNA or COL4A1 siRNA and cells transfected with control siRNA in cell count , proliferation assay result , apoptosis assay result , and BCL2 immunoblotting band intensity ( standardized against β-actin band intensity ) , respectively . Logistic regression analyses and chi-squared tests were carried out to ascertain genotypic and allelic association with MI in CHD patients . All p values were two-sided .
|
People who carry certain variants in their DNA are genetically predisposed to suffer from coronary heart disease ( CHD ) caused by abnormal tissue buildup ( known as atherosclerosis ) and blood clotting in the blood vessels of the heart . One of the DNA variants reported to increase CHD risk is named single nucleotide polymorphism rs4773144 . It is still unclear as to why this DNA variant has an effect on CHD risk . In this study , by studying blood vessel cells from many people , we found that the DNA variant affects the production of two collagen genes and vascular cell survival . By examining atherosclerotic tissues from many patients , we discovered that the atherosclerotic tissues of patients who carry the rs4773144 variant are structurally more likely to break down and cause blood clotting which can lead to a heart attack . Furthermore , by studying a group of CHD patients , we noticed that those who carry the rs4773144 variant do have higher rates of heart attack . These findings are useful for understanding why this DNA variant has an impact on CHD risk and suggest that preserving adequate production of these two collagen genes may reduce the risk of heart attack in CHD patients , a potential strategy for development of therapeutics for the disease .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"luciferase",
"molecular",
"probe",
"techniques",
"enzymes",
"gene",
"regulation",
"cell",
"processes",
"immunoblotting",
"enzymology",
"collagens",
"coronary",
"heart",
"disease",
"microscopy",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"cardiology",
"small",
"interfering",
"rnas",
"proteins",
"oxidoreductases",
"gene",
"expression",
"scanning",
"electron",
"microscopy",
"molecular",
"biology",
"myocardial",
"infarction",
"biochemistry",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"apoptosis",
"genetics",
"electron",
"microscopy",
"biology",
"and",
"life",
"sciences",
"non-coding",
"rna",
"vascular",
"medicine"
] |
2016
|
Coronary-Heart-Disease-Associated Genetic Variant at the COL4A1/COL4A2 Locus Affects COL4A1/COL4A2 Expression, Vascular Cell Survival, Atherosclerotic Plaque Stability and Risk of Myocardial Infarction
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.